[llvm] [AArch64] Add additional cost coverage for SVE fptosi.sat and fptoui.sat. NFC (PR #192095)
David Green via llvm-commits
llvm-commits at lists.llvm.org
Tue Apr 14 10:58:33 PDT 2026
https://github.com/davemgreen created https://github.com/llvm/llvm-project/pull/192095
None
>From f292ff0463e80d5457042ca4a125f7352d7493f7 Mon Sep 17 00:00:00 2001
From: David Green <david.green at arm.com>
Date: Tue, 14 Apr 2026 18:58:20 +0100
Subject: [PATCH] [AArch64] Add additional cost coverage for SVE fptosi.sat and
fptoui.sat. NFC
---
.../Analysis/CostModel/AArch64/fptoi_sat.ll | 166 ----------------
.../CostModel/AArch64/sve-fptoi_sat.ll | 179 ++++++++++++++++++
2 files changed, 179 insertions(+), 166 deletions(-)
create mode 100644 llvm/test/Analysis/CostModel/AArch64/sve-fptoi_sat.ll
diff --git a/llvm/test/Analysis/CostModel/AArch64/fptoi_sat.ll b/llvm/test/Analysis/CostModel/AArch64/fptoi_sat.ll
index b614c82fc2453..4d9c0c59224b4 100644
--- a/llvm/test/Analysis/CostModel/AArch64/fptoi_sat.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/fptoi_sat.ll
@@ -385,169 +385,3 @@ define void @fp16() {
ret void
}
-
-
-declare i1 @llvm.fptosi.sat.i1.f32(float)
-declare i1 @llvm.fptoui.sat.i1.f32(float)
-declare i8 @llvm.fptosi.sat.i8.f32(float)
-declare i8 @llvm.fptoui.sat.i8.f32(float)
-declare i16 @llvm.fptosi.sat.i16.f32(float)
-declare i16 @llvm.fptoui.sat.i16.f32(float)
-declare i32 @llvm.fptosi.sat.i32.f32(float)
-declare i32 @llvm.fptoui.sat.i32.f32(float)
-declare i64 @llvm.fptosi.sat.i64.f32(float)
-declare i64 @llvm.fptoui.sat.i64.f32(float)
-
-declare i1 @llvm.fptosi.sat.i1.f64(double)
-declare i1 @llvm.fptoui.sat.i1.f64(double)
-declare i8 @llvm.fptosi.sat.i8.f64(double)
-declare i8 @llvm.fptoui.sat.i8.f64(double)
-declare i16 @llvm.fptosi.sat.i16.f64(double)
-declare i16 @llvm.fptoui.sat.i16.f64(double)
-declare i32 @llvm.fptosi.sat.i32.f64(double)
-declare i32 @llvm.fptoui.sat.i32.f64(double)
-declare i64 @llvm.fptosi.sat.i64.f64(double)
-declare i64 @llvm.fptoui.sat.i64.f64(double)
-
-declare <2 x i1> @llvm.fptosi.sat.v2i1.v2f32(<2 x float>)
-declare <2 x i1> @llvm.fptoui.sat.v2i1.v2f32(<2 x float>)
-declare <2 x i8> @llvm.fptosi.sat.v2i8.v2f32(<2 x float>)
-declare <2 x i8> @llvm.fptoui.sat.v2i8.v2f32(<2 x float>)
-declare <2 x i16> @llvm.fptosi.sat.v2i16.v2f32(<2 x float>)
-declare <2 x i16> @llvm.fptoui.sat.v2i16.v2f32(<2 x float>)
-declare <2 x i32> @llvm.fptosi.sat.v2i32.v2f32(<2 x float>)
-declare <2 x i32> @llvm.fptoui.sat.v2i32.v2f32(<2 x float>)
-declare <2 x i64> @llvm.fptosi.sat.v2i64.v2f32(<2 x float>)
-declare <2 x i64> @llvm.fptoui.sat.v2i64.v2f32(<2 x float>)
-
-declare <2 x i1> @llvm.fptosi.sat.v2i1.v2f64(<2 x double>)
-declare <2 x i1> @llvm.fptoui.sat.v2i1.v2f64(<2 x double>)
-declare <2 x i8> @llvm.fptosi.sat.v2i8.v2f64(<2 x double>)
-declare <2 x i8> @llvm.fptoui.sat.v2i8.v2f64(<2 x double>)
-declare <2 x i16> @llvm.fptosi.sat.v2i16.v2f64(<2 x double>)
-declare <2 x i16> @llvm.fptoui.sat.v2i16.v2f64(<2 x double>)
-declare <2 x i32> @llvm.fptosi.sat.v2i32.v2f64(<2 x double>)
-declare <2 x i32> @llvm.fptoui.sat.v2i32.v2f64(<2 x double>)
-declare <2 x i64> @llvm.fptosi.sat.v2i64.v2f64(<2 x double>)
-declare <2 x i64> @llvm.fptoui.sat.v2i64.v2f64(<2 x double>)
-
-declare <4 x i1> @llvm.fptosi.sat.v4i1.v4f32(<4 x float>)
-declare <4 x i1> @llvm.fptoui.sat.v4i1.v4f32(<4 x float>)
-declare <4 x i8> @llvm.fptosi.sat.v4i8.v4f32(<4 x float>)
-declare <4 x i8> @llvm.fptoui.sat.v4i8.v4f32(<4 x float>)
-declare <4 x i16> @llvm.fptosi.sat.v4i16.v4f32(<4 x float>)
