[llvm] [AArch64] Fix SVE cost model for various math intrinsics (PR #184358)
David Sherwood via llvm-commits
llvm-commits at lists.llvm.org
Wed Mar 4 02:45:08 PST 2026
https://github.com/david-arm updated https://github.com/llvm/llvm-project/pull/184358
>From 8ca0fc9506c69a0742249f3d796b66d92d97c129 Mon Sep 17 00:00:00 2001
From: David Sherwood <david.sherwood at arm.com>
Date: Tue, 3 Mar 2026 17:10:40 +0000
Subject: [PATCH 1/3] Add tests
---
.../CostModel/AArch64/sve-intrinsics.ll | 304 ++++++++++++++++++
.../AArch64/veclib-intrinsic-calls.ll | 256 +++++++--------
2 files changed, 432 insertions(+), 128 deletions(-)
diff --git a/llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll b/llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll
index 0836ef1b352c9..c3242bd56f133 100644
--- a/llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll
@@ -1447,6 +1447,310 @@ define void @match() #3 {
ret void
}
+define void @sqrt() #1 {
+; CHECK-VSCALE-1-LABEL: 'sqrt'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'sqrt'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'sqrt'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
+ %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
+ %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
+ %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @fabs() #1 {
+; CHECK-VSCALE-1-LABEL: 'fabs'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'fabs'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'fabs'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
+ %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
+ %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
+ %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @ceil() #1 {
+; CHECK-VSCALE-1-LABEL: 'ceil'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'ceil'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'ceil'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
+ %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
+ %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
+ %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @floor() #1 {
+; CHECK-VSCALE-1-LABEL: 'floor'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'floor'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'floor'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
+ %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
+ %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
+ %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @nearbyint() #1 {
+; CHECK-VSCALE-1-LABEL: 'nearbyint'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'nearbyint'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'nearbyint'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
+ %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
+ %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
+ %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @round() #1 {
+; CHECK-VSCALE-1-LABEL: 'round'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'round'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'round'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
+ %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
+ %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
+ %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @rint() #1 {
+; CHECK-VSCALE-1-LABEL: 'rint'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'rint'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'rint'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
+ %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
+ %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
+ %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @roundeven() #1 {
+; CHECK-VSCALE-1-LABEL: 'roundeven'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'roundeven'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'roundeven'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
+ %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
+ %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
+ %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @trunc() #1 {
+; CHECK-VSCALE-1-LABEL: 'trunc'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'trunc'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'trunc'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
+ %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
+ %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
+ %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
+
+ ret void
+}
+
+define void @abs() #1 {
+; CHECK-VSCALE-1-LABEL: 'abs'
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; CHECK-VSCALE-2-LABEL: 'abs'
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+; TYPE_BASED_ONLY-LABEL: 'abs'
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
+ %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
+ %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
+ %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
+ %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
+
+ ret void
+}
+
declare <vscale x 16 x i1> @llvm.get.active.lane.mask.nxv16i1.i64(i64, i64)
declare <vscale x 8 x i1> @llvm.get.active.lane.mask.nxv8i1.i64(i64, i64)
declare <vscale x 4 x i1> @llvm.get.active.lane.mask.nxv4i1.i64(i64, i64)
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/veclib-intrinsic-calls.ll b/llvm/test/Transforms/LoopVectorize/AArch64/veclib-intrinsic-calls.ll
index 946c51ab922c0..abf4dead6fcd4 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/veclib-intrinsic-calls.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/veclib-intrinsic-calls.ll
@@ -38,7 +38,7 @@ define void @acos_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @acos_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1:[0-9]+]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_acos(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_acos(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @acos_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1:[0-9]+]] {
@@ -46,7 +46,7 @@ define void @acos_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @acos_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1:[0-9]+]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svacos_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svacos_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -86,7 +86,7 @@ define void @acos_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @acos_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_acosf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_acosf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @acos_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -94,7 +94,7 @@ define void @acos_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @acos_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svacos_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svacos_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -137,7 +137,7 @@ define void @asin_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @asin_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_asin(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_asin(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @asin_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -145,7 +145,7 @@ define void @asin_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @asin_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svasin_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svasin_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -185,7 +185,7 @@ define void @asin_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @asin_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_asinf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_asinf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @asin_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -193,7 +193,7 @@ define void @asin_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @asin_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svasin_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svasin_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -236,7 +236,7 @@ define void @atan_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @atan_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_atan(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_atan(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @atan_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -244,7 +244,7 @@ define void @atan_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @atan_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svatan_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svatan_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -284,7 +284,7 @@ define void @atan_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @atan_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_atanf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_atanf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @atan_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -292,7 +292,7 @@ define void @atan_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @atan_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svatan_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svatan_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -335,7 +335,7 @@ define void @atan2_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @atan2_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxvv_atan2(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxvv_atan2(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @atan2_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -343,7 +343,7 @@ define void @atan2_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @atan2_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svatan2_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svatan2_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -383,7 +383,7 @@ define void @atan2_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @atan2_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxvv_atan2f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxvv_atan2f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @atan2_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -391,7 +391,7 @@ define void @atan2_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @atan2_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svatan2_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svatan2_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -433,7 +433,7 @@ define void @ceil_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @ceil_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @ceil_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -441,7 +441,7 @@ define void @ceil_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @ceil_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -480,7 +480,7 @@ define void @ceil_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @ceil_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @ceil_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -488,7 +488,7 @@ define void @ceil_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @ceil_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -531,7 +531,7 @@ define void @copysign_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @copysign_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.copysign.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.copysign.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
; ARMPL-NEON-LABEL: define void @copysign_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -539,7 +539,7 @@ define void @copysign_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @copysign_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.copysign.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.copysign.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -579,7 +579,7 @@ define void @copysign_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @copysign_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.copysign.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.copysign.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
; ARMPL-NEON-LABEL: define void @copysign_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -587,7 +587,7 @@ define void @copysign_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @copysign_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.copysign.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.copysign.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -630,7 +630,7 @@ define void @cos_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @cos_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_cos(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_cos(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @cos_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -638,7 +638,7 @@ define void @cos_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @cos_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svcos_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svcos_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -678,7 +678,7 @@ define void @cos_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @cos_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_cosf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_cosf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @cos_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -686,7 +686,7 @@ define void @cos_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @cos_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svcos_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svcos_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -729,7 +729,7 @@ define void @cosh_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @cosh_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_cosh(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_cosh(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @cosh_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -737,7 +737,7 @@ define void @cosh_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @cosh_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svcosh_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svcosh_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -777,7 +777,7 @@ define void @cosh_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @cosh_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_coshf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_coshf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @cosh_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -785,7 +785,7 @@ define void @cosh_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @cosh_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svcosh_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svcosh_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -828,7 +828,7 @@ define void @exp_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @exp_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_exp(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_exp(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @exp_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -836,7 +836,7 @@ define void @exp_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @exp_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svexp_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svexp_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -876,7 +876,7 @@ define void @exp_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @exp_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_expf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_expf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @exp_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -884,7 +884,7 @@ define void @exp_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @exp_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svexp_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svexp_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -927,7 +927,7 @@ define void @exp10_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @exp10_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_exp10(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_exp10(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @exp10_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -935,7 +935,7 @@ define void @exp10_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @exp10_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svexp10_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svexp10_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -975,7 +975,7 @@ define void @exp10_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @exp10_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_exp10f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_exp10f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @exp10_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -983,7 +983,7 @@ define void @exp10_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @exp10_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svexp10_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svexp10_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1026,7 +1026,7 @@ define void @exp2_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @exp2_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_exp2(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_exp2(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @exp2_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1034,7 +1034,7 @@ define void @exp2_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @exp2_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svexp2_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svexp2_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1074,7 +1074,7 @@ define void @exp2_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @exp2_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_exp2f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_exp2f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @exp2_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1082,7 +1082,7 @@ define void @exp2_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @exp2_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svexp2_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svexp2_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1124,7 +1124,7 @@ define void @fabs_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @fabs_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @fabs_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1132,7 +1132,7 @@ define void @fabs_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @fabs_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -1171,7 +1171,7 @@ define void @fabs_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @fabs_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @fabs_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1179,7 +1179,7 @@ define void @fabs_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @fabs_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -1221,7 +1221,7 @@ define void @floor_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @floor_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @floor_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1229,7 +1229,7 @@ define void @floor_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @floor_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -1268,7 +1268,7 @@ define void @floor_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @floor_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @floor_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1276,7 +1276,7 @@ define void @floor_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @floor_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -1319,7 +1319,7 @@ define void @fma_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @fma_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.fma.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.fma.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
; ARMPL-NEON-LABEL: define void @fma_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1327,7 +1327,7 @@ define void @fma_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @fma_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.fma.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.fma.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -1367,7 +1367,7 @@ define void @fma_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @fma_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.fma.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.fma.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
