[llvm-branch-commits] [llvm] release/22.x: [RISCV] Fix i64 gather/scatter cost on rv32 (#176105) (PR #176899)
Cullen Rhodes via llvm-branch-commits
llvm-branch-commits at lists.llvm.org
Wed Jan 28 01:23:25 PST 2026
https://github.com/c-rhodes updated https://github.com/llvm/llvm-project/pull/176899
>From 06b159af459b76dfd5f50c97fb067b5ea67f6f70 Mon Sep 17 00:00:00 2001
From: Luke Lau <luke at igalia.com>
Date: Thu, 15 Jan 2026 16:07:52 +0800
Subject: [PATCH 1/2] [RISCV] Add tests for rv32 gather/scatter costs. NFC
There's a divergence with the rv32 costs that I plan on fixing in
another patch, so this precommits the tests for them.
The zve32f RUN lines were split off into another file so the check prefixes
are easier to reason about.
The -riscv-v-vector-bits-max RUN lines were also removed to simplify the
check prefixes since I'm not sure if they were intentionally testing any
specific logic.
(cherry picked from commit 3ad6d350c44f54482a86a7eb488732093eaed372)
---
.../CostModel/RISCV/scalable-gather-zve32f.ll | 112 ++++++++
.../CostModel/RISCV/scalable-gather.ll | 243 +++++++++--------
.../RISCV/scalable-scatter-zve32f.ll | 125 +++++++++
.../CostModel/RISCV/scalable-scatter.ll | 245 ++++++++++--------
4 files changed, 500 insertions(+), 225 deletions(-)
create mode 100644 llvm/test/Analysis/CostModel/RISCV/scalable-gather-zve32f.ll
create mode 100644 llvm/test/Analysis/CostModel/RISCV/scalable-scatter-zve32f.ll
diff --git a/llvm/test/Analysis/CostModel/RISCV/scalable-gather-zve32f.ll b/llvm/test/Analysis/CostModel/RISCV/scalable-gather-zve32f.ll
new file mode 100644
index 0000000000000..20749d07c44fa
--- /dev/null
+++ b/llvm/test/Analysis/CostModel/RISCV/scalable-gather-zve32f.ll
@@ -0,0 +1,112 @@
+; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --version 6
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 -mattr=+zve32f,+zvl128b < %s | FileCheck %s --check-prefixes=RV32
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+zve32f,+zvl128b < %s | FileCheck %s --check-prefixes=RV64
+
+define void @masked_gather() {
+; RV32-LABEL: 'masked_gather'
+; RV32-NEXT: Cost Model: Invalid cost for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; RV64-LABEL: 'masked_gather'
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+ %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
+ %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
+ %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
+ %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
+
+ %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> undef, i32 4, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
+ %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> undef, i32 4, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
+ %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> undef, i32 4, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
+ %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> undef, i32 4, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
+ %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> undef, i32 4, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
+
+ %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
+ %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
+ %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
+ %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
+
+ %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> undef, i32 4, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
+ %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> undef, i32 4, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
+ %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> undef, i32 4, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
+ %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> undef, i32 4, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
+ %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> undef, i32 4, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
+
+ %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> undef, i32 2, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
+ %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> undef, i32 2, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
+ %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> undef, i32 2, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
+ %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> undef, i32 2, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
+ %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> undef, i32 2, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
+ %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> undef, i32 2, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
+
+ %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> undef, i32 1, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
+ %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> undef, i32 1, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
+ %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> undef, i32 1, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
+ %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> undef, i32 1, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
+ %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> undef, i32 1, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
+ %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
+ %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
+
+ ret void
+}
diff --git a/llvm/test/Analysis/CostModel/RISCV/scalable-gather.ll b/llvm/test/Analysis/CostModel/RISCV/scalable-gather.ll
index 997ec12b714f6..e4faaceea38d8 100644
--- a/llvm/test/Analysis/CostModel/RISCV/scalable-gather.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/scalable-gather.ll
@@ -1,86 +1,79 @@
; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfh < %s | FileCheck %s
-; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfhmin < %s | FileCheck %s
-; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfh -riscv-v-vector-bits-max=256 < %s | FileCheck %s --check-prefixes=CHECK,MAX256