-declare <4 x i16> @llvm.fptoui.sat.v4i16.v4f32(<4 x float>)
-declare <4 x i32> @llvm.fptosi.sat.v4i32.v4f32(<4 x float>)
-declare <4 x i32> @llvm.fptoui.sat.v4i32.v4f32(<4 x float>)
-declare <4 x i64> @llvm.fptosi.sat.v4i64.v4f32(<4 x float>)
-declare <4 x i64> @llvm.fptoui.sat.v4i64.v4f32(<4 x float>)
-
-declare <4 x i1> @llvm.fptosi.sat.v4i1.v4f64(<4 x double>)
-declare <4 x i1> @llvm.fptoui.sat.v4i1.v4f64(<4 x double>)
-declare <4 x i8> @llvm.fptosi.sat.v4i8.v4f64(<4 x double>)
-declare <4 x i8> @llvm.fptoui.sat.v4i8.v4f64(<4 x double>)
-declare <4 x i16> @llvm.fptosi.sat.v4i16.v4f64(<4 x double>)
-declare <4 x i16> @llvm.fptoui.sat.v4i16.v4f64(<4 x double>)
-declare <4 x i32> @llvm.fptosi.sat.v4i32.v4f64(<4 x double>)
-declare <4 x i32> @llvm.fptoui.sat.v4i32.v4f64(<4 x double>)
-declare <4 x i64> @llvm.fptosi.sat.v4i64.v4f64(<4 x double>)
-declare <4 x i64> @llvm.fptoui.sat.v4i64.v4f64(<4 x double>)
-
-declare <8 x i1> @llvm.fptosi.sat.v8i1.v8f32(<8 x float>)
-declare <8 x i1> @llvm.fptoui.sat.v8i1.v8f32(<8 x float>)
-declare <8 x i8> @llvm.fptosi.sat.v8i8.v8f32(<8 x float>)
-declare <8 x i8> @llvm.fptoui.sat.v8i8.v8f32(<8 x float>)
-declare <8 x i16> @llvm.fptosi.sat.v8i16.v8f32(<8 x float>)
-declare <8 x i16> @llvm.fptoui.sat.v8i16.v8f32(<8 x float>)
-declare <8 x i32> @llvm.fptosi.sat.v8i32.v8f32(<8 x float>)
-declare <8 x i32> @llvm.fptoui.sat.v8i32.v8f32(<8 x float>)
-declare <8 x i64> @llvm.fptosi.sat.v8i64.v8f32(<8 x float>)
-declare <8 x i64> @llvm.fptoui.sat.v8i64.v8f32(<8 x float>)
-
-declare <8 x i1> @llvm.fptosi.sat.v8i1.v8f64(<8 x double>)
-declare <8 x i1> @llvm.fptoui.sat.v8i1.v8f64(<8 x double>)
-declare <8 x i8> @llvm.fptosi.sat.v8i8.v8f64(<8 x double>)
-declare <8 x i8> @llvm.fptoui.sat.v8i8.v8f64(<8 x double>)
-declare <8 x i16> @llvm.fptosi.sat.v8i16.v8f64(<8 x double>)
-declare <8 x i16> @llvm.fptoui.sat.v8i16.v8f64(<8 x double>)
-declare <8 x i32> @llvm.fptosi.sat.v8i32.v8f64(<8 x double>)
-declare <8 x i32> @llvm.fptoui.sat.v8i32.v8f64(<8 x double>)
-declare <8 x i64> @llvm.fptosi.sat.v8i64.v8f64(<8 x double>)
-declare <8 x i64> @llvm.fptoui.sat.v8i64.v8f64(<8 x double>)
-
-declare <16 x i1> @llvm.fptosi.sat.v16i1.v16f32(<16 x float>)
-declare <16 x i1> @llvm.fptoui.sat.v16i1.v16f32(<16 x float>)
-declare <16 x i8> @llvm.fptosi.sat.v16i8.v16f32(<16 x float>)
-declare <16 x i8> @llvm.fptoui.sat.v16i8.v16f32(<16 x float>)
-declare <16 x i16> @llvm.fptosi.sat.v16i16.v16f32(<16 x float>)
-declare <16 x i16> @llvm.fptoui.sat.v16i16.v16f32(<16 x float>)
-declare <16 x i32> @llvm.fptosi.sat.v16i32.v16f32(<16 x float>)
-declare <16 x i32> @llvm.fptoui.sat.v16i32.v16f32(<16 x float>)
-declare <16 x i64> @llvm.fptosi.sat.v16i64.v16f32(<16 x float>)
-declare <16 x i64> @llvm.fptoui.sat.v16i64.v16f32(<16 x float>)
-
-declare <16 x i1> @llvm.fptosi.sat.v16i1.v16f64(<16 x double>)
-declare <16 x i1> @llvm.fptoui.sat.v16i1.v16f64(<16 x double>)
-declare <16 x i8> @llvm.fptosi.sat.v16i8.v16f64(<16 x double>)
-declare <16 x i8> @llvm.fptoui.sat.v16i8.v16f64(<16 x double>)
-declare <16 x i16> @llvm.fptosi.sat.v16i16.v16f64(<16 x double>)
-declare <16 x i16> @llvm.fptoui.sat.v16i16.v16f64(<16 x double>)
-declare <16 x i32> @llvm.fptosi.sat.v16i32.v16f64(<16 x double>)
-declare <16 x i32> @llvm.fptoui.sat.v16i32.v16f64(<16 x double>)
-declare <16 x i64> @llvm.fptosi.sat.v16i64.v16f64(<16 x double>)
-declare <16 x i64> @llvm.fptoui.sat.v16i64.v16f64(<16 x double>)
-
-declare i1 @llvm.fptosi.sat.i1.f16(half)
-declare i1 @llvm.fptoui.sat.i1.f16(half)
-declare i8 @llvm.fptosi.sat.i8.f16(half)