; ARMPL-NEON-LABEL: define void @fma_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1375,7 +1375,7 @@ define void @fma_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @fma_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.fma.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.fma.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -1418,7 +1418,7 @@ define void @log_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @log_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_log(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_log(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @log_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1426,7 +1426,7 @@ define void @log_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @log_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svlog_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svlog_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1466,7 +1466,7 @@ define void @log_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @log_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_logf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_logf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @log_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1474,7 +1474,7 @@ define void @log_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @log_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svlog_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svlog_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1517,7 +1517,7 @@ define void @log10_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @log10_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_log10(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_log10(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @log10_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1525,7 +1525,7 @@ define void @log10_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @log10_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svlog10_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svlog10_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1565,7 +1565,7 @@ define void @log10_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @log10_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_log10f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_log10f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @log10_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1573,7 +1573,7 @@ define void @log10_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @log10_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svlog10_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svlog10_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1616,7 +1616,7 @@ define void @log2_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @log2_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_log2(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_log2(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @log2_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1624,7 +1624,7 @@ define void @log2_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @log2_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svlog2_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svlog2_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1664,7 +1664,7 @@ define void @log2_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @log2_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_log2f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_log2f(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @log2_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1672,7 +1672,7 @@ define void @log2_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @log2_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svlog2_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svlog2_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -1714,7 +1714,7 @@ define void @maxnum_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @maxnum_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.maxnum.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.maxnum.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
; ARMPL-NEON-LABEL: define void @maxnum_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1722,7 +1722,7 @@ define void @maxnum_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @maxnum_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.maxnum.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.maxnum.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -1761,7 +1761,7 @@ define void @maxnum_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @maxnum_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.maxnum.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.maxnum.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
; ARMPL-NEON-LABEL: define void @maxnum_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1769,7 +1769,7 @@ define void @maxnum_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @maxnum_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.maxnum.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.maxnum.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -1811,7 +1811,7 @@ define void @minnum_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @minnum_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.minnum.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.minnum.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
; ARMPL-NEON-LABEL: define void @minnum_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1819,7 +1819,7 @@ define void @minnum_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @minnum_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.minnum.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.minnum.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -1858,7 +1858,7 @@ define void @minnum_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @minnum_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.minnum.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.minnum.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
; ARMPL-NEON-LABEL: define void @minnum_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1866,7 +1866,7 @@ define void @minnum_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @minnum_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.minnum.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.minnum.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -1912,7 +1912,7 @@ define void @modf_f64(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; SLEEF-SVE-LABEL: define void @modf_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP12:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.modf.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP7:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.modf.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
; SLEEF-SVE: [[CALL:%.*]] = tail call { double, double } @llvm.modf.f64(double [[IN:%.*]])
;
; ARMPL-NEON-LABEL: define void @modf_f64
@@ -1922,7 +1922,7 @@ define void @modf_f64(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; ARMPL-SVE-LABEL: define void @modf_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP12:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @armpl_svmodf_f64_stret_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP7:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @armpl_svmodf_f64_stret_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
; ARMPL-SVE: [[CALL:%.*]] = tail call { double, double } @llvm.modf.f64(double [[IN:%.*]])
;
entry:
@@ -1970,7 +1970,7 @@ define void @modf_f32(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; SLEEF-SVE-LABEL: define void @modf_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP12:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.modf.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP7:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.modf.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
; SLEEF-SVE: [[CALL:%.*]] = tail call { float, float } @llvm.modf.f32(float [[IN:%.*]])
;
; ARMPL-NEON-LABEL: define void @modf_f32
@@ -1980,7 +1980,7 @@ define void @modf_f32(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; ARMPL-SVE-LABEL: define void @modf_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP12:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @armpl_svmodf_f32_stret_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP7:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @armpl_svmodf_f32_stret_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
; ARMPL-SVE: [[CALL:%.*]] = tail call { float, float } @llvm.modf.f32(float [[IN:%.*]])
;
entry:
@@ -2027,7 +2027,7 @@ define void @nearbyint_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @nearbyint_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @nearbyint_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2035,7 +2035,7 @@ define void @nearbyint_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @nearbyint_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -2074,7 +2074,7 @@ define void @nearbyint_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @nearbyint_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @nearbyint_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2082,7 +2082,7 @@ define void @nearbyint_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @nearbyint_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -2125,7 +2125,7 @@ define void @pow_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @pow_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxvv_pow(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxvv_pow(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @pow_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2133,7 +2133,7 @@ define void @pow_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @pow_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.pow.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.pow.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x double> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -2173,7 +2173,7 @@ define void @pow_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @pow_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxvv_powf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxvv_powf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @pow_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2181,7 +2181,7 @@ define void @pow_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @pow_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.pow.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.pow.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x float> [[WIDE_MASKED_LOAD]])
;
entry:
br label %for.body
@@ -2223,7 +2223,7 @@ define void @rint_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @rint_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @rint_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2231,7 +2231,7 @@ define void @rint_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @rint_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -2270,7 +2270,7 @@ define void @rint_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @rint_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @rint_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2278,7 +2278,7 @@ define void @rint_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @rint_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -2320,7 +2320,7 @@ define void @round_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @round_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @round_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2328,7 +2328,7 @@ define void @round_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @round_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -2367,7 +2367,7 @@ define void @round_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @round_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @round_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2375,7 +2375,7 @@ define void @round_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @round_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -2418,7 +2418,7 @@ define void @sin_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @sin_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_sin(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_sin(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @sin_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2426,7 +2426,7 @@ define void @sin_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @sin_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svsin_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svsin_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -2466,7 +2466,7 @@ define void @sin_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @sin_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_sinf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_sinf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @sin_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2474,7 +2474,7 @@ define void @sin_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @sin_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svsin_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svsin_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -2520,7 +2520,7 @@ define void @sincos_f64(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; SLEEF-SVE-LABEL: define void @sincos_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP12:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.sincos.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP7:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.sincos.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
; SLEEF-SVE: [[CALL:%.*]] = tail call { double, double } @llvm.sincos.f64(double [[IN:%.*]])
;
; ARMPL-NEON-LABEL: define void @sincos_f64
@@ -2530,7 +2530,7 @@ define void @sincos_f64(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; ARMPL-SVE-LABEL: define void @sincos_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP12:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @armpl_svcexpi_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP7:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @armpl_svcexpi_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
; ARMPL-SVE: [[CALL:%.*]] = tail call { double, double } @llvm.sincos.f64(double [[IN:%.*]])
;
entry:
@@ -2578,7 +2578,7 @@ define void @sincos_f32(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; SLEEF-SVE-LABEL: define void @sincos_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP12:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.sincospi.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP7:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.sincospi.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
; SLEEF-SVE: [[CALL:%.*]] = tail call { float, float } @llvm.sincospi.f32(float [[IN:%.*]])
;
; ARMPL-NEON-LABEL: define void @sincos_f32
@@ -2588,7 +2588,7 @@ define void @sincos_f32(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; ARMPL-SVE-LABEL: define void @sincos_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP12:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @armpl_svcexpipi_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP7:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @armpl_svcexpipi_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
; ARMPL-SVE: [[CALL:%.*]] = tail call { float, float } @llvm.sincospi.f32(float [[IN:%.*]])
;
entry:
@@ -2639,7 +2639,7 @@ define void @sincospi_f64(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; SLEEF-SVE-LABEL: define void @sincospi_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP12:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.sincospi.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP7:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @llvm.sincospi.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
; SLEEF-SVE: [[CALL:%.*]] = tail call { double, double } @llvm.sincospi.f64(double [[IN:%.*]])
;
; ARMPL-NEON-LABEL: define void @sincospi_f64
@@ -2649,7 +2649,7 @@ define void @sincospi_f64(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; ARMPL-SVE-LABEL: define void @sincospi_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP12:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @armpl_svcexpipi_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP7:%.*]] = call { <vscale x 2 x double>, <vscale x 2 x double> } @armpl_svcexpipi_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
; ARMPL-SVE: [[CALL:%.*]] = tail call { double, double } @llvm.sincospi.f64(double [[IN:%.*]])
;
entry:
@@ -2697,7 +2697,7 @@ define void @sincospi_f32(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; SLEEF-SVE-LABEL: define void @sincospi_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP12:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.sincospi.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP7:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @llvm.sincospi.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
; SLEEF-SVE: [[CALL:%.*]] = tail call { float, float } @llvm.sincospi.f32(float [[IN:%.*]])
;
; ARMPL-NEON-LABEL: define void @sincospi_f32
@@ -2707,7 +2707,7 @@ define void @sincospi_f32(ptr noalias %in.ptr, ptr %out_a.ptr, ptr %out_b.ptr) {
;
; ARMPL-SVE-LABEL: define void @sincospi_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_A_PTR:%.*]], ptr [[OUT_B_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP12:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @armpl_svcexpipi_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP7:%.*]] = call { <vscale x 4 x float>, <vscale x 4 x float> } @armpl_svcexpipi_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
; ARMPL-SVE: [[CALL:%.*]] = tail call { float, float } @llvm.sincospi.f32(float [[IN:%.*]])
;
entry:
@@ -2755,7 +2755,7 @@ define void @sinh_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @sinh_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_sinh(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_sinh(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @sinh_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2763,7 +2763,7 @@ define void @sinh_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @sinh_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svsinh_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svsinh_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -2803,7 +2803,7 @@ define void @sinh_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @sinh_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_sinhf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_sinhf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @sinh_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2811,7 +2811,7 @@ define void @sinh_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @sinh_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svsinh_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svsinh_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -2853,7 +2853,7 @@ define void @sqrt_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @sqrt_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @sqrt_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2861,7 +2861,7 @@ define void @sqrt_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @sqrt_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -2900,7 +2900,7 @@ define void @sqrt_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @sqrt_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @sqrt_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2908,7 +2908,7 @@ define void @sqrt_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @sqrt_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -2951,7 +2951,7 @@ define void @tan_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @tan_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_tan(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_tan(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @tan_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2959,7 +2959,7 @@ define void @tan_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @tan_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svtan_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svtan_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -2999,7 +2999,7 @@ define void @tan_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @tan_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_tanf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_tanf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @tan_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -3007,7 +3007,7 @@ define void @tan_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @tan_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svtan_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svtan_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -3050,7 +3050,7 @@ define void @tanh_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @tanh_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_tanh(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @_ZGVsMxv_tanh(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @tanh_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -3058,7 +3058,7 @@ define void @tanh_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @tanh_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @armpl_svtanh_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @armpl_svtanh_f64_x(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]], <vscale x 2 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -3098,7 +3098,7 @@ define void @tanh_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @tanh_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_tanhf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @_ZGVsMxv_tanhf(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
; ARMPL-NEON-LABEL: define void @tanh_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -3106,7 +3106,7 @@ define void @tanh_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @tanh_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @armpl_svtanh_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @armpl_svtanh_f32_x(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]], <vscale x 4 x i1> [[ACTIVE_LANE_MASK:%.*]])
;
entry:
br label %for.body
@@ -3148,7 +3148,7 @@ define void @trunc_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @trunc_f64
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @trunc_f64
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -3156,7 +3156,7 @@ define void @trunc_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @trunc_f64
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
@@ -3195,7 +3195,7 @@ define void @trunc_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; SLEEF-SVE-LABEL: define void @trunc_f32
; SLEEF-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; SLEEF-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; SLEEF-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
; ARMPL-NEON-LABEL: define void @trunc_f32
; ARMPL-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -3203,7 +3203,7 @@ define void @trunc_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; ARMPL-SVE-LABEL: define void @trunc_f32
; ARMPL-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; ARMPL-SVE: [[TMP8:%.*]] = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
+; ARMPL-SVE: [[TMP3:%.*]] = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> [[WIDE_MASKED_LOAD:%.*]])
;
entry:
br label %for.body
>From 21a2103f5b57caaec33aa5f7537dea2de2ce857a Mon Sep 17 00:00:00 2001
From: David Sherwood <david.sherwood at arm.com>
Date: Tue, 3 Mar 2026 17:10:49 +0000
Subject: [PATCH 2/3] [AArch64] Fix SVE cost model for various math intrinsics
The implementation of getIntrinsicInstrCost in BasicTTIImpl
assumes that for some intrinsics if we're using custom
lowering for the equivalent DAG node that the cost needs to
be 2, instead of 1 for legal ops. However, even though we
use custom lowering for these scalable vector operations
when SVE is available, we still end up generating the same
efficient codegen as fixed-width. This patch deals with a
few obvious intrinsics that we know get lowered to something
sensible and return the same cost as NEON, i.e. 1.
