[llvm] [PPC] generate stxvw4x/lxvw4x on P7 (PR #87049)
via llvm-commits
llvm-commits at lists.llvm.org
Fri Mar 29 03:33:28 PDT 2024
llvmbot wrote:
<!--LLVM PR SUMMARY COMMENT-->
@llvm/pr-subscribers-backend-powerpc
Author: Chen Zheng (chenzheng1030)
<details>
<summary>Changes</summary>
My understanding is that we should also use `stxvw4x`/`lxvw4x` on P7 even for unaligned load/stores.
P8 ISA and P7 ISA indicates that these two instructions are able to handle the unaligned addresses well.
---
Patch is 23.37 KiB, truncated to 20.00 KiB below, full version: https://github.com/llvm/llvm-project/pull/87049.diff
10 Files Affected:
- (modified) llvm/lib/Target/PowerPC/PPCISelLowering.cpp (+4-5)
- (modified) llvm/test/CodeGen/PowerPC/aix-vec-arg-spills-mir.ll (+27-34)
- (modified) llvm/test/CodeGen/PowerPC/aix-vec-arg-spills.ll (+49-55)
- (modified) llvm/test/CodeGen/PowerPC/memcpy-vec.ll (+4-7)
- (modified) llvm/test/CodeGen/PowerPC/unal-altivec-wint.ll (+3-3)
- (modified) llvm/test/CodeGen/PowerPC/unal-altivec2.ll (+1-1)
- (modified) llvm/test/CodeGen/PowerPC/unal-vec-ldst.ll (+12-48)
- (modified) llvm/test/CodeGen/PowerPC/unal-vec-negarith.ll (+1-1)
- (modified) llvm/test/CodeGen/PowerPC/unaligned.ll (+2-6)
- (modified) llvm/test/CodeGen/PowerPC/vsx.ll (+3-15)
``````````diff
diff --git a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
index 7436b202fba0d9..289e0bc29c4a55 100644
--- a/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
+++ b/llvm/lib/Target/PowerPC/PPCISelLowering.cpp
@@ -15965,8 +15965,8 @@ SDValue PPCTargetLowering::PerformDAGCombine(SDNode *N,
Align ABIAlignment = DAG.getDataLayout().getABITypeAlign(Ty);
if (LD->isUnindexed() && VT.isVector() &&
((Subtarget.hasAltivec() && ISD::isNON_EXTLoad(N) &&
- // P8 and later hardware should just use LOAD.
- !Subtarget.hasP8Vector() &&
+ // Hardware has VSX should just use LOAD.
+ !Subtarget.hasVSX() &&
(VT == MVT::v16i8 || VT == MVT::v8i16 || VT == MVT::v4i32 ||
VT == MVT::v4f32))) &&
LD->getAlign() < ABIAlignment) {
@@ -17250,8 +17250,7 @@ bool PPCTargetLowering::getTgtMemIntrinsic(IntrinsicInfo &Info,
EVT PPCTargetLowering::getOptimalMemOpType(
const MemOp &Op, const AttributeList &FuncAttributes) const {
if (getTargetMachine().getOptