[AArch64] Update cost model for extracting halves from 128+ bit vectors #155601
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Previously, only 128-bit "NEON" vectors were given sensible costs. Cores with vscale>1 can use SVE's EXT instruction to perform a fixed-length subvector extract.
llvmbot
added
backend:AArch64
llvm:analysis
|
@llvm/pr-subscribers-backend-aarch64 @llvm/pr-subscribers-llvm-analysis Author: Gaëtan Bossu (gbossu) ChangesPreviously, only 128-bit "NEON" vectors were given sensible costs. This is a follow-up from the codegen patches at #152554. They show that with the help of MOVPRFX, we can do subvector extract with roughly one instruction. We now at least give sensible costs for extracting 128-bit halves from a 256-bit vector. Patch is 32.34 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/155601.diff 2 Files Affected:
diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
index 490f6391c15a0..120f29a1aa22c 100644
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -5750,11 +5750,13 @@ AArch64TTIImpl::getShuffleCost(TTI::ShuffleKind Kind, VectorType *DstTy,
Kind = improveShuffleKindFromMask(Kind, Mask, SrcTy, Index, SubTp);
bool IsExtractSubvector = Kind == TTI::SK_ExtractSubvector;
- // A subvector extract can be implemented with an ext (or trivial extract, if
- // from lane 0). This currently only handles low or high extracts to prevent
- // SLP vectorizer regressions.
+ // A subvector extract can be implemented with a NEON/SVE ext (or trivial
+ // extract, if from lane 0). This currently only handles low or high extracts
+ // to prevent SLP vectorizer regressions.
+ // Note that SVE's ext instruciton is destructive, but it can be fused with
+ // a movprfx to act like a constructive instruction.
if (IsExtractSubvector && LT.second.isFixedLengthVector()) {
- if (LT.second.is128BitVector() &&
+ if (LT.second.getFixedSizeInBits() >= AArch64::SVEBitsPerBlock &&
cast<FixedVectorType>(SubTp)->getNumElements() ==
LT.second.getVectorNumElements() / 2) {
if (Index == 0)
diff --git a/llvm/test/Analysis/CostModel/AArch64/shuffle-extract.ll b/llvm/test/Analysis/CostModel/AArch64/shuffle-extract.ll
index 867c35ab79446..8199f26b9c11c 100644
--- a/llvm/test/Analysis/CostModel/AArch64/shuffle-extract.ll
+++ b/llvm/test/Analysis/CostModel/AArch64/shuffle-extract.ll
@@ -1,10 +1,10 @@
; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py
-; RUN: opt < %s -mtriple=aarch64--linux-gnu -passes="print<cost-model>" -cost-kind=all 2>&1 -disable-output | FileCheck %s
+; RUN: opt < %s -mtriple=aarch64 -mattr=+sve -passes="print<cost-model>" -cost-kind=all 2>&1 -disable-output | FileCheck %s
target datalayout = "e-m:e-i8:8:32-i16:16:32-i64:64-i128:128-n32:64-S128"
-define void @extract_half() {
-; CHECK-LABEL: 'extract_half'
+define void @extract_half_vscale_any() {
+; CHECK-LABEL: 'extract_half_vscale_any'
; CHECK-NEXT: Cost Model: Found costs of 0 for: %v2i8_lo = shufflevector <2 x i8> poison, <2 x i8> poison, <1 x i32> zeroinitializer
; CHECK-NEXT: Cost Model: Found costs of 1 for: %v2i8_hi = shufflevector <2 x i8> poison, <2 x i8> poison, <1 x i32> <i32 1>
; CHECK-NEXT: Cost Model: Found costs of 0 for: %v4i8_lo = shufflevector <4 x i8> poison, <4 x i8> poison, <2 x i32> <i32 0, i32 1>
@@ -16,6 +16,9 @@ define void @extract_half() {
; CHECK-NEXT: Cost Model: Found costs of 0 for: %v16i8_lo = shufflevector <16 x i8> poison, <16 x i8> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
; CHECK-NEXT: Cost Model: Found costs of 8 for: %v16i8_mi = shufflevector <16 x i8> poison, <16 x i8> poison, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
; CHECK-NEXT: Cost Model: Found costs of 1 for: %v16i8_hi = shufflevector <16 x i8> poison, <16 x i8> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v32i8_lo = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of 16 for: %v32i8_mi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; CHECK-NEXT: Cost Model: Found costs of 16 for: %v32i8_hi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
; CHECK-NEXT: Cost Model: Found costs of 0 for: %v2i16_lo = shufflevector <2 x i16> poison, <2 x i16> poison, <1 x i32> zeroinitializer
; CHECK-NEXT: Cost Model: Found costs of 1 for: %v2i16_hi = shufflevector <2 x i16> poison, <2 x i16> poison, <1 x i32> <i32 1>
; CHECK-NEXT: Cost Model: Found costs of 0 for: %v4i16_lo = shufflevector <4 x i16> poison, <4 x i16> poison, <2 x i32> <i32 0, i32 1>
@@ -59,6 +62,9 @@ define void @extract_half() {
%v16i8_lo = shufflevector <16 x i8> poison, <16 x i8> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%v16i8_mi = shufflevector <16 x i8> poison, <16 x i8> poison, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
%v16i8_hi = shufflevector <16 x i8> poison, <16 x i8> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %v32i8_lo = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %v32i8_mi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+ %v32i8_hi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
%v2i16_lo = shufflevector <2 x i16> poison, <2 x i16> poison, <1 x i32> <i32 0>
%v2i16_hi = shufflevector <2 x i16> poison, <2 x i16> poison, <1 x i32> <i32 1>
@@ -96,8 +102,129 @@ define void @extract_half() {
ret void
}
-define void @extract_qtr() {
-; CHECK-LABEL: 'extract_qtr'
+; Similar as above, but for a 256-bit SVE core.
+; Vectors smaller than 256-bit aren't tested again.
+define void @extract_half_vscale2() vscale_range(2,2) {
+; CHECK-LABEL: 'extract_half_vscale2'
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v32i8_lo = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of RThru:64 CodeSize:32 Lat:64 SizeLat:64 for: %v32i8_mi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v32i8_hi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v64i8_lo = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; CHECK-NEXT: Cost Model: Found costs of RThru:128 CodeSize:64 Lat:128 SizeLat:128 for: %v64i8_mi = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
+; CHECK-NEXT: Cost Model: Found costs of RThru:128 CodeSize:64 Lat:128 SizeLat:128 for: %v64i8_hi = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v16i16_lo = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT: Cost Model: Found costs of RThru:32 CodeSize:16 Lat:32 SizeLat:32 for: %v16i16_mi = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v16i16_hi = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v32i16_lo = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of RThru:64 CodeSize:32 Lat:64 SizeLat:64 for: %v32i16_mi = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: Cost Model: Found costs of RThru:64 CodeSize:32 Lat:64 SizeLat:64 for: %v32i16_hi = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v8i32_lo = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT: Cost Model: Found costs of RThru:16 CodeSize:8 Lat:16 SizeLat:16 for: %v8i32_mi = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v8i32_hi = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v16i32_lo = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT: Cost Model: Found costs of RThru:32 CodeSize:16 Lat:32 SizeLat:32 for: %v16i32_mi = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
+; CHECK-NEXT: Cost Model: Found costs of RThru:32 CodeSize:16 Lat:32 SizeLat:32 for: %v16i32_hi = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v4i64_lo = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT: Cost Model: Found costs of RThru:8 CodeSize:4 Lat:8 SizeLat:8 for: %v4i64_mi = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 1, i32 2>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v4i64_hi = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v8i64_lo = shufflevector <8 x i64> poison, <8 x i64> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT: Cost Model: Found costs of RThru:16 CodeSize:8 Lat:16 SizeLat:16 for: %v8i64_mi = shufflevector <8 x i64> poison, <8 x i64> poison, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; CHECK-NEXT: Cost Model: Found costs of RThru:16 CodeSize:8 Lat:16 SizeLat:16 for: %v8i64_hi = shufflevector <8 x i64> poison, <8 x i64> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT: Cost Model: Found costs of RThru:0 CodeSize:1 Lat:1 SizeLat:1 for: ret void
+;
+ %v32i8_lo = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %v32i8_mi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+ %v32i8_hi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+ %v64i8_lo = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+ %v64i8_mi = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
+ %v64i8_hi = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+
+ %v16i16_lo = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+ %v16i16_mi = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
+ %v16i16_hi = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %v32i16_lo = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+ %v32i16_mi = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19>
+ %v32i16_hi = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+
+ %v8i32_lo = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+ %v8i32_mi = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+ %v8i32_hi = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+ %v16i32_lo = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+ %v16i32_mi = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
+ %v16i32_hi = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+
+ %v4i64_lo = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 0, i32 1>
+ %v4i64_mi = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 1, i32 2>
+ %v4i64_hi = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 2, i32 3>
+ %v8i64_lo = shufflevector <8 x i64> poison, <8 x i64> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+ %v8i64_mi = shufflevector <8 x i64> poison, <8 x i64> poison, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+ %v8i64_hi = shufflevector <8 x i64> poison, <8 x i64> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+
+ ret void
+}
+
+; Similar as above, but for a 512-bit SVE core.