---
.../AArch64/AArch64TargetTransformInfo.cpp | 28 ++
.../CostModel/AArch64/sve-intrinsics.ll | 259 +++++++++---------
.../Analysis/CostModel/AArch64/sve-math.ll | 2 +-
.../Analysis/CostModel/AArch64/sve-min-max.ll | 136 ++++-----
.../pr151664-cost-hoisted-vector-scalable.ll | 27 +-
.../AArch64/veclib-intrinsic-calls.ll | 60 ++--
6 files changed, 284 insertions(+), 228 deletions(-)
diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
index ff957e0582aad..7d24986b7981a 100644
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -696,6 +696,13 @@ AArch64TTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
break;
}
case Intrinsic::abs: {
+ if (isa<ScalableVectorType>(RetTy) && ST->isSVEorStreamingSVEAvailable()) {
+ auto LT = getTypeLegalizationCost(RetTy);
+ return LT.first;
+ }
+
+ // TODO: Why can't we just always return
+ // getTypeLegalizationCost(RetTy).first for all types?
static const auto ValidAbsTys = {MVT::v8i8, MVT::v16i8, MVT::v4i16,
MVT::v8i16, MVT::v2i32, MVT::v4i32,
MVT::v2i64};
@@ -1114,6 +1121,27 @@ AArch64TTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
return getCallInstrCost(nullptr, RetTy, ICA.getArgTypes(), CostKind);
break;
}
+ case Intrinsic::sqrt:
+ case Intrinsic::fabs:
+ case Intrinsic::ceil:
+ case Intrinsic::floor:
+ case Intrinsic::nearbyint:
+ case Intrinsic::round:
+ case Intrinsic::rint:
+ case Intrinsic::roundeven:
+ case Intrinsic::trunc:
+ case Intrinsic::minnum:
+ case Intrinsic::maxnum:
+ case Intrinsic::minimum:
+ case Intrinsic::maximum:
+ case Intrinsic::minimumnum:
+ case Intrinsic::maximumnum: {
+ if (isa<ScalableVectorType>(RetTy) && ST->isSVEorStreamingSVEAvailable()) {
+ auto LT = getTypeLegalizationCost(RetTy);
+ return LT.first;
+ }
+ break;
+ }
default:
break;
}
diff --git a/llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll b/llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll
index c3242bd56f133..2a81448db1143 100644
--- a/llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll
@@ -182,10 +182,10 @@ define void @reductions(<vscale x 4 x i32> %v0, <vscale x 4 x i64> %v1, <vscale
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:3 CodeSize:3 Lat:5 SizeLat:3 for: %fadd_nxv4f64 = call fast double @llvm.vector.reduce.fadd.nxv4f64(double 0.000000e+00, <vscale x 4 x double> %v3)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of Invalid for: %fmin_nxv1f32 = call fast float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fmin_nxv4f32 = call fast float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v2)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %fmin_nxv4f64 = call fast double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v3)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 3 for: %fmin_nxv4f64 = call fast double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v3)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of Invalid for: %fmax_nxv1f32 = call fast float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fmax_nxv4f32 = call fast float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v2)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %fmax_nxv4f64 = call fast double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v3)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 3 for: %fmax_nxv4f64 = call fast double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v3)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'reductions'
@@ -221,10 +221,10 @@ define void @reductions(<vscale x 4 x i32> %v0, <vscale x 4 x i64> %v1, <vscale
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:3 CodeSize:3 Lat:5 SizeLat:3 for: %fadd_nxv4f64 = call fast double @llvm.vector.reduce.fadd.nxv4f64(double 0.000000e+00, <vscale x 4 x double> %v3)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of Invalid for: %fmin_nxv1f32 = call fast float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fmin_nxv4f32 = call fast float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v2)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %fmin_nxv4f64 = call fast double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v3)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 3 for: %fmin_nxv4f64 = call fast double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v3)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of Invalid for: %fmax_nxv1f32 = call fast float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fmax_nxv4f32 = call fast float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v2)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %fmax_nxv4f64 = call fast double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v3)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 3 for: %fmax_nxv4f64 = call fast double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v3)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'reductions'
@@ -260,10 +260,10 @@ define void @reductions(<vscale x 4 x i32> %v0, <vscale x 4 x i64> %v1, <vscale
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:3 CodeSize:3 Lat:5 SizeLat:3 for: %fadd_nxv4f64 = call fast double @llvm.vector.reduce.fadd.nxv4f64(double 0.000000e+00, <vscale x 4 x double> %v3)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of Invalid for: %fmin_nxv1f32 = call fast float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fmin_nxv4f32 = call fast float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v2)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %fmin_nxv4f64 = call fast double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v3)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 3 for: %fmin_nxv4f64 = call fast double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v3)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of Invalid for: %fmax_nxv1f32 = call fast float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fmax_nxv4f32 = call fast float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v2)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %fmax_nxv4f64 = call fast double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v3)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 3 for: %fmax_nxv4f64 = call fast double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v3)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%add_nxv1i32 = call i32 @llvm.vector.reduce.add.nxv1i32(<vscale x 1 x i32> poison)
@@ -1449,24 +1449,24 @@ define void @match() #3 {
define void @sqrt() #1 {
; CHECK-VSCALE-1-LABEL: 'sqrt'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'sqrt'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'sqrt'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv4f32 = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %sqrt.nxv2f64 = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %sqrt.nxv4f64 = call <vscale x 4 x double> @llvm.sqrt.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%sqrt.nxv8f16 = call <vscale x 8 x half> @llvm.sqrt.nxv8f16(<vscale x 8 x half> poison)
@@ -1479,24 +1479,24 @@ define void @sqrt() #1 {
define void @fabs() #1 {
; CHECK-VSCALE-1-LABEL: 'fabs'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'fabs'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'fabs'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv4f32 = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %fabs.nxv2f64 = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %fabs.nxv4f64 = call <vscale x 4 x double> @llvm.fabs.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%fabs.nxv8f16 = call <vscale x 8 x half> @llvm.fabs.nxv8f16(<vscale x 8 x half> poison)
@@ -1509,24 +1509,24 @@ define void @fabs() #1 {
define void @ceil() #1 {
; CHECK-VSCALE-1-LABEL: 'ceil'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'ceil'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'ceil'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv4f32 = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %ceil.nxv2f64 = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %ceil.nxv4f64 = call <vscale x 4 x double> @llvm.ceil.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%ceil.nxv8f16 = call <vscale x 8 x half> @llvm.ceil.nxv8f16(<vscale x 8 x half> poison)
@@ -1539,24 +1539,24 @@ define void @ceil() #1 {
define void @floor() #1 {
; CHECK-VSCALE-1-LABEL: 'floor'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'floor'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'floor'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %floor.nxv4f32 = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %floor.nxv2f64 = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %floor.nxv4f64 = call <vscale x 4 x double> @llvm.floor.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%floor.nxv8f16 = call <vscale x 8 x half> @llvm.floor.nxv8f16(<vscale x 8 x half> poison)
@@ -1569,24 +1569,24 @@ define void @floor() #1 {
define void @nearbyint() #1 {
; CHECK-VSCALE-1-LABEL: 'nearbyint'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'nearbyint'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'nearbyint'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv4f32 = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %nearbyint.nxv2f64 = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %nearbyint.nxv4f64 = call <vscale x 4 x double> @llvm.nearbyint.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%nearbyint.nxv8f16 = call <vscale x 8 x half> @llvm.nearbyint.nxv8f16(<vscale x 8 x half> poison)
@@ -1599,24 +1599,24 @@ define void @nearbyint() #1 {
define void @round() #1 {
; CHECK-VSCALE-1-LABEL: 'round'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'round'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'round'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %round.nxv4f32 = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %round.nxv2f64 = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %round.nxv4f64 = call <vscale x 4 x double> @llvm.round.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%round.nxv8f16 = call <vscale x 8 x half> @llvm.round.nxv8f16(<vscale x 8 x half> poison)
@@ -1629,24 +1629,24 @@ define void @round() #1 {
define void @rint() #1 {
; CHECK-VSCALE-1-LABEL: 'rint'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'rint'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'rint'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %rint.nxv4f32 = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %rint.nxv2f64 = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %rint.nxv4f64 = call <vscale x 4 x double> @llvm.rint.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%rint.nxv8f16 = call <vscale x 8 x half> @llvm.rint.nxv8f16(<vscale x 8 x half> poison)