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 -mattr=+v,+f,+d,+zfh,+zvfh < %s | FileCheck %s --check-prefixes=CHECK,SUPPORTED,RV32
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfh < %s | FileCheck %s --check-prefixes=CHECK,SUPPORTED,RV64
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 -mattr=+v,+f,+d,+zfh,+zvfhmin < %s | FileCheck %s --check-prefixes=CHECK,SUPPORTED,RV32
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfhmin < %s | FileCheck %s --check-prefixes=CHECK,SUPPORTED,RV64
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 < %s | FileCheck %s --check-prefixes=CHECK,UNSUPPORTED
; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 < %s | FileCheck %s --check-prefixes=CHECK,UNSUPPORTED
-; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+zve32f,+zvl128b,+f,+d,+zfh,+zvfh < %s | FileCheck %s --check-prefixes=CHECK,UNSUPPORTED
define void @masked_gather_aligned() {
-; GENERIC-LABEL: 'masked_gather_aligned'
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> undef, i32 4, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> undef, i32 4, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> undef, i32 4, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> undef, i32 4, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> undef, i32 4, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> undef, i32 4, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> undef, i32 4, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> undef, i32 4, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> undef, i32 4, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> undef, i32 4, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> undef, i32 2, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> undef, i32 2, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> undef, i32 2, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> undef, i32 2, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> undef, i32 2, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> undef, i32 2, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> undef, i32 1, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> undef, i32 1, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> undef, i32 1, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> undef, i32 1, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> undef, i32 1, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; RV32-LABEL: 'masked_gather_aligned'
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
-; MAX256-LABEL: 'masked_gather_aligned'
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; RV64-LABEL: 'masked_gather_aligned'
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; UNSUPPORTED-LABEL: 'masked_gather_aligned'
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
@@ -114,10 +107,6 @@ define void @masked_gather_aligned() {
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
-; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
-; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
-; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
-; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
; UNSUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
@@ -157,32 +146,18 @@ define void @masked_gather_aligned() {
%V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
%V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
-
- %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
- %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
- %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
- %V1PTR= call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
ret void
}
define void @masked_gather_aligned_f16() {
-; GENERIC-LABEL: 'masked_gather_aligned_f16'
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32F16 = call <vscale x 32 x half> @llvm.masked.gather.nxv32f16.nxv32p0(<vscale x 32 x ptr> undef, i32 2, <vscale x 32 x i1> undef, <vscale x 32 x half> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F16 = call <vscale x 16 x half> @llvm.masked.gather.nxv16f16.nxv16p0(<vscale x 16 x ptr> undef, i32 2, <vscale x 16 x i1> undef, <vscale x 16 x half> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F16 = call <vscale x 8 x half> @llvm.masked.gather.nxv8f16.nxv8p0(<vscale x 8 x ptr> undef, i32 2, <vscale x 8 x i1> undef, <vscale x 8 x half> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F16 = call <vscale x 4 x half> @llvm.masked.gather.nxv4f16.nxv4p0(<vscale x 4 x ptr> undef, i32 2, <vscale x 4 x i1> undef, <vscale x 4 x half> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F16 = call <vscale x 2 x half> @llvm.masked.gather.nxv2f16.nxv2p0(<vscale x 2 x ptr> undef, i32 2, <vscale x 2 x i1> undef, <vscale x 2 x half> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F16 = call <vscale x 1 x half> @llvm.masked.gather.nxv1f16.nxv1p0(<vscale x 1 x ptr> undef, i32 2, <vscale x 1 x i1> undef, <vscale x 1 x half> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
-;
-; MAX256-LABEL: 'masked_gather_aligned_f16'
-; MAX256-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32F16 = call <vscale x 32 x half> @llvm.masked.gather.nxv32f16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x half> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F16 = call <vscale x 16 x half> @llvm.masked.gather.nxv16f16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x half> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F16 = call <vscale x 8 x half> @llvm.masked.gather.nxv8f16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x half> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F16 = call <vscale x 4 x half> @llvm.masked.gather.nxv4f16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x half> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F16 = call <vscale x 2 x half> @llvm.masked.gather.nxv2f16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x half> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F16 = call <vscale x 1 x half> @llvm.masked.gather.nxv1f16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x half> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; SUPPORTED-LABEL: 'masked_gather_aligned_f16'