+define void @extract_half_vscale4() vscale_range(4,4) {
+; CHECK-LABEL: 'extract_half_vscale4'
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v32i8_lo = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of RThru:64 CodeSize:32 Lat:64 SizeLat:64 for: %v32i8_mi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v32i8_hi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v64i8_lo = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; CHECK-NEXT: Cost Model: Found costs of RThru:128 CodeSize:64 Lat:128 SizeLat:128 for: %v64i8_mi = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v64i8_hi = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v16i16_lo = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT: Cost Model: Found costs of RThru:32 CodeSize:16 Lat:32 SizeLat:32 for: %v16i16_mi = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v16i16_hi = shufflevector <16 x i16> poison, <16 x i16> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v32i16_lo = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of RThru:64 CodeSize:32 Lat:64 SizeLat:64 for: %v32i16_mi = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v32i16_hi = shufflevector <32 x i16> poison, <32 x i16> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v8i32_lo = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
+; CHECK-NEXT: Cost Model: Found costs of RThru:16 CodeSize:8 Lat:16 SizeLat:16 for: %v8i32_mi = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 2, i32 3, i32 4, i32 5>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v8i32_hi = shufflevector <8 x i32> poison, <8 x i32> poison, <4 x i32> <i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v16i32_lo = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
+; CHECK-NEXT: Cost Model: Found costs of RThru:32 CodeSize:16 Lat:32 SizeLat:32 for: %v16i32_mi = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v16i32_hi = shufflevector <16 x i32> poison, <16 x i32> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
+; CHECK-NEXT: Cost Model: Found costs of 0 for: %v4i64_lo = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 0, i32 1>
+; CHECK-NEXT: Cost Model: Found costs of RThru:8 CodeSize:4 Lat:8 SizeLat:8 for: %v4i64_mi = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 1, i32 2>
+; CHECK-NEXT: Cost Model: Found costs of 1 for: %v4i64_hi = shufflevector <4 x i64> poison, <4 x i64> poison, <2 x i32> <i32 2, i32 3>
+; CHECK-NEXT: Cost Model: Found cos...
[truncated]
|
SamTebbs33
reviewed
Aug 27, 2025
I'll create a new one for VLS SVE.
sdesmalen-arm
approved these changes
Sep 1, 2025
| ; VSCALE-4-LABEL: 'extract_half_hi' | ||
| ; VSCALE-4-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v16i8_hi = shufflevector <16 x i8> poison, <16 x i8> poison, <8 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15> | ||
| ; VSCALE-4-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v32i8_hi = shufflevector <32 x i8> poison, <32 x i8> poison, <16 x i32> <i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31> | ||
| ; VSCALE-4-NEXT: Cost Model: Found an estimated cost of 1 for instruction: %v64i8_hi = shufflevector <64 x i8> poison, <64 x i8> poison, <32 x i32> <i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63> |
There was a problem hiding this comment.
Just an observation that this could also be a cost of 0, because if the extracted vector is a multiple of the legalised input type, and the index is a multiple of the number of elements of the legalised input type, then it would just be selecting the high registers (rather than requiring an ext).
paulwalker-arm
approved these changes
Sep 3, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Previously, only 128-bit "NEON" vectors were given sensible costs.
Cores with vscale>1 can use SVE's EXT instruction to perform a
fixed-length subvector extract.
This is a follow-up from the codegen patches at #152554. They show that with the help of MOVPRFX, we can do subvector extract with roughly one instruction. We now at least give sensible costs for extracting 128-bit halves from a 256-bit vector.