@@ -1659,24 +1659,24 @@ define void @rint() #1 {
define void @roundeven() #1 {
; CHECK-VSCALE-1-LABEL: 'roundeven'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'roundeven'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'roundeven'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv4f32 = call <vscale x 4 x float> @llvm.roundeven.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %roundeven.nxv2f64 = call <vscale x 2 x double> @llvm.roundeven.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %roundeven.nxv4f64 = call <vscale x 4 x double> @llvm.roundeven.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%roundeven.nxv8f16 = call <vscale x 8 x half> @llvm.roundeven.nxv8f16(<vscale x 8 x half> poison)
@@ -1689,24 +1689,24 @@ define void @roundeven() #1 {
define void @trunc() #1 {
; CHECK-VSCALE-1-LABEL: 'trunc'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'trunc'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'trunc'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv4f32 = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %trunc.nxv2f64 = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> poison)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %trunc.nxv4f64 = call <vscale x 4 x double> @llvm.trunc.nxv4f64(<vscale x 4 x double> poison)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%trunc.nxv8f16 = call <vscale x 8 x half> @llvm.trunc.nxv8f16(<vscale x 8 x half> poison)
@@ -1719,27 +1719,27 @@ define void @trunc() #1 {
define void @abs() #1 {
; CHECK-VSCALE-1-LABEL: 'abs'
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
-; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 4 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 1 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
+; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
; CHECK-VSCALE-1-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; CHECK-VSCALE-2-LABEL: 'abs'
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
-; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 4 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 1 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
+; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
; CHECK-VSCALE-2-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
; TYPE_BASED_ONLY-LABEL: 'abs'
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
-; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 4 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %abs.nxv8i16 = call <vscale x 8 x i16> @llvm.abs.nxv8i16(<vscale x 8 x i16> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %abs.nxv4i32 = call <vscale x 4 x i32> @llvm.abs.nxv4i32(<vscale x 4 x i32> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 1 for: %abs.nxv2i64 = call <vscale x 2 x i64> @llvm.abs.nxv2i64(<vscale x 2 x i64> poison, i1 true)
+; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of 2 for: %abs.nxv4i64 = call <vscale x 4 x i64> @llvm.abs.nxv4i64(<vscale x 4 x i64> poison, i1 true)
; TYPE_BASED_ONLY-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%abs.nxv16i8 = call <vscale x 16 x i8> @llvm.abs.nxv16i8(<vscale x 16 x i8> poison, i1 true)
@@ -1751,6 +1751,7 @@ define void @abs() #1 {
ret void
}
+
declare <vscale x 16 x i1> @llvm.get.active.lane.mask.nxv16i1.i64(i64, i64)
declare <vscale x 8 x i1> @llvm.get.active.lane.mask.nxv8i1.i64(i64, i64)
declare <vscale x 4 x i1> @llvm.get.active.lane.mask.nxv4i1.i64(i64, i64)
diff --git a/llvm/test/Analysis/CostModel/AArch64/sve-math.ll b/llvm/test/Analysis/CostModel/AArch64/sve-math.ll
index c75f266e0983e..6ad60805a99ad 100644
--- a/llvm/test/Analysis/CostModel/AArch64/sve-math.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/sve-math.ll
@@ -16,7 +16,7 @@ define <vscale x 2 x double> @fadd_v2f64(<vscale x 2 x double> %a, <vscale x 2 x
define <vscale x 2 x double> @sqrt_v2f64(<vscale x 2 x double> %a) {
; CHECK-LABEL: 'sqrt_v2f64'
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %r = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> %a)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %r = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> %a)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret <vscale x 2 x double> %r
;
%r = call <vscale x 2 x double> @llvm.sqrt.v2f64(<vscale x 2 x double> %a)
diff --git a/llvm/test/Analysis/CostModel/AArch64/sve-min-max.ll b/llvm/test/Analysis/CostModel/AArch64/sve-min-max.ll
index 5d1113388015b..08f3b577f77e6 100644
--- a/llvm/test/Analysis/CostModel/AArch64/sve-min-max.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/sve-min-max.ll
@@ -154,15 +154,15 @@ define void @smax() {
define void @minnum() {
; CHECK-LABEL: 'minnum'
; CHECK-NEXT: Cost Model: Found costs of Invalid for: %V1f32 = call <vscale x 1 x float> @llvm.minnum.nxv1f32(<vscale x 1 x float> undef, <vscale x 1 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f32 = call <vscale x 2 x float> @llvm.minnum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f32 = call <vscale x 4 x float> @llvm.minnum.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V8f32 = call <vscale x 8 x float> @llvm.minnum.nxv8f32(<vscale x 8 x float> undef, <vscale x 8 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f64 = call <vscale x 2 x double> @llvm.minnum.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x double> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V4f64 = call <vscale x 4 x double> @llvm.minnum.nxv4f64(<vscale x 4 x double> undef, <vscale x 4 x double> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f16 = call <vscale x 2 x half> @llvm.minnum.nxv2f16(<vscale x 2 x half> undef, <vscale x 2 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f16 = call <vscale x 4 x half> @llvm.minnum.nxv4f16(<vscale x 4 x half> undef, <vscale x 4 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V8f16 = call <vscale x 8 x half> @llvm.minnum.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V16f16 = call <vscale x 16 x half> @llvm.minnum.nxv16f16(<vscale x 16 x half> undef, <vscale x 16 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f32 = call <vscale x 2 x float> @llvm.minnum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V4f32 = call <vscale x 4 x float> @llvm.minnum.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V8f32 = call <vscale x 8 x float> @llvm.minnum.nxv8f32(<vscale x 8 x float> undef, <vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f64 = call <vscale x 2 x double> @llvm.minnum.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f64 = call <vscale x 4 x double> @llvm.minnum.nxv4f64(<vscale x 4 x double> undef, <vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f16 = call <vscale x 2 x half> @llvm.minnum.nxv2f16(<vscale x 2 x half> undef, <vscale x 2 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V4f16 = call <vscale x 4 x half> @llvm.minnum.nxv4f16(<vscale x 4 x half> undef, <vscale x 4 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V8f16 = call <vscale x 8 x half> @llvm.minnum.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V16f16 = call <vscale x 16 x half> @llvm.minnum.nxv16f16(<vscale x 16 x half> undef, <vscale x 16 x half> undef)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%V1f32 = call <vscale x 1 x float> @llvm.minnum.nxv1f32(<vscale x 1 x float> undef, <vscale x 1 x float> undef)
@@ -181,15 +181,15 @@ define void @minnum() {
define void @maxnum() {
; CHECK-LABEL: 'maxnum'
; CHECK-NEXT: Cost Model: Found costs of Invalid for: %V1f32 = call <vscale x 1 x float> @llvm.maxnum.nxv1f32(<vscale x 1 x float> undef, <vscale x 1 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f32 = call <vscale x 2 x float> @llvm.maxnum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f32 = call <vscale x 4 x float> @llvm.maxnum.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V8f32 = call <vscale x 8 x float> @llvm.maxnum.nxv8f32(<vscale x 8 x float> undef, <vscale x 8 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f64 = call <vscale x 2 x double> @llvm.maxnum.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x double> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V4f64 = call <vscale x 4 x double> @llvm.maxnum.nxv4f64(<vscale x 4 x double> undef, <vscale x 4 x double> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f16 = call <vscale x 2 x half> @llvm.maxnum.nxv2f16(<vscale x 2 x half> undef, <vscale x 2 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f16 = call <vscale x 4 x half> @llvm.maxnum.nxv4f16(<vscale x 4 x half> undef, <vscale x 4 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V8f16 = call <vscale x 8 x half> @llvm.maxnum.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V16f16 = call <vscale x 16 x half> @llvm.maxnum.nxv16f16(<vscale x 16 x half> undef, <vscale x 16 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f32 = call <vscale x 2 x float> @llvm.maxnum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V4f32 = call <vscale x 4 x float> @llvm.maxnum.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V8f32 = call <vscale x 8 x float> @llvm.maxnum.nxv8f32(<vscale x 8 x float> undef, <vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f64 = call <vscale x 2 x double> @llvm.maxnum.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f64 = call <vscale x 4 x double> @llvm.maxnum.nxv4f64(<vscale x 4 x double> undef, <vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f16 = call <vscale x 2 x half> @llvm.maxnum.nxv2f16(<vscale x 2 x half> undef, <vscale x 2 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V4f16 = call <vscale x 4 x half> @llvm.maxnum.nxv4f16(<vscale x 4 x half> undef, <vscale x 4 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V8f16 = call <vscale x 8 x half> @llvm.maxnum.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V16f16 = call <vscale x 16 x half> @llvm.maxnum.nxv16f16(<vscale x 16 x half> undef, <vscale x 16 x half> undef)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%V1f32 = call <vscale x 1 x float> @llvm.maxnum.nxv1f32(<vscale x 1 x float> undef, <vscale x 1 x float> undef)
@@ -207,15 +207,15 @@ define void @maxnum() {
define void @minimum() {
; CHECK-LABEL: 'minimum'