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32F16 = call <vscale x 32 x half> @llvm.masked.gather.nxv32f16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x half> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F16 = call <vscale x 16 x half> @llvm.masked.gather.nxv16f16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x half> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F16 = call <vscale x 8 x half> @llvm.masked.gather.nxv8f16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x half> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F16 = call <vscale x 4 x half> @llvm.masked.gather.nxv4f16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x half> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F16 = call <vscale x 2 x half> @llvm.masked.gather.nxv2f16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x half> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F16 = call <vscale x 1 x half> @llvm.masked.gather.nxv1f16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x half> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; UNSUPPORTED-LABEL: 'masked_gather_aligned_f16'
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V32F16 = call <vscale x 32 x half> @llvm.masked.gather.nxv32f16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x half> undef)
@@ -229,10 +204,6 @@ define void @masked_gather_unaligned() {
; CHECK-NEXT: Cost Model: Invalid cost for instruction: %V4I16.u = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
; CHECK-NEXT: Cost Model: Invalid cost for instruction: %V2I16.u = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
; CHECK-NEXT: Cost Model: Invalid cost for instruction: %V1I16.u = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
-; CHECK-NEXT: Cost Model: Invalid cost for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
-; CHECK-NEXT: Cost Model: Invalid cost for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
-; CHECK-NEXT: Cost Model: Invalid cost for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
-; CHECK-NEXT: Cost Model: Invalid cost for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
%V8F64.u = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> undef, i32 2, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
@@ -264,11 +235,6 @@ define void @masked_gather_unaligned() {
%V2I16.u = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
%V1I16.u = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
- %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, i32 1, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
- %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, i32 1, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
- %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
- %V1PTR= call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
-
ret void
}
@@ -291,3 +257,54 @@ define void @masked_gather_unaligned_f16() {
ret void
}
+
+define void @masked_gather_ptr_align4() {
+; RV32-LABEL: 'masked_gather_ptr_align4'
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; RV64-LABEL: 'masked_gather_ptr_align4'
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; UNSUPPORTED-LABEL: 'masked_gather_ptr_align4'
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
+; UNSUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+ %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
+ %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
+ %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
+ %V1PTR= call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
+ ret void
+}
+
+define void @masked_gather_ptr_align8() {
+; SUPPORTED-LABEL: 'masked_gather_ptr_align8'
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; UNSUPPORTED-LABEL: 'masked_gather_ptr_align8'
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V1PTR = call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
+; UNSUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+ %V8PTR = call <vscale x 8 x ptr> @llvm.masked.gather.nxv8p0.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x ptr> undef)
+ %V4PTR = call <vscale x 4 x ptr> @llvm.masked.gather.nxv4p0.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x ptr> undef)
+ %V2PTR = call <vscale x 2 x ptr> @llvm.masked.gather.nxv2p0.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x ptr> undef)
+ %V1PTR= call <vscale x 1 x ptr> @llvm.masked.gather.nxv1p0.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x ptr> undef)
+ ret void
+}
diff --git a/llvm/test/Analysis/CostModel/RISCV/scalable-scatter-zve32f.ll b/llvm/test/Analysis/CostModel/RISCV/scalable-scatter-zve32f.ll
new file mode 100644
index 0000000000000..982c227de553a
--- /dev/null
+++ b/llvm/test/Analysis/CostModel/RISCV/scalable-scatter-zve32f.ll
@@ -0,0 +1,125 @@
+; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --version 6
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 -mattr=+zve32f,+zvl128b < %s | FileCheck %s --check-prefixes=RV32
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+zve32f,+zvl128b < %s | FileCheck %s --check-prefixes=RV64
+
+define void @masked_scatter() {
+; RV32-LABEL: 'masked_scatter'
+; RV32-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 128 for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; RV64-LABEL: 'masked_scatter'
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+ call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef)
+
+ call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> undef, i32 4, <vscale x 16 x i1> undef)
+ call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> undef, i32 4, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> undef, i32 4, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> undef, i32 4, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> undef, i32 4, <vscale x 1 x i1> undef)
+
+ call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef)