-declare i8 @llvm.fptoui.sat.i8.f16(half)
-declare i16 @llvm.fptosi.sat.i16.f16(half)
-declare i16 @llvm.fptoui.sat.i16.f16(half)
-declare i32 @llvm.fptosi.sat.i32.f16(half)
-declare i32 @llvm.fptoui.sat.i32.f16(half)
-declare i64 @llvm.fptosi.sat.i64.f16(half)
-declare i64 @llvm.fptoui.sat.i64.f16(half)
-
-declare <2 x i1> @llvm.fptosi.sat.v2i1.v2f16(<2 x half>)
-declare <2 x i1> @llvm.fptoui.sat.v2i1.v2f16(<2 x half>)
-declare <2 x i8> @llvm.fptosi.sat.v2i8.v2f16(<2 x half>)
-declare <2 x i8> @llvm.fptoui.sat.v2i8.v2f16(<2 x half>)
-declare <2 x i16> @llvm.fptosi.sat.v2i16.v2f16(<2 x half>)
-declare <2 x i16> @llvm.fptoui.sat.v2i16.v2f16(<2 x half>)
-declare <2 x i32> @llvm.fptosi.sat.v2i32.v2f16(<2 x half>)
-declare <2 x i32> @llvm.fptoui.sat.v2i32.v2f16(<2 x half>)
-declare <2 x i64> @llvm.fptosi.sat.v2i64.v2f16(<2 x half>)
-declare <2 x i64> @llvm.fptoui.sat.v2i64.v2f16(<2 x half>)
-
-declare <4 x i1> @llvm.fptosi.sat.v4i1.v4f16(<4 x half>)
-declare <4 x i1> @llvm.fptoui.sat.v4i1.v4f16(<4 x half>)
-declare <4 x i8> @llvm.fptosi.sat.v4i8.v4f16(<4 x half>)
-declare <4 x i8> @llvm.fptoui.sat.v4i8.v4f16(<4 x half>)
-declare <4 x i16> @llvm.fptosi.sat.v4i16.v4f16(<4 x half>)
-declare <4 x i16> @llvm.fptoui.sat.v4i16.v4f16(<4 x half>)
-declare <4 x i32> @llvm.fptosi.sat.v4i32.v4f16(<4 x half>)
-declare <4 x i32> @llvm.fptoui.sat.v4i32.v4f16(<4 x half>)
-declare <4 x i64> @llvm.fptosi.sat.v4i64.v4f16(<4 x half>)
-declare <4 x i64> @llvm.fptoui.sat.v4i64.v4f16(<4 x half>)
-
-declare <8 x i1> @llvm.fptosi.sat.v8i1.v8f16(<8 x half>)
-declare <8 x i1> @llvm.fptoui.sat.v8i1.v8f16(<8 x half>)
-declare <8 x i8> @llvm.fptosi.sat.v8i8.v8f16(<8 x half>)
-declare <8 x i8> @llvm.fptoui.sat.v8i8.v8f16(<8 x half>)
-declare <8 x i16> @llvm.fptosi.sat.v8i16.v8f16(<8 x half>)
-declare <8 x i16> @llvm.fptoui.sat.v8i16.v8f16(<8 x half>)
-declare <8 x i32> @llvm.fptosi.sat.v8i32.v8f16(<8 x half>)
-declare <8 x i32> @llvm.fptoui.sat.v8i32.v8f16(<8 x half>)
-declare <8 x i64> @llvm.fptosi.sat.v8i64.v8f16(<8 x half>)
-declare <8 x i64> @llvm.fptoui.sat.v8i64.v8f16(<8 x half>)
-
-declare <16 x i1> @llvm.fptosi.sat.v16i1.v16f16(<16 x half>)
-declare <16 x i1> @llvm.fptoui.sat.v16i1.v16f16(<16 x half>)
-declare <16 x i8> @llvm.fptosi.sat.v16i8.v16f16(<16 x half>)
-declare <16 x i8> @llvm.fptoui.sat.v16i8.v16f16(<16 x half>)
-declare <16 x i16> @llvm.fptosi.sat.v16i16.v16f16(<16 x half>)
-declare <16 x i16> @llvm.fptoui.sat.v16i16.v16f16(<16 x half>)
-declare <16 x i32> @llvm.fptosi.sat.v16i32.v16f16(<16 x half>)
-declare <16 x i32> @llvm.fptoui.sat.v16i32.v16f16(<16 x half>)
-declare <16 x i64> @llvm.fptosi.sat.v16i64.v16f16(<16 x half>)
-declare <16 x i64> @llvm.fptoui.sat.v16i64.v16f16(<16 x half>)
diff --git a/llvm/test/Analysis/CostModel/AArch64/sve-fptoi_sat.ll b/llvm/test/Analysis/CostModel/AArch64/sve-fptoi_sat.ll
new file mode 100644
index 0000000000000..5aa2573639df2
--- /dev/null
+++ b/llvm/test/Analysis/CostModel/AArch64/sve-fptoi_sat.ll
@@ -0,0 +1,179 @@
+; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
+; RUN: opt -passes="print<cost-model>" -cost-kind=all 2>&1 -disable-output -mtriple=aarch64 -mattr=+sve2 %s | FileCheck %s
+
+target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128"
+
+define void @casts() {
+; CHECK-LABEL: 'casts'