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f32 = call <vscale x 2 x float> @llvm.minimum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f32 = call <vscale x 4 x float> @llvm.minimum.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V8f32 = call <vscale x 8 x float> @llvm.minimum.nxv8f32(<vscale x 8 x float> undef, <vscale x 8 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f64 = call <vscale x 2 x double> @llvm.minimum.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x double> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V4f64 = call <vscale x 4 x double> @llvm.minimum.nxv4f64(<vscale x 4 x double> undef, <vscale x 4 x double> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f16 = call <vscale x 2 x half> @llvm.minimum.nxv2f16(<vscale x 2 x half> undef, <vscale x 2 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f16 = call <vscale x 4 x half> @llvm.minimum.nxv4f16(<vscale x 4 x half> undef, <vscale x 4 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V8f16 = call <vscale x 8 x half> @llvm.minimum.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V16f16 = call <vscale x 16 x half> @llvm.minimum.nxv16f16(<vscale x 16 x half> undef, <vscale x 16 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f32 = call <vscale x 2 x float> @llvm.minimum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V4f32 = call <vscale x 4 x float> @llvm.minimum.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V8f32 = call <vscale x 8 x float> @llvm.minimum.nxv8f32(<vscale x 8 x float> undef, <vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f64 = call <vscale x 2 x double> @llvm.minimum.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f64 = call <vscale x 4 x double> @llvm.minimum.nxv4f64(<vscale x 4 x double> undef, <vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f16 = call <vscale x 2 x half> @llvm.minimum.nxv2f16(<vscale x 2 x half> undef, <vscale x 2 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V4f16 = call <vscale x 4 x half> @llvm.minimum.nxv4f16(<vscale x 4 x half> undef, <vscale x 4 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V8f16 = call <vscale x 8 x half> @llvm.minimum.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V16f16 = call <vscale x 16 x half> @llvm.minimum.nxv16f16(<vscale x 16 x half> undef, <vscale x 16 x half> undef)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%V2f32 = call <vscale x 2 x float> @llvm.minimum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
@@ -232,15 +232,15 @@ define void @minimum() {
define void @maximum() {
; CHECK-LABEL: 'maximum'
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f32 = call <vscale x 2 x float> @llvm.maximum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f32 = call <vscale x 4 x float> @llvm.maximum.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V8f32 = call <vscale x 8 x float> @llvm.maximum.nxv8f32(<vscale x 8 x float> undef, <vscale x 8 x float> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f64 = call <vscale x 2 x double> @llvm.maximum.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x double> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V4f64 = call <vscale x 4 x double> @llvm.maximum.nxv4f64(<vscale x 4 x double> undef, <vscale x 4 x double> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V2f16 = call <vscale x 2 x half> @llvm.maximum.nxv2f16(<vscale x 2 x half> undef, <vscale x 2 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f16 = call <vscale x 4 x half> @llvm.maximum.nxv4f16(<vscale x 4 x half> undef, <vscale x 4 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %V8f16 = call <vscale x 8 x half> @llvm.maximum.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x half> undef)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %V16f16 = call <vscale x 16 x half> @llvm.maximum.nxv16f16(<vscale x 16 x half> undef, <vscale x 16 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f32 = call <vscale x 2 x float> @llvm.maximum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V4f32 = call <vscale x 4 x float> @llvm.maximum.nxv4f32(<vscale x 4 x float> undef, <vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V8f32 = call <vscale x 8 x float> @llvm.maximum.nxv8f32(<vscale x 8 x float> undef, <vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f64 = call <vscale x 2 x double> @llvm.maximum.nxv2f64(<vscale x 2 x double> undef, <vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V4f64 = call <vscale x 4 x double> @llvm.maximum.nxv4f64(<vscale x 4 x double> undef, <vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V2f16 = call <vscale x 2 x half> @llvm.maximum.nxv2f16(<vscale x 2 x half> undef, <vscale x 2 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V4f16 = call <vscale x 4 x half> @llvm.maximum.nxv4f16(<vscale x 4 x half> undef, <vscale x 4 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %V8f16 = call <vscale x 8 x half> @llvm.maximum.nxv8f16(<vscale x 8 x half> undef, <vscale x 8 x half> undef)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %V16f16 = call <vscale x 16 x half> @llvm.maximum.nxv16f16(<vscale x 16 x half> undef, <vscale x 16 x half> undef)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
%V2f32 = call <vscale x 2 x float> @llvm.maximum.nxv2f32(<vscale x 2 x float> undef, <vscale x 2 x float> undef)
@@ -257,23 +257,23 @@ define void @maximum() {
define void @minimumnum() {
; CHECK-LABEL: 'minimumnum'
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %1 = call <vscale x 2 x float> @llvm.minimumnum.nxv2f32(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %2 = call <vscale x 4 x float> @llvm.minimumnum.nxv4f32(<vscale x 4 x float> poison, <vscale x 4 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %3 = call <vscale x 8 x float> @llvm.minimumnum.nxv8f32(<vscale x 8 x float> poison, <vscale x 8 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %4 = call <vscale x 16 x float> @llvm.minimumnum.nxv16f32(<vscale x 16 x float> poison, <vscale x 16 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %5 = call <vscale x 2 x double> @llvm.minimumnum.nxv2f64(<vscale x 2 x double> poison, <vscale x 2 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %6 = call <vscale x 4 x double> @llvm.minimumnum.nxv4f64(<vscale x 4 x double> poison, <vscale x 4 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %7 = call <vscale x 8 x double> @llvm.minimumnum.nxv8f64(<vscale x 8 x double> poison, <vscale x 8 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %8 = call <vscale x 16 x double> @llvm.minimumnum.nxv16f64(<vscale x 16 x double> poison, <vscale x 16 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %1 = call <vscale x 2 x float> @llvm.minimumnum.nxv2f32(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %2 = call <vscale x 4 x float> @llvm.minimumnum.nxv4f32(<vscale x 4 x float> poison, <vscale x 4 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %3 = call <vscale x 8 x float> @llvm.minimumnum.nxv8f32(<vscale x 8 x float> poison, <vscale x 8 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of 4 for: %4 = call <vscale x 16 x float> @llvm.minimumnum.nxv16f32(<vscale x 16 x float> poison, <vscale x 16 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %5 = call <vscale x 2 x double> @llvm.minimumnum.nxv2f64(<vscale x 2 x double> poison, <vscale x 2 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %6 = call <vscale x 4 x double> @llvm.minimumnum.nxv4f64(<vscale x 4 x double> poison, <vscale x 4 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of 4 for: %7 = call <vscale x 8 x double> @llvm.minimumnum.nxv8f64(<vscale x 8 x double> poison, <vscale x 8 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of 8 for: %8 = call <vscale x 16 x double> @llvm.minimumnum.nxv16f64(<vscale x 16 x double> poison, <vscale x 16 x double> poison)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %10 = call <vscale x 2 x half> @llvm.minimumnum.nxv2f16(<vscale x 2 x half> poison, <vscale x 2 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %11 = call <vscale x 4 x half> @llvm.minimumnum.nxv4f16(<vscale x 4 x half> poison, <vscale x 4 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %12 = call <vscale x 8 x half> @llvm.minimumnum.nxv8f16(<vscale x 8 x half> poison, <vscale x 8 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %13 = call <vscale x 16 x half> @llvm.minimumnum.nxv16f16(<vscale x 16 x half> poison, <vscale x 16 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %14 = call <vscale x 2 x bfloat> @llvm.minimumnum.nxv2bf16(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %15 = call <vscale x 4 x bfloat> @llvm.minimumnum.nxv4bf16(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %16 = call <vscale x 8 x bfloat> @llvm.minimumnum.nxv8bf16(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %17 = call <vscale x 16 x bfloat> @llvm.minimumnum.nxv16bf16(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %10 = call <vscale x 2 x half> @llvm.minimumnum.nxv2f16(<vscale x 2 x half> poison, <vscale x 2 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %11 = call <vscale x 4 x half> @llvm.minimumnum.nxv4f16(<vscale x 4 x half> poison, <vscale x 4 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %12 = call <vscale x 8 x half> @llvm.minimumnum.nxv8f16(<vscale x 8 x half> poison, <vscale x 8 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %13 = call <vscale x 16 x half> @llvm.minimumnum.nxv16f16(<vscale x 16 x half> poison, <vscale x 16 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %14 = call <vscale x 2 x bfloat> @llvm.minimumnum.nxv2bf16(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %15 = call <vscale x 4 x bfloat> @llvm.minimumnum.nxv4bf16(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %16 = call <vscale x 8 x bfloat> @llvm.minimumnum.nxv8bf16(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %17 = call <vscale x 16 x bfloat> @llvm.minimumnum.nxv16bf16(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> poison)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
call <vscale x 2 x float> @llvm.minimumnum(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
@@ -298,23 +298,23 @@ define void @minimumnum() {
define void @maximumnum() {
; CHECK-LABEL: 'maximumnum'