+
+ call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> undef, i32 4, <vscale x 16 x i1> undef)
+ call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> undef, i32 4, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> undef, i32 4, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> undef, i32 4, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> undef, i32 4, <vscale x 1 x i1> undef)
+
+ call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> undef, i32 2, <vscale x 32 x i1> undef)
+ call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> undef, i32 2, <vscale x 16 x i1> undef)
+ call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> undef, i32 2, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> undef, i32 2, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> undef, i32 2, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> undef, i32 2, <vscale x 1 x i1> undef)
+
+ call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> undef, i32 1, <vscale x 64 x i1> undef)
+ call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> undef, i32 1, <vscale x 32 x i1> undef)
+ call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> undef, i32 1, <vscale x 16 x i1> undef)
+ call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> undef, i32 1, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> undef, i32 1, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef)
+
+ call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef)
+
+ ret void
+}
diff --git a/llvm/test/Analysis/CostModel/RISCV/scalable-scatter.ll b/llvm/test/Analysis/CostModel/RISCV/scalable-scatter.ll
index 69abcdea176cc..204fee14b9ce1 100644
--- a/llvm/test/Analysis/CostModel/RISCV/scalable-scatter.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/scalable-scatter.ll
@@ -1,86 +1,79 @@
; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfh < %s | FileCheck %s
-; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfhmin < %s | FileCheck %s
-; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfh -riscv-v-vector-bits-max=256 < %s | FileCheck %s --check-prefixes=CHECK,MAX256
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 -mattr=+v,+f,+d,+zfh,+zvfh < %s | FileCheck %s --check-prefixes=CHECK,SUPPORTED,RV32
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfh < %s | FileCheck %s --check-prefixes=CHECK,SUPPORTED,RV64
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 -mattr=+v,+f,+d,+zfh,+zvfhmin < %s | FileCheck %s --check-prefixes=CHECK,SUPPORTED,RV32
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v,+f,+d,+zfh,+zvfhmin < %s | FileCheck %s --check-prefixes=CHECK,SUPPORTED,RV64
+; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 < %s | FileCheck %s --check-prefixes=CHECK,UNSUPPORTED
; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 < %s | FileCheck %s --check-prefixes=CHECK,UNSUPPORTED
-; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+zve32f,+zvl128b,+f,+d,+zfh,+zvfh < %s | FileCheck %s --check-prefixes=CHECK,UNSUPPORTED
define void @masked_scatter_aligned() {
-; GENERIC-LABEL: 'masked_scatter_aligned'
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> undef, i32 4, <vscale x 16 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> undef, i32 4, <vscale x 8 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> undef, i32 4, <vscale x 4 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> undef, i32 4, <vscale x 2 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> undef, i32 4, <vscale x 1 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> undef, i32 4, <vscale x 16 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> undef, i32 4, <vscale x 8 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> undef, i32 4, <vscale x 4 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> undef, i32 4, <vscale x 2 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> undef, i32 4, <vscale x 1 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> undef, i32 2, <vscale x 32 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> undef, i32 2, <vscale x 16 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> undef, i32 2, <vscale x 8 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> undef, i32 2, <vscale x 4 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> undef, i32 2, <vscale x 2 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> undef, i32 2, <vscale x 1 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 128 for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> undef, i32 1, <vscale x 64 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> undef, i32 1, <vscale x 32 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> undef, i32 1, <vscale x 16 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> undef, i32 1, <vscale x 8 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> undef, i32 1, <vscale x 4 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; RV32-LABEL: 'masked_scatter_aligned'
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 128 for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
-; MAX256-LABEL: 'masked_scatter_aligned'
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 128 for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; RV64-LABEL: 'masked_scatter_aligned'
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 128 for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; UNSUPPORTED-LABEL: 'masked_scatter_aligned'
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
@@ -114,10 +107,6 @@ define void @masked_scatter_aligned() {
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
-; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
-; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
-; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
-; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
; UNSUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef)
@@ -157,32 +146,20 @@ define void @masked_scatter_aligned() {
call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef)