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32s1 = call <vscale x 2 x i1> @llvm.fptosi.sat.nxv2i1.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32u1 = call <vscale x 2 x i1> @llvm.fptoui.sat.nxv2i1.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32s8 = call <vscale x 2 x i8> @llvm.fptosi.sat.nxv2i8.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32u8 = call <vscale x 2 x i8> @llvm.fptoui.sat.nxv2i8.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32s16 = call <vscale x 2 x i16> @llvm.fptosi.sat.nxv2i16.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32u16 = call <vscale x 2 x i16> @llvm.fptoui.sat.nxv2i16.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32s32 = call <vscale x 2 x i32> @llvm.fptosi.sat.nxv2i32.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32u32 = call <vscale x 2 x i32> @llvm.fptoui.sat.nxv2i32.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of RThru:6 CodeSize:5 Lat:6 SizeLat:5 for: %v2f32s64 = call <vscale x 2 x i64> @llvm.fptosi.sat.nxv2i64.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f32u64 = call <vscale x 2 x i64> @llvm.fptoui.sat.nxv2i64.nxv2f32(<vscale x 2 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64s1 = call <vscale x 2 x i1> @llvm.fptosi.sat.nxv2i1.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64u1 = call <vscale x 2 x i1> @llvm.fptoui.sat.nxv2i1.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64s8 = call <vscale x 2 x i8> @llvm.fptosi.sat.nxv2i8.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64u8 = call <vscale x 2 x i8> @llvm.fptoui.sat.nxv2i8.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64s16 = call <vscale x 2 x i16> @llvm.fptosi.sat.nxv2i16.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64u16 = call <vscale x 2 x i16> @llvm.fptoui.sat.nxv2i16.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64s32 = call <vscale x 2 x i32> @llvm.fptosi.sat.nxv2i32.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64u32 = call <vscale x 2 x i32> @llvm.fptoui.sat.nxv2i32.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64s64 = call <vscale x 2 x i64> @llvm.fptosi.sat.nxv2i64.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v2f64u64 = call <vscale x 2 x i64> @llvm.fptoui.sat.nxv2i64.nxv2f64(<vscale x 2 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v4f32s1 = call <vscale x 4 x i1> @llvm.fptosi.sat.nxv4i1.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v4f32u1 = call <vscale x 4 x i1> @llvm.fptoui.sat.nxv4i1.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v4f32s8 = call <vscale x 4 x i8> @llvm.fptosi.sat.nxv4i8.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v4f32u8 = call <vscale x 4 x i8> @llvm.fptoui.sat.nxv4i8.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v4f32s16 = call <vscale x 4 x i16> @llvm.fptosi.sat.nxv4i16.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v4f32u16 = call <vscale x 4 x i16> @llvm.fptoui.sat.nxv4i16.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v4f32s32 = call <vscale x 4 x i32> @llvm.fptosi.sat.nxv4i32.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 3 for: %v4f32u32 = call <vscale x 4 x i32> @llvm.fptoui.sat.nxv4i32.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of RThru:10 CodeSize:9 Lat:10 SizeLat:9 for: %v4f32s64 = call <vscale x 4 x i64> @llvm.fptosi.sat.nxv4i64.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f32u64 = call <vscale x 4 x i64> @llvm.fptoui.sat.nxv4i64.nxv4f32(<vscale x 4 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64s1 = call <vscale x 4 x i1> @llvm.fptosi.sat.nxv4i1.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64u1 = call <vscale x 4 x i1> @llvm.fptoui.sat.nxv4i1.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64s8 = call <vscale x 4 x i8> @llvm.fptosi.sat.nxv4i8.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64u8 = call <vscale x 4 x i8> @llvm.fptoui.sat.nxv4i8.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64s16 = call <vscale x 4 x i16> @llvm.fptosi.sat.nxv4i16.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64u16 = call <vscale x 4 x i16> @llvm.fptoui.sat.nxv4i16.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64s32 = call <vscale x 4 x i32> @llvm.fptosi.sat.nxv4i32.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64u32 = call <vscale x 4 x i32> @llvm.fptoui.sat.nxv4i32.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64s64 = call <vscale x 4 x i64> @llvm.fptosi.sat.nxv4i64.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v4f64u64 = call <vscale x 4 x i64> @llvm.fptoui.sat.nxv4i64.nxv4f64(<vscale x 4 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v8f32s1 = call <vscale x 8 x i1> @llvm.fptosi.sat.nxv8i1.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v8f32u1 = call <vscale x 8 x i1> @llvm.fptoui.sat.nxv8i1.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v8f32s8 = call <vscale x 8 x i8> @llvm.fptosi.sat.nxv8i8.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v8f32u8 = call <vscale x 8 x i8> @llvm.fptoui.sat.nxv8i8.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v8f32s16 = call <vscale x 8 x i16> @llvm.fptosi.sat.nxv8i16.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v8f32u16 = call <vscale x 8 x i16> @llvm.fptoui.sat.nxv8i16.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v8f32s32 = call <vscale x 8 x i32> @llvm.fptosi.sat.nxv8i32.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 6 for: %v8f32u32 = call <vscale x 8 x i32> @llvm.fptoui.sat.nxv8i32.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of RThru:20 CodeSize:18 Lat:20 SizeLat:18 for: %v8f32s64 = call <vscale x 8 x i64> @llvm.fptosi.sat.nxv8i64.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f32u64 = call <vscale x 8 x i64> @llvm.fptoui.sat.nxv8i64.nxv8f32(<vscale x 8 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64s1 = call <vscale x 8 x i1> @llvm.fptosi.sat.nxv8i1.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64u1 = call <vscale x 8 x i1> @llvm.fptoui.sat.nxv8i1.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64s8 = call <vscale x 8 x i8> @llvm.fptosi.sat.nxv8i8.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64u8 = call <vscale x 8 x i8> @llvm.fptoui.sat.nxv8i8.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64s16 = call <vscale x 8 x i16> @llvm.fptosi.sat.nxv8i16.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64u16 = call <vscale x 8 x i16> @llvm.fptoui.sat.nxv8i16.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64s32 = call <vscale x 8 x i32> @llvm.fptosi.sat.nxv8i32.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64u32 = call <vscale x 8 x i32> @llvm.fptoui.sat.nxv8i32.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64s64 = call <vscale x 8 x i64> @llvm.fptosi.sat.nxv8i64.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v8f64u64 = call <vscale x 8 x i64> @llvm.fptoui.sat.nxv8i64.nxv8f64(<vscale x 8 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v16f32s1 = call <vscale x 16 x i1> @llvm.fptosi.sat.nxv16i1.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v16f32u1 = call <vscale x 16 x i1> @llvm.fptoui.sat.nxv16i1.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v16f32s8 = call <vscale x 16 x i8> @llvm.fptosi.sat.nxv16i8.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v16f32u8 = call <vscale x 16 x i8> @llvm.fptoui.sat.nxv16i8.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v16f32s16 = call <vscale x 16 x i16> @llvm.fptosi.sat.nxv16i16.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v16f32u16 = call <vscale x 16 x i16> @llvm.fptoui.sat.nxv16i16.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v16f32s32 = call <vscale x 16 x i32> @llvm.fptosi.sat.nxv16i32.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 12 for: %v16f32u32 = call <vscale x 16 x i32> @llvm.fptoui.sat.nxv16i32.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of RThru:40 CodeSize:36 Lat:40 SizeLat:36 for: %v16f32s64 = call <vscale x 16 x i64> @llvm.fptosi.sat.nxv16i64.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f32u64 = call <vscale x 16 x i64> @llvm.fptoui.sat.nxv16i64.nxv16f32(<vscale x 16 x float> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64s1 = call <vscale x 16 x i1> @llvm.fptosi.sat.nxv16i1.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64u1 = call <vscale x 16 x i1> @llvm.fptoui.sat.nxv16i1.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64s8 = call <vscale x 16 x i8> @llvm.fptosi.sat.nxv16i8.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64u8 = call <vscale x 16 x i8> @llvm.fptoui.sat.nxv16i8.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64s16 = call <vscale x 16 x i16> @llvm.fptosi.sat.nxv16i16.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64u16 = call <vscale x 16 x i16> @llvm.fptoui.sat.nxv16i16.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64s32 = call <vscale x 16 x i32> @llvm.fptosi.sat.nxv16i32.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64u32 = call <vscale x 16 x i32> @llvm.fptoui.sat.nxv16i32.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64s64 = call <vscale x 16 x i64> @llvm.fptosi.sat.nxv16i64.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of 24 for: %v16f64u64 = call <vscale x 16 x i64> @llvm.fptoui.sat.nxv16i64.nxv16f64(<vscale x 16 x double> undef)
+; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %v2f32s1 = call <vscale x 2 x i1> @llvm.fptosi.sat.nxv2i1.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32u1 = call <vscale x 2 x i1> @llvm.fptoui.sat.nxv2i1.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32s8 = call <vscale x 2 x i8> @llvm.fptosi.sat.nxv2i8.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32u8 = call <vscale x 2 x i8> @llvm.fptoui.sat.nxv2i8.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32s16 = call <vscale x 2 x i16> @llvm.fptosi.sat.nxv2i16.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32u16 = call <vscale x 2 x i16> @llvm.fptoui.sat.nxv2i16.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32s32 = call <vscale x 2 x i32> @llvm.fptosi.sat.nxv2i32.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32u32 = call <vscale x 2 x i32> @llvm.fptoui.sat.nxv2i32.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32s64 = call <vscale x 2 x i64> @llvm.fptosi.sat.nxv2i64.nxv2f32(<vscale x 2 x float> undef)
+ %v2f32u64 = call <vscale x 2 x i64> @llvm.fptoui.sat.nxv2i64.nxv2f32(<vscale x 2 x float> undef)
+
+ %v2f64s1 = call <vscale x 2 x i1> @llvm.fptosi.sat.nxv2i1.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64u1 = call <vscale x 2 x i1> @llvm.fptoui.sat.nxv2i1.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64s8 = call <vscale x 2 x i8> @llvm.fptosi.sat.nxv2i8.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64u8 = call <vscale x 2 x i8> @llvm.fptoui.sat.nxv2i8.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64s16 = call <vscale x 2 x i16> @llvm.fptosi.sat.nxv2i16.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64u16 = call <vscale x 2 x i16> @llvm.fptoui.sat.nxv2i16.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64s32 = call <vscale x 2 x i32> @llvm.fptosi.sat.nxv2i32.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64u32 = call <vscale x 2 x i32> @llvm.fptoui.sat.nxv2i32.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64s64 = call <vscale x 2 x i64> @llvm.fptosi.sat.nxv2i64.nxv2f64(<vscale x 2 x double> undef)