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %1 = call <vscale x 2 x float> @llvm.maximumnum.nxv2f32(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %2 = call <vscale x 4 x float> @llvm.maximumnum.nxv4f32(<vscale x 4 x float> poison, <vscale x 4 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %3 = call <vscale x 8 x float> @llvm.maximumnum.nxv8f32(<vscale x 8 x float> poison, <vscale x 8 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %4 = call <vscale x 16 x float> @llvm.maximumnum.nxv16f32(<vscale x 16 x float> poison, <vscale x 16 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %5 = call <vscale x 2 x double> @llvm.maximumnum.nxv2f64(<vscale x 2 x double> poison, <vscale x 2 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %6 = call <vscale x 4 x double> @llvm.maximumnum.nxv4f64(<vscale x 4 x double> poison, <vscale x 4 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %7 = call <vscale x 8 x double> @llvm.maximumnum.nxv8f64(<vscale x 8 x double> poison, <vscale x 8 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %8 = call <vscale x 16 x double> @llvm.maximumnum.nxv16f64(<vscale x 16 x double> poison, <vscale x 16 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %1 = call <vscale x 2 x float> @llvm.maximumnum.nxv2f32(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %2 = call <vscale x 4 x float> @llvm.maximumnum.nxv4f32(<vscale x 4 x float> poison, <vscale x 4 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %3 = call <vscale x 8 x float> @llvm.maximumnum.nxv8f32(<vscale x 8 x float> poison, <vscale x 8 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of 4 for: %4 = call <vscale x 16 x float> @llvm.maximumnum.nxv16f32(<vscale x 16 x float> poison, <vscale x 16 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %5 = call <vscale x 2 x double> @llvm.maximumnum.nxv2f64(<vscale x 2 x double> poison, <vscale x 2 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %6 = call <vscale x 4 x double> @llvm.maximumnum.nxv4f64(<vscale x 4 x double> poison, <vscale x 4 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of 4 for: %7 = call <vscale x 8 x double> @llvm.maximumnum.nxv8f64(<vscale x 8 x double> poison, <vscale x 8 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of 8 for: %8 = call <vscale x 16 x double> @llvm.maximumnum.nxv16f64(<vscale x 16 x double> poison, <vscale x 16 x double> poison)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %10 = call <vscale x 2 x half> @llvm.maximumnum.nxv2f16(<vscale x 2 x half> poison, <vscale x 2 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %11 = call <vscale x 4 x half> @llvm.maximumnum.nxv4f16(<vscale x 4 x half> poison, <vscale x 4 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %12 = call <vscale x 8 x half> @llvm.maximumnum.nxv8f16(<vscale x 8 x half> poison, <vscale x 8 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %13 = call <vscale x 16 x half> @llvm.maximumnum.nxv16f16(<vscale x 16 x half> poison, <vscale x 16 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %14 = call <vscale x 2 x bfloat> @llvm.maximumnum.nxv2bf16(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %15 = call <vscale x 4 x bfloat> @llvm.maximumnum.nxv4bf16(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %16 = call <vscale x 8 x bfloat> @llvm.maximumnum.nxv8bf16(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of Invalid for: %17 = call <vscale x 16 x bfloat> @llvm.maximumnum.nxv16bf16(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %10 = call <vscale x 2 x half> @llvm.maximumnum.nxv2f16(<vscale x 2 x half> poison, <vscale x 2 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %11 = call <vscale x 4 x half> @llvm.maximumnum.nxv4f16(<vscale x 4 x half> poison, <vscale x 4 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %12 = call <vscale x 8 x half> @llvm.maximumnum.nxv8f16(<vscale x 8 x half> poison, <vscale x 8 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %13 = call <vscale x 16 x half> @llvm.maximumnum.nxv16f16(<vscale x 16 x half> poison, <vscale x 16 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %14 = call <vscale x 2 x bfloat> @llvm.maximumnum.nxv2bf16(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %15 = call <vscale x 4 x bfloat> @llvm.maximumnum.nxv4bf16(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %16 = call <vscale x 8 x bfloat> @llvm.maximumnum.nxv8bf16(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of 2 for: %17 = call <vscale x 16 x bfloat> @llvm.maximumnum.nxv16bf16(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> poison)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
call <vscale x 2 x float> @llvm.maximumnum(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/pr151664-cost-hoisted-vector-scalable.ll b/llvm/test/Transforms/LoopVectorize/AArch64/pr151664-cost-hoisted-vector-scalable.ll
index 73dbefeb10413..b634d5150a4f1 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/pr151664-cost-hoisted-vector-scalable.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/pr151664-cost-hoisted-vector-scalable.ll
@@ -7,22 +7,29 @@ define void @cost_hoisted_vector_code(ptr %p, float %arg) {
; CHECK-NEXT: [[ENTRY:.*:]]
; CHECK-NEXT: br label %[[VECTOR_PH:.*]]
; CHECK: [[VECTOR_PH]]:
-; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x float> poison, float [[ARG]], i64 0
-; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x float> [[BROADCAST_SPLATINSERT]], <4 x float> poison, <4 x i32> zeroinitializer
-; CHECK-NEXT: [[TMP0:%.*]] = call <4 x float> @llvm.minimumnum.v4f32(<4 x float> [[BROADCAST_SPLAT]], <4 x float> zeroinitializer)
+; CHECK-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
+; CHECK-NEXT: [[TMP5:%.*]] = shl nuw i64 [[TMP0]], 2
+; CHECK-NEXT: [[TMP2:%.*]] = shl nuw i64 [[TMP5]], 1
+; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 -1, [[TMP2]]
+; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 -1, [[N_MOD_VF]]
+; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 4 x float> poison, float [[ARG]], i64 0
+; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 4 x float> [[BROADCAST_SPLATINSERT]], <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
+; CHECK-NEXT: [[TMP3:%.*]] = add i64 1, [[N_VEC]]
+; CHECK-NEXT: [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.minimumnum.nxv4f32(<vscale x 4 x float> [[BROADCAST_SPLAT]], <vscale x 4 x float> zeroinitializer)
; CHECK-NEXT: br label %[[VECTOR_BODY:.*]]
; CHECK: [[VECTOR_BODY]]:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT: [[OFFSET_IDX:%.*]] = add i64 1, [[INDEX]]
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr float, ptr [[P]], i64 [[OFFSET_IDX]]
-; CHECK-NEXT: [[TMP2:%.*]] = getelementptr float, ptr [[TMP1]], i64 4
-; CHECK-NEXT: store <4 x float> [[TMP0]], ptr [[TMP1]], align 4
-; CHECK-NEXT: store <4 x float> [[TMP0]], ptr [[TMP2]], align 4
-; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
-; CHECK-NEXT: [[TMP3:%.*]] = icmp eq i64 [[INDEX_NEXT]], -8
-; CHECK-NEXT: br i1 [[TMP3]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT: [[TMP6:%.*]] = getelementptr float, ptr [[TMP1]], i64 [[TMP5]]
+; CHECK-NEXT: store <vscale x 4 x float> [[TMP4]], ptr [[TMP1]], align 4
+; CHECK-NEXT: store <vscale x 4 x float> [[TMP4]], ptr [[TMP6]], align 4
+; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP2]]
+; CHECK-NEXT: [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
+; CHECK-NEXT: br i1 [[TMP7]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
-; CHECK-NEXT: br label %[[SCALAR_PH:.*]]
+; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 -1, [[N_VEC]]
+; CHECK-NEXT: br i1 [[CMP_N]], [[EXIT:label %.*]], label %[[SCALAR_PH:.*]]
; CHECK: [[SCALAR_PH]]:
;
entry:
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/veclib-intrinsic-calls.ll b/llvm/test/Transforms/LoopVectorize/AArch64/veclib-intrinsic-calls.ll
index abf4dead6fcd4..f8cbde019eebe 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/veclib-intrinsic-calls.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/veclib-intrinsic-calls.ll
@@ -425,7 +425,8 @@ define void @ceil_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @ceil_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.ceil.v2f64(<2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.ceil.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.ceil.f64(double [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @ceil_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -472,7 +473,8 @@ define void @ceil_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @ceil_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.ceil.v4f32(<4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.ceil.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.ceil.f32(float [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @ceil_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1116,7 +1118,8 @@ define void @fabs_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @fabs_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.fabs.v2f64(<2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.fabs.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.fabs.f64(double [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @fabs_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1163,7 +1166,8 @@ define void @fabs_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @fabs_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.fabs.v4f32(<4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.fabs.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.fabs.f32(float [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @fabs_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1213,7 +1217,8 @@ define void @floor_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @floor_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.floor.v2f64(<2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.floor.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.floor.f64(double [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @floor_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1260,7 +1265,8 @@ define void @floor_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @floor_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.floor.v4f32(<4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.floor.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.floor.f32(float [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @floor_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1706,7 +1712,8 @@ define void @maxnum_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @maxnum_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.maxnum.v2f64(<2 x double> [[WIDE_LOAD:%.*]], <2 x double> [[WIDE_LOAD]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.maxnum.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]], <vscale x 2 x double> [[WIDE_LOAD]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.maxnum.f64(double [[IN:%.*]], double [[IN]])