call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef)
- call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> undef, i32 8, <vscale x 8 x i1> undef)
- call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> undef, i32 8, <vscale x 4 x i1> undef)
- call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> undef, i32 8, <vscale x 2 x i1> undef)
- call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> undef, i32 8, <vscale x 1 x i1> undef)
-
ret void
}
define void @masked_scatter_aligned_f16() {
-; GENERIC-LABEL: 'masked_scatter_aligned_f16'
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32f16.nxv32p0(<vscale x 32 x half> undef, <vscale x 32 x ptr> undef, i32 2, <vscale x 32 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f16.nxv16p0(<vscale x 16 x half> undef, <vscale x 16 x ptr> undef, i32 2, <vscale x 16 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f16.nxv8p0(<vscale x 8 x half> undef, <vscale x 8 x ptr> undef, i32 2, <vscale x 8 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f16.nxv4p0(<vscale x 4 x half> undef, <vscale x 4 x ptr> undef, i32 2, <vscale x 4 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f16.nxv2p0(<vscale x 2 x half> undef, <vscale x 2 x ptr> undef, i32 2, <vscale x 2 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f16.nxv1p0(<vscale x 1 x half> undef, <vscale x 1 x ptr> undef, i32 2, <vscale x 1 x i1> undef)
-; GENERIC-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
-; MAX256-LABEL: 'masked_scatter_aligned_f16'
-; MAX256-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32f16.nxv32p0(<vscale x 32 x half> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f16.nxv16p0(<vscale x 16 x half> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f16.nxv8p0(<vscale x 8 x half> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f16.nxv4p0(<vscale x 4 x half> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f16.nxv2p0(<vscale x 2 x half> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f16.nxv1p0(<vscale x 1 x half> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
-; MAX256-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; SUPPORTED-LABEL: 'masked_scatter_aligned_f16'
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32f16.nxv32p0(<vscale x 32 x half> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f16.nxv16p0(<vscale x 16 x half> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f16.nxv8p0(<vscale x 8 x half> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f16.nxv4p0(<vscale x 4 x half> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f16.nxv2p0(<vscale x 2 x half> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f16.nxv1p0(<vscale x 1 x half> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; UNSUPPORTED-LABEL: 'masked_scatter_aligned_f16'
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv32f16.nxv32p0(<vscale x 32 x half> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
@@ -229,10 +206,6 @@ define void @masked_scatter_unaligned() {
; CHECK-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
; CHECK-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
; CHECK-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
-; CHECK-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
-; CHECK-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
-; CHECK-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
-; CHECK-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> undef, i32 2, <vscale x 8 x i1> undef)
@@ -264,11 +237,6 @@ define void @masked_scatter_unaligned() {
call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef)
call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef)
- call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> undef, i32 1, <vscale x 8 x i1> undef)
- call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> undef, i32 1, <vscale x 4 x i1> undef)
- call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> undef, i32 1, <vscale x 2 x i1> undef)
- call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> undef, i32 1, <vscale x 1 x i1> undef)
-
ret void
}
@@ -291,3 +259,56 @@ define void @masked_scatter_unaligned_f16() {
ret void
}
+
+define void @masked_scatter_ptr_align4() {
+; RV32-LABEL: 'masked_scatter_ptr_align4'
+; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV32-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; RV64-LABEL: 'masked_scatter_ptr_align4'
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; RV64-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; RV64-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; UNSUPPORTED-LABEL: 'masked_scatter_ptr_align4'
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; UNSUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+ call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+
+ ret void
+}
+
+define void @masked_scatter_ptr_align8() {
+; SUPPORTED-LABEL: 'masked_scatter_ptr_align8'
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+; UNSUPPORTED-LABEL: 'masked_scatter_ptr_align8'
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; UNSUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+;
+ call void @llvm.masked.scatter.nxv8p0.nxv8p0(<vscale x 8 x ptr> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+ call void @llvm.masked.scatter.nxv4p0.nxv4p0(<vscale x 4 x ptr> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+ call void @llvm.masked.scatter.nxv2p0.nxv2p0(<vscale x 2 x ptr> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+ call void @llvm.masked.scatter.nxv1p0.nxv1p0(<vscale x 1 x ptr> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+
+ ret void
+}
>From 1a9171828f7102fc8caf3d81b559f9bb645bd768 Mon Sep 17 00:00:00 2001
From: Luke Lau <luke at igalia.com>
Date: Tue, 20 Jan 2026 16:00:05 +0800
Subject: [PATCH 2/2] [RISCV] Fix i64 gather/scatter cost on rv32 (#176105)
Fixes #175909
We compute the cost of a gather/scatter by multiplying the cost of the
scalar element type memory op by the estimated number of elements. On
rv32 though a scalar i64 load costs 2, even if we have zve64x.