+ %v2f64u64 = call <vscale x 2 x i64> @llvm.fptoui.sat.nxv2i64.nxv2f64(<vscale x 2 x double> undef)
+
+ %v4f32s1 = call <vscale x 4 x i1> @llvm.fptosi.sat.nxv4i1.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32u1 = call <vscale x 4 x i1> @llvm.fptoui.sat.nxv4i1.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32s8 = call <vscale x 4 x i8> @llvm.fptosi.sat.nxv4i8.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32u8 = call <vscale x 4 x i8> @llvm.fptoui.sat.nxv4i8.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32s16 = call <vscale x 4 x i16> @llvm.fptosi.sat.nxv4i16.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32u16 = call <vscale x 4 x i16> @llvm.fptoui.sat.nxv4i16.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32s32 = call <vscale x 4 x i32> @llvm.fptosi.sat.nxv4i32.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32u32 = call <vscale x 4 x i32> @llvm.fptoui.sat.nxv4i32.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32s64 = call <vscale x 4 x i64> @llvm.fptosi.sat.nxv4i64.nxv4f32(<vscale x 4 x float> undef)
+ %v4f32u64 = call <vscale x 4 x i64> @llvm.fptoui.sat.nxv4i64.nxv4f32(<vscale x 4 x float> undef)
+
+ %v4f64s1 = call <vscale x 4 x i1> @llvm.fptosi.sat.nxv4i1.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64u1 = call <vscale x 4 x i1> @llvm.fptoui.sat.nxv4i1.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64s8 = call <vscale x 4 x i8> @llvm.fptosi.sat.nxv4i8.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64u8 = call <vscale x 4 x i8> @llvm.fptoui.sat.nxv4i8.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64s16 = call <vscale x 4 x i16> @llvm.fptosi.sat.nxv4i16.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64u16 = call <vscale x 4 x i16> @llvm.fptoui.sat.nxv4i16.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64s32 = call <vscale x 4 x i32> @llvm.fptosi.sat.nxv4i32.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64u32 = call <vscale x 4 x i32> @llvm.fptoui.sat.nxv4i32.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64s64 = call <vscale x 4 x i64> @llvm.fptosi.sat.nxv4i64.nxv4f64(<vscale x 4 x double> undef)
+ %v4f64u64 = call <vscale x 4 x i64> @llvm.fptoui.sat.nxv4i64.nxv4f64(<vscale x 4 x double> undef)
+
+ %v8f32s1 = call <vscale x 8 x i1> @llvm.fptosi.sat.nxv8i1.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32u1 = call <vscale x 8 x i1> @llvm.fptoui.sat.nxv8i1.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32s8 = call <vscale x 8 x i8> @llvm.fptosi.sat.nxv8i8.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32u8 = call <vscale x 8 x i8> @llvm.fptoui.sat.nxv8i8.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32s16 = call <vscale x 8 x i16> @llvm.fptosi.sat.nxv8i16.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32u16 = call <vscale x 8 x i16> @llvm.fptoui.sat.nxv8i16.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32s32 = call <vscale x 8 x i32> @llvm.fptosi.sat.nxv8i32.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32u32 = call <vscale x 8 x i32> @llvm.fptoui.sat.nxv8i32.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32s64 = call <vscale x 8 x i64> @llvm.fptosi.sat.nxv8i64.nxv8f32(<vscale x 8 x float> undef)