;
; SLEEF-NEON-LABEL: define void @maxnum_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1753,7 +1760,8 @@ define void @maxnum_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @maxnum_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.maxnum.v4f32(<4 x float> [[WIDE_LOAD:%.*]], <4 x float> [[WIDE_LOAD]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.maxnum.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]], <vscale x 4 x float> [[WIDE_LOAD]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.maxnum.f32(float [[IN:%.*]], float [[IN]])
;
; SLEEF-NEON-LABEL: define void @maxnum_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1803,7 +1811,8 @@ define void @minnum_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @minnum_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.minnum.v2f64(<2 x double> [[WIDE_LOAD:%.*]], <2 x double> [[WIDE_LOAD]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.minnum.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]], <vscale x 2 x double> [[WIDE_LOAD]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.minnum.f64(double [[IN:%.*]], double [[IN]])
;
; SLEEF-NEON-LABEL: define void @minnum_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -1850,7 +1859,8 @@ define void @minnum_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @minnum_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.minnum.v4f32(<4 x float> [[WIDE_LOAD:%.*]], <4 x float> [[WIDE_LOAD]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.minnum.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]], <vscale x 4 x float> [[WIDE_LOAD]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.minnum.f32(float [[IN:%.*]], float [[IN]])
;
; SLEEF-NEON-LABEL: define void @minnum_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2019,7 +2029,8 @@ define void @nearbyint_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @nearbyint_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.nearbyint.v2f64(<2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.nearbyint.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.nearbyint.f64(double [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @nearbyint_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2066,7 +2077,8 @@ define void @nearbyint_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @nearbyint_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.nearbyint.v4f32(<4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.nearbyint.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.nearbyint.f32(float [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @nearbyint_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2215,7 +2227,8 @@ define void @rint_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @rint_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.rint.v2f64(<2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.rint.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.rint.f64(double [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @rint_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2262,7 +2275,8 @@ define void @rint_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @rint_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.rint.v4f32(<4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.rint.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.rint.f32(float [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @rint_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2312,7 +2326,8 @@ define void @round_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @round_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.round.v2f64(<2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.round.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.round.f64(double [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @round_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2359,7 +2374,8 @@ define void @round_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @round_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.round.v4f32(<4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.round.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.round.f32(float [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @round_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2845,7 +2861,8 @@ define void @sqrt_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @sqrt_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.sqrt.v2f64(<2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.sqrt.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.sqrt.f64(double [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @sqrt_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -2892,7 +2909,8 @@ define void @sqrt_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @sqrt_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.sqrt.v4f32(<4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.sqrt.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.sqrt.f32(float [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @sqrt_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -3140,7 +3158,8 @@ define void @trunc_f64(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @trunc_f64
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <2 x double> @llvm.trunc.v2f64(<2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 2 x double> @llvm.trunc.nxv2f64(<vscale x 2 x double> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call double @llvm.trunc.f64(double [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @trunc_f64
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
@@ -3187,7 +3206,8 @@ define void @trunc_f32(ptr noalias %in.ptr, ptr %out.ptr) {
;
; LIBMVEC-SVE-LABEL: define void @trunc_f32
; LIBMVEC-SVE-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
-; LIBMVEC-SVE: [[TMP1:%.*]] = call <4 x float> @llvm.trunc.v4f32(<4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[TMP5:%.*]] = call <vscale x 4 x float> @llvm.trunc.nxv4f32(<vscale x 4 x float> [[WIDE_LOAD:%.*]])
+; LIBMVEC-SVE: [[CALL:%.*]] = tail call float @llvm.trunc.f32(float [[IN:%.*]])
;
; SLEEF-NEON-LABEL: define void @trunc_f32
; SLEEF-NEON-SAME: (ptr noalias [[IN_PTR:%.*]], ptr [[OUT_PTR:%.*]]) #[[ATTR1]] {
>From e6bf96ff846a4afac17a87b25ce1e50dda42eb93 Mon Sep 17 00:00:00 2001
From: David Sherwood <david.sherwood at arm.com>
Date: Wed, 4 Mar 2026 10:42:48 +0000
Subject: [PATCH 3/3] Remove maximumnum/minimumnum
---
.../AArch64/AArch64TargetTransformInfo.cpp | 4 +-
.../Analysis/CostModel/AArch64/sve-min-max.ll | 64 +++++++++----------
.../pr151664-cost-hoisted-vector-scalable.ll | 27 +++-----
3 files changed, 43 insertions(+), 52 deletions(-)
diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
index 7d24986b7981a..5a98fb950ac76 100644
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -1133,9 +1133,7 @@ AArch64TTIImpl::getIntrinsicInstrCost(const IntrinsicCostAttributes &ICA,
case Intrinsic::minnum:
case Intrinsic::maxnum:
case Intrinsic::minimum:
- case Intrinsic::maximum:
- case Intrinsic::minimumnum:
- case Intrinsic::maximumnum: {
+ case Intrinsic::maximum: {
if (isa<ScalableVectorType>(RetTy) && ST->isSVEorStreamingSVEAvailable()) {
auto LT = getTypeLegalizationCost(RetTy);
return LT.first;
diff --git a/llvm/test/Analysis/CostModel/AArch64/sve-min-max.ll b/llvm/test/Analysis/CostModel/AArch64/sve-min-max.ll
index 08f3b577f77e6..e84e476e5d824 100644
--- a/llvm/test/Analysis/CostModel/AArch64/sve-min-max.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/sve-min-max.ll
@@ -257,23 +257,23 @@ define void @maximum() {
define void @minimumnum() {
; CHECK-LABEL: 'minimumnum'
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %1 = call <vscale x 2 x float> @llvm.minimumnum.nxv2f32(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %2 = call <vscale x 4 x float> @llvm.minimumnum.nxv4f32(<vscale x 4 x float> poison, <vscale x 4 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %3 = call <vscale x 8 x float> @llvm.minimumnum.nxv8f32(<vscale x 8 x float> poison, <vscale x 8 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %4 = call <vscale x 16 x float> @llvm.minimumnum.nxv16f32(<vscale x 16 x float> poison, <vscale x 16 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %5 = call <vscale x 2 x double> @llvm.minimumnum.nxv2f64(<vscale x 2 x double> poison, <vscale x 2 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %6 = call <vscale x 4 x double> @llvm.minimumnum.nxv4f64(<vscale x 4 x double> poison, <vscale x 4 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %7 = call <vscale x 8 x double> @llvm.minimumnum.nxv8f64(<vscale x 8 x double> poison, <vscale x 8 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of 8 for: %8 = call <vscale x 16 x double> @llvm.minimumnum.nxv16f64(<vscale x 16 x double> poison, <vscale x 16 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %1 = call <vscale x 2 x float> @llvm.minimumnum.nxv2f32(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %2 = call <vscale x 4 x float> @llvm.minimumnum.nxv4f32(<vscale x 4 x float> poison, <vscale x 4 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %3 = call <vscale x 8 x float> @llvm.minimumnum.nxv8f32(<vscale x 8 x float> poison, <vscale x 8 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %4 = call <vscale x 16 x float> @llvm.minimumnum.nxv16f32(<vscale x 16 x float> poison, <vscale x 16 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %5 = call <vscale x 2 x double> @llvm.minimumnum.nxv2f64(<vscale x 2 x double> poison, <vscale x 2 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %6 = call <vscale x 4 x double> @llvm.minimumnum.nxv4f64(<vscale x 4 x double> poison, <vscale x 4 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %7 = call <vscale x 8 x double> @llvm.minimumnum.nxv8f64(<vscale x 8 x double> poison, <vscale x 8 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %8 = call <vscale x 16 x double> @llvm.minimumnum.nxv16f64(<vscale x 16 x double> poison, <vscale x 16 x double> poison)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %10 = call <vscale x 2 x half> @llvm.minimumnum.nxv2f16(<vscale x 2 x half> poison, <vscale