This causes the cost to diverge between a vector of f64 and vector of
i64, even though both are the same. This fixes it by just using
TTI::TCC_Basic as the scalar memory op cost. The element type is checked
to be legal at this point.
I think we have the same issue for the strided op cost, but we don't
have test coverage for it yet.
(cherry picked from commit 0c1257cd46456513016b106d964dc5ad47c6289b)
---
.../Target/RISCV/RISCVTargetTransformInfo.cpp | 7 +-
.../CostModel/RISCV/scalable-gather.ll | 100 ++++++------------
.../CostModel/RISCV/scalable-scatter.ll | 100 ++++++------------
3 files changed, 68 insertions(+), 139 deletions(-)
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index e812d092c3ea0..c0119325285c7 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -1184,7 +1184,6 @@ RISCVTTIImpl::getGatherScatterOpCost(const MemIntrinsicCostAttributes &MICA,
unsigned Opcode = IsLoad ? Instruction::Load : Instruction::Store;
Type *DataTy = MICA.getDataType();
Align Alignment = MICA.getAlignment();
- const Instruction *I = MICA.getInst();
if (CostKind != TTI::TCK_RecipThroughput)
return BaseT::getMemIntrinsicInstrCost(MICA, CostKind);
@@ -1198,11 +1197,8 @@ RISCVTTIImpl::getGatherScatterOpCost(const MemIntrinsicCostAttributes &MICA,
// scalable vectors, we use an estimate on that number since we don't
// know exactly what VL will be.
auto &VTy = *cast<VectorType>(DataTy);
- InstructionCost MemOpCost =
- getMemoryOpCost(Opcode, VTy.getElementType(), Alignment, 0, CostKind,
- {TTI::OK_AnyValue, TTI::OP_None}, I);
unsigned NumLoads = getEstimatedVLFor(&VTy);
- return NumLoads * MemOpCost;
+ return NumLoads * TTI::TCC_Basic;
}
InstructionCost RISCVTTIImpl::getExpandCompressMemoryOpCost(
@@ -1269,6 +1265,7 @@ RISCVTTIImpl::getStridedMemoryOpCost(const MemIntrinsicCostAttributes &MICA,
// Cost is proportional to the number of memory operations implied. For
// scalable vectors, we use an estimate on that number since we don't
// know exactly what VL will be.
+ // FIXME: This will overcost for i64 on rv32 with +zve64x.
auto &VTy = *cast<VectorType>(DataTy);
InstructionCost MemOpCost =
getMemoryOpCost(Opcode, VTy.getElementType(), Alignment, 0, CostKind,
diff --git a/llvm/test/Analysis/CostModel/RISCV/scalable-gather.ll b/llvm/test/Analysis/CostModel/RISCV/scalable-gather.ll
index e4faaceea38d8..c60c372c717f7 100644
--- a/llvm/test/Analysis/CostModel/RISCV/scalable-gather.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/scalable-gather.ll
@@ -7,73 +7,39 @@
; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 < %s | FileCheck %s --check-prefixes=CHECK,UNSUPPORTED
define void @masked_gather_aligned() {
-; RV32-LABEL: 'masked_gather_aligned'
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
-;
-; RV64-LABEL: 'masked_gather_aligned'
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; SUPPORTED-LABEL: 'masked_gather_aligned'
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F64 = call <vscale x 4 x double> @llvm.masked.gather.nxv4f64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x double> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F64 = call <vscale x 2 x double> @llvm.masked.gather.nxv2f64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x double> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F64 = call <vscale x 1 x double> @llvm.masked.gather.nxv1f64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x double> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16F32 = call <vscale x 16 x float> @llvm.masked.gather.nxv16f32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x float> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8F32 = call <vscale x 8 x float> @llvm.masked.gather.nxv8f32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x float> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4F32 = call <vscale x 4 x float> @llvm.masked.gather.nxv4f32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x float> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2F32 = call <vscale x 2 x float> @llvm.masked.gather.nxv2f32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x float> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1F32 = call <vscale x 1 x float> @llvm.masked.gather.nxv1f32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x float> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I64 = call <vscale x 8 x i64> @llvm.masked.gather.nxv8i64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x i64> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I64 = call <vscale x 4 x i64> @llvm.masked.gather.nxv4i64.nxv4p0(<vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef, <vscale x 4 x i64> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I64 = call <vscale x 2 x