+ %v8f32u64 = call <vscale x 8 x i64> @llvm.fptoui.sat.nxv8i64.nxv8f32(<vscale x 8 x float> undef)
+
+ %v8f64s1 = call <vscale x 8 x i1> @llvm.fptosi.sat.nxv8i1.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64u1 = call <vscale x 8 x i1> @llvm.fptoui.sat.nxv8i1.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64s8 = call <vscale x 8 x i8> @llvm.fptosi.sat.nxv8i8.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64u8 = call <vscale x 8 x i8> @llvm.fptoui.sat.nxv8i8.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64s16 = call <vscale x 8 x i16> @llvm.fptosi.sat.nxv8i16.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64u16 = call <vscale x 8 x i16> @llvm.fptoui.sat.nxv8i16.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64s32 = call <vscale x 8 x i32> @llvm.fptosi.sat.nxv8i32.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64u32 = call <vscale x 8 x i32> @llvm.fptoui.sat.nxv8i32.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64s64 = call <vscale x 8 x i64> @llvm.fptosi.sat.nxv8i64.nxv8f64(<vscale x 8 x double> undef)
+ %v8f64u64 = call <vscale x 8 x i64> @llvm.fptoui.sat.nxv8i64.nxv8f64(<vscale x 8 x double> undef)
+
+ %v16f32s1 = call <vscale x 16 x i1> @llvm.fptosi.sat.nxv16i1.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32u1 = call <vscale x 16 x i1> @llvm.fptoui.sat.nxv16i1.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32s8 = call <vscale x 16 x i8> @llvm.fptosi.sat.nxv16i8.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32u8 = call <vscale x 16 x i8> @llvm.fptoui.sat.nxv16i8.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32s16 = call <vscale x 16 x i16> @llvm.fptosi.sat.nxv16i16.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32u16 = call <vscale x 16 x i16> @llvm.fptoui.sat.nxv16i16.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32s32 = call <vscale x 16 x i32> @llvm.fptosi.sat.nxv16i32.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32u32 = call <vscale x 16 x i32> @llvm.fptoui.sat.nxv16i32.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32s64 = call <vscale x 16 x i64> @llvm.fptosi.sat.nxv16i64.nxv16f32(<vscale x 16 x float> undef)
+ %v16f32u64 = call <vscale x 16 x i64> @llvm.fptoui.sat.nxv16i64.nxv16f32(<vscale x 16 x float> undef)
+
+ %v16f64s1 = call <vscale x 16 x i1> @llvm.fptosi.sat.nxv16i1.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64u1 = call <vscale x 16 x i1> @llvm.fptoui.sat.nxv16i1.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64s8 = call <vscale x 16 x i8> @llvm.fptosi.sat.nxv16i8.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64u8 = call <vscale x 16 x i8> @llvm.fptoui.sat.nxv16i8.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64s16 = call <vscale x 16 x i16> @llvm.fptosi.sat.nxv16i16.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64u16 = call <vscale x 16 x i16> @llvm.fptoui.sat.nxv16i16.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64s32 = call <vscale x 16 x i32> @llvm.fptosi.sat.nxv16i32.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64u32 = call <vscale x 16 x i32> @llvm.fptoui.sat.nxv16i32.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64s64 = call <vscale x 16 x i64> @llvm.fptosi.sat.nxv16i64.nxv16f64(<vscale x 16 x double> undef)
+ %v16f64u64 = call <vscale x 16 x i64> @llvm.fptoui.sat.nxv16i64.nxv16f64(<vscale x 16 x double> undef)
+
+ ret void
+}
More information about the llvm-commits
mailing list