x 2 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %11 = call <vscale x 4 x half> @llvm.minimumnum.nxv4f16(<vscale x 4 x half> poison, <vscale x 4 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %12 = call <vscale x 8 x half> @llvm.minimumnum.nxv8f16(<vscale x 8 x half> poison, <vscale x 8 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %13 = call <vscale x 16 x half> @llvm.minimumnum.nxv16f16(<vscale x 16 x half> poison, <vscale x 16 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %14 = call <vscale x 2 x bfloat> @llvm.minimumnum.nxv2bf16(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %15 = call <vscale x 4 x bfloat> @llvm.minimumnum.nxv4bf16(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %16 = call <vscale x 8 x bfloat> @llvm.minimumnum.nxv8bf16(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %17 = call <vscale x 16 x bfloat> @llvm.minimumnum.nxv16bf16(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %10 = call <vscale x 2 x half> @llvm.minimumnum.nxv2f16(<vscale x 2 x half> poison, <vscale x 2 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %11 = call <vscale x 4 x half> @llvm.minimumnum.nxv4f16(<vscale x 4 x half> poison, <vscale x 4 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %12 = call <vscale x 8 x half> @llvm.minimumnum.nxv8f16(<vscale x 8 x half> poison, <vscale x 8 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %13 = call <vscale x 16 x half> @llvm.minimumnum.nxv16f16(<vscale x 16 x half> poison, <vscale x 16 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %14 = call <vscale x 2 x bfloat> @llvm.minimumnum.nxv2bf16(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %15 = call <vscale x 4 x bfloat> @llvm.minimumnum.nxv4bf16(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %16 = call <vscale x 8 x bfloat> @llvm.minimumnum.nxv8bf16(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %17 = call <vscale x 16 x bfloat> @llvm.minimumnum.nxv16bf16(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> poison)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
call <vscale x 2 x float> @llvm.minimumnum(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
@@ -298,23 +298,23 @@ define void @minimumnum() {
define void @maximumnum() {
; CHECK-LABEL: 'maximumnum'
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %1 = call <vscale x 2 x float> @llvm.maximumnum.nxv2f32(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %2 = call <vscale x 4 x float> @llvm.maximumnum.nxv4f32(<vscale x 4 x float> poison, <vscale x 4 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %3 = call <vscale x 8 x float> @llvm.maximumnum.nxv8f32(<vscale x 8 x float> poison, <vscale x 8 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %4 = call <vscale x 16 x float> @llvm.maximumnum.nxv16f32(<vscale x 16 x float> poison, <vscale x 16 x float> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %5 = call <vscale x 2 x double> @llvm.maximumnum.nxv2f64(<vscale x 2 x double> poison, <vscale x 2 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %6 = call <vscale x 4 x double> @llvm.maximumnum.nxv4f64(<vscale x 4 x double> poison, <vscale x 4 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of 4 for: %7 = call <vscale x 8 x double> @llvm.maximumnum.nxv8f64(<vscale x 8 x double> poison, <vscale x 8 x double> poison)
-; CHECK-NEXT: Cost Model: Found costs of 8 for: %8 = call <vscale x 16 x double> @llvm.maximumnum.nxv16f64(<vscale x 16 x double> poison, <vscale x 16 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %1 = call <vscale x 2 x float> @llvm.maximumnum.nxv2f32(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %2 = call <vscale x 4 x float> @llvm.maximumnum.nxv4f32(<vscale x 4 x float> poison, <vscale x 4 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %3 = call <vscale x 8 x float> @llvm.maximumnum.nxv8f32(<vscale x 8 x float> poison, <vscale x 8 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %4 = call <vscale x 16 x float> @llvm.maximumnum.nxv16f32(<vscale x 16 x float> poison, <vscale x 16 x float> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %5 = call <vscale x 2 x double> @llvm.maximumnum.nxv2f64(<vscale x 2 x double> poison, <vscale x 2 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %6 = call <vscale x 4 x double> @llvm.maximumnum.nxv4f64(<vscale x 4 x double> poison, <vscale x 4 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %7 = call <vscale x 8 x double> @llvm.maximumnum.nxv8f64(<vscale x 8 x double> poison, <vscale x 8 x double> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %8 = call <vscale x 16 x double> @llvm.maximumnum.nxv16f64(<vscale x 16 x double> poison, <vscale x 16 x double> poison)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %10 = call <vscale x 2 x half> @llvm.maximumnum.nxv2f16(<vscale x 2 x half> poison, <vscale x 2 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %11 = call <vscale x 4 x half> @llvm.maximumnum.nxv4f16(<vscale x 4 x half> poison, <vscale x 4 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %12 = call <vscale x 8 x half> @llvm.maximumnum.nxv8f16(<vscale x 8 x half> poison, <vscale x 8 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %13 = call <vscale x 16 x half> @llvm.maximumnum.nxv16f16(<vscale x 16 x half> poison, <vscale x 16 x half> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %14 = call <vscale x 2 x bfloat> @llvm.maximumnum.nxv2bf16(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %15 = call <vscale x 4 x bfloat> @llvm.maximumnum.nxv4bf16(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of 1 for: %16 = call <vscale x 8 x bfloat> @llvm.maximumnum.nxv8bf16(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> poison)
-; CHECK-NEXT: Cost Model: Found costs of 2 for: %17 = call <vscale x 16 x bfloat> @llvm.maximumnum.nxv16bf16(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %10 = call <vscale x 2 x half> @llvm.maximumnum.nxv2f16(<vscale x 2 x half> poison, <vscale x 2 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %11 = call <vscale x 4 x half> @llvm.maximumnum.nxv4f16(<vscale x 4 x half> poison, <vscale x 4 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %12 = call <vscale x 8 x half> @llvm.maximumnum.nxv8f16(<vscale x 8 x half> poison, <vscale x 8 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %13 = call <vscale x 16 x half> @llvm.maximumnum.nxv16f16(<vscale x 16 x half> poison, <vscale x 16 x half> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %14 = call <vscale x 2 x bfloat> @llvm.maximumnum.nxv2bf16(<vscale x 2 x bfloat> poison, <vscale x 2 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %15 = call <vscale x 4 x bfloat> @llvm.maximumnum.nxv4bf16(<vscale x 4 x bfloat> poison, <vscale x 4 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %16 = call <vscale x 8 x bfloat> @llvm.maximumnum.nxv8bf16(<vscale x 8 x bfloat> poison, <vscale x 8 x bfloat> poison)
+; CHECK-NEXT: Cost Model: Found costs of Invalid for: %17 = call <vscale x 16 x bfloat> @llvm.maximumnum.nxv16bf16(<vscale x 16 x bfloat> poison, <vscale x 16 x bfloat> poison)
; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
;
call <vscale x 2 x float> @llvm.maximumnum(<vscale x 2 x float> poison, <vscale x 2 x float> poison)
diff --git a/llvm/test/Transforms/LoopVectorize/AArch64/pr151664-cost-hoisted-vector-scalable.ll b/llvm/test/Transforms/LoopVectorize/AArch64/pr151664-cost-hoisted-vector-scalable.ll
index b634d5150a4f1..73dbefeb10413 100644
--- a/llvm/test/Transforms/LoopVectorize/AArch64/pr151664-cost-hoisted-vector-scalable.ll
+++ b/llvm/test/Transforms/LoopVectorize/AArch64/pr151664-cost-hoisted-vector-scalable.ll
@@ -7,29 +7,22 @@ define void @cost_hoisted_vector_code(ptr %p, float %arg) {
; CHECK-NEXT: [[ENTRY:.*:]]
; CHECK-NEXT: br label %[[VECTOR_PH:.*]]
; CHECK: [[VECTOR_PH]]:
-; CHECK-NEXT: [[TMP0:%.*]] = call i64 @llvm.vscale.i64()
-; CHECK-NEXT: [[TMP5:%.*]] = shl nuw i64 [[TMP0]], 2
-; CHECK-NEXT: [[TMP2:%.*]] = shl nuw i64 [[TMP5]], 1
-; CHECK-NEXT: [[N_MOD_VF:%.*]] = urem i64 -1, [[TMP2]]
-; CHECK-NEXT: [[N_VEC:%.*]] = sub i64 -1, [[N_MOD_VF]]
-; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <vscale x 4 x float> poison, float [[ARG]], i64 0
-; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <vscale x 4 x float> [[BROADCAST_SPLATINSERT]], <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
-; CHECK-NEXT: [[TMP3:%.*]] = add i64 1, [[N_VEC]]
-; CHECK-NEXT: [[TMP4:%.*]] = call <vscale x 4 x float> @llvm.minimumnum.nxv4f32(<vscale x 4 x float> [[BROADCAST_SPLAT]], <vscale x 4 x float> zeroinitializer)
+; CHECK-NEXT: [[BROADCAST_SPLATINSERT:%.*]] = insertelement <4 x float> poison, float [[ARG]], i64 0
+; CHECK-NEXT: [[BROADCAST_SPLAT:%.*]] = shufflevector <4 x float> [[BROADCAST_SPLATINSERT]], <4 x float> poison, <4 x i32> zeroinitializer
+; CHECK-NEXT: [[TMP0:%.*]] = call <4 x float> @llvm.minimumnum.v4f32(<4 x float> [[BROADCAST_SPLAT]], <4 x float> zeroinitializer)
; CHECK-NEXT: br label %[[VECTOR_BODY:.*]]
; CHECK: [[VECTOR_BODY]]:
; CHECK-NEXT: [[INDEX:%.*]] = phi i64 [ 0, %[[VECTOR_PH]] ], [ [[INDEX_NEXT:%.*]], %[[VECTOR_BODY]] ]
; CHECK-NEXT: [[OFFSET_IDX:%.*]] = add i64 1, [[INDEX]]
; CHECK-NEXT: [[TMP1:%.*]] = getelementptr float, ptr [[P]], i64 [[OFFSET_IDX]]
-; CHECK-NEXT: [[TMP6:%.*]] = getelementptr float, ptr [[TMP1]], i64 [[TMP5]]
-; CHECK-NEXT: store <vscale x 4 x float> [[TMP4]], ptr [[TMP1]], align 4
-; CHECK-NEXT: store <vscale x 4 x float> [[TMP4]], ptr [[TMP6]], align 4
-; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], [[TMP2]]
-; CHECK-NEXT: [[TMP7:%.*]] = icmp eq i64 [[INDEX_NEXT]], [[N_VEC]]
-; CHECK-NEXT: br i1 [[TMP7]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
+; CHECK-NEXT: [[TMP2:%.*]] = getelementptr float, ptr [[TMP1]], i64 4
+; CHECK-NEXT: store <4 x float> [[TMP0]], ptr [[TMP1]], align 4
+; CHECK-NEXT: store <4 x float> [[TMP0]], ptr [[TMP2]], align 4
+; CHECK-NEXT: [[INDEX_NEXT]] = add nuw i64 [[INDEX]], 8
+; CHECK-NEXT: [[TMP3:%.*]] = icmp eq i64 [[INDEX_NEXT]], -8
+; CHECK-NEXT: br i1 [[TMP3]], label %[[MIDDLE_BLOCK:.*]], label %[[VECTOR_BODY]], !llvm.loop [[LOOP0:![0-9]+]]
; CHECK: [[MIDDLE_BLOCK]]:
-; CHECK-NEXT: [[CMP_N:%.*]] = icmp eq i64 -1, [[N_VEC]]
-; CHECK-NEXT: br i1 [[CMP_N]], [[EXIT:label %.*]], label %[[SCALAR_PH:.*]]
+; CHECK-NEXT: br label %[[SCALAR_PH:.*]]
; CHECK: [[SCALAR_PH]]:
;
entry:
More information about the llvm-commits
mailing list