i64> @llvm.masked.gather.nxv2i64.nxv2p0(<vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef, <vscale x 2 x i64> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I64 = call <vscale x 1 x i64> @llvm.masked.gather.nxv1i64.nxv1p0(<vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef, <vscale x 1 x i64> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I32 = call <vscale x 16 x i32> @llvm.masked.gather.nxv16i32.nxv16p0(<vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef, <vscale x 16 x i32> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I32 = call <vscale x 8 x i32> @llvm.masked.gather.nxv8i32.nxv8p0(<vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef, <vscale x 8 x i32> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I32 = call <vscale x 4 x i32> @llvm.masked.gather.nxv4i32.nxv4p0(<vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef, <vscale x 4 x i32> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I32 = call <vscale x 2 x i32> @llvm.masked.gather.nxv2i32.nxv2p0(<vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef, <vscale x 2 x i32> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I32 = call <vscale x 1 x i32> @llvm.masked.gather.nxv1i32.nxv1p0(<vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef, <vscale x 1 x i32> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I16 = call <vscale x 32 x i16> @llvm.masked.gather.nxv32i16.nxv32p0(<vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef, <vscale x 32 x i16> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I16 = call <vscale x 16 x i16> @llvm.masked.gather.nxv16i16.nxv16p0(<vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef, <vscale x 16 x i16> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I16 = call <vscale x 8 x i16> @llvm.masked.gather.nxv8i16.nxv8p0(<vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef, <vscale x 8 x i16> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I16 = call <vscale x 4 x i16> @llvm.masked.gather.nxv4i16.nxv4p0(<vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef, <vscale x 4 x i16> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I16 = call <vscale x 2 x i16> @llvm.masked.gather.nxv2i16.nxv2p0(<vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef, <vscale x 2 x i16> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I16 = call <vscale x 1 x i16> @llvm.masked.gather.nxv1i16.nxv1p0(<vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef, <vscale x 1 x i16> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 128 for instruction: %V64I8 = call <vscale x 64 x i8> @llvm.masked.gather.nxv64i8.nxv64p0(<vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef, <vscale x 64 x i8> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 64 for instruction: %V32I8 = call <vscale x 32 x i8> @llvm.masked.gather.nxv32i8.nxv32p0(<vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef, <vscale x 32 x i8> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: %V16I8 = call <vscale x 16 x i8> @llvm.masked.gather.nxv16i8.nxv16p0(<vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef, <vscale x 16 x i8> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V8I8 = call <vscale x 8 x i8> @llvm.masked.gather.nxv8i8.nxv8p0(<vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef, <vscale x 8 x i8> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V4I8 = call <vscale x 4 x i8> @llvm.masked.gather.nxv4i8.nxv4p0(<vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef, <vscale x 4 x i8> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V2I8 = call <vscale x 2 x i8> @llvm.masked.gather.nxv2i8.nxv2p0(<vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef, <vscale x 2 x i8> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1I8 = call <vscale x 1 x i8> @llvm.masked.gather.nxv1i8.nxv1p0(<vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef, <vscale x 1 x i8> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; UNSUPPORTED-LABEL: 'masked_gather_aligned'
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: %V8F64 = call <vscale x 8 x double> @llvm.masked.gather.nxv8f64.nxv8p0(<vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef, <vscale x 8 x double> undef)
diff --git a/llvm/test/Analysis/CostModel/RISCV/scalable-scatter.ll b/llvm/test/Analysis/CostModel/RISCV/scalable-scatter.ll
index 204fee14b9ce1..6b8188a4c1a68 100644
--- a/llvm/test/Analysis/CostModel/RISCV/scalable-scatter.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/scalable-scatter.ll
@@ -7,73 +7,39 @@
; RUN: opt -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 < %s | FileCheck %s --check-prefixes=CHECK,UNSUPPORTED
define void @masked_scatter_aligned() {
-; RV32-LABEL: 'masked_scatter_aligned'
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 128 for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
-; RV32-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
-;
-; RV64-LABEL: 'masked_scatter_aligned'
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 128 for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
-; RV64-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
+; SUPPORTED-LABEL: 'masked_scatter_aligned'
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f64.nxv4p0(<vscale x 4 x double> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f64.nxv2p0(<vscale x 2 x double> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f64.nxv1p0(<vscale x 1 x double> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16f32.nxv16p0(<vscale x 16 x float> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8f32.nxv8p0(<vscale x 8 x float> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4f32.nxv4p0(<vscale x 4 x float> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2f32.nxv2p0(<vscale x 2 x float> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1f32.nxv1p0(<vscale x 1 x float> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i64.nxv8p0(<vscale x 8 x i64> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i64.nxv4p0(<vscale x 4 x i64> undef, <vscale x 4 x ptr> align 8 undef, <vscale x 4 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i64.nxv2p0(<vscale x 2 x i64> undef, <vscale x 2 x ptr> align 8 undef, <vscale x 2 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i64.nxv1p0(<vscale x 1 x i64> undef, <vscale x 1 x ptr> align 8 undef, <vscale x 1 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i32.nxv16p0(<vscale x 16 x i32> undef, <vscale x 16 x ptr> align 4 undef, <vscale x 16 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i32.nxv8p0(<vscale x 8 x i32> undef, <vscale x 8 x ptr> align 4 undef, <vscale x 8 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i32.nxv4p0(<vscale x 4 x i32> undef, <vscale x 4 x ptr> align 4 undef, <vscale x 4 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i32.nxv2p0(<vscale x 2 x i32> undef, <vscale x 2 x ptr> align 4 undef, <vscale x 2 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i32.nxv1p0(<vscale x 1 x i32> undef, <vscale x 1 x ptr> align 4 undef, <vscale x 1 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i16.nxv32p0(<vscale x 32 x i16> undef, <vscale x 32 x ptr> align 2 undef, <vscale x 32 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i16.nxv16p0(<vscale x 16 x i16> undef, <vscale x 16 x ptr> align 2 undef, <vscale x 16 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i16.nxv8p0(<vscale x 8 x i16> undef, <vscale x 8 x ptr> align 2 undef, <vscale x 8 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i16.nxv4p0(<vscale x 4 x i16> undef, <vscale x 4 x ptr> align 2 undef, <vscale x 4 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i16.nxv2p0(<vscale x 2 x i16> undef, <vscale x 2 x ptr> align 2 undef, <vscale x 2 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i16.nxv1p0(<vscale x 1 x i16> undef, <vscale x 1 x ptr> align 2 undef, <vscale x 1 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 128 for instruction: call void @llvm.masked.scatter.nxv64i8.nxv64p0(<vscale x 64 x i8> undef, <vscale x 64 x ptr> align 1 undef, <vscale x 64 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 64 for instruction: call void @llvm.masked.scatter.nxv32i8.nxv32p0(<vscale x 32 x i8> undef, <vscale x 32 x ptr> align 1 undef, <vscale x 32 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 32 for instruction: call void @llvm.masked.scatter.nxv16i8.nxv16p0(<vscale x 16 x i8> undef, <vscale x 16 x ptr> align 1 undef, <vscale x 16 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 16 for instruction: call void @llvm.masked.scatter.nxv8i8.nxv8p0(<vscale x 8 x i8> undef, <vscale x 8 x ptr> align 1 undef, <vscale x 8 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 8 for instruction: call void @llvm.masked.scatter.nxv4i8.nxv4p0(<vscale x 4 x i8> undef, <vscale x 4 x ptr> align 1 undef, <vscale x 4 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 4 for instruction: call void @llvm.masked.scatter.nxv2i8.nxv2p0(<vscale x 2 x i8> undef, <vscale x 2 x ptr> align 1 undef, <vscale x 2 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 2 for instruction: call void @llvm.masked.scatter.nxv1i8.nxv1p0(<vscale x 1 x i8> undef, <vscale x 1 x ptr> align 1 undef, <vscale x 1 x i1> undef)
+; SUPPORTED-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret void
;
; UNSUPPORTED-LABEL: 'masked_scatter_aligned'
; UNSUPPORTED-NEXT: Cost Model: Invalid cost for instruction: call void @llvm.masked.scatter.nxv8f64.nxv8p0(<vscale x 8 x double> undef, <vscale x 8 x ptr> align 8 undef, <vscale x 8 x i1> undef)
More information about the llvm-branch-commits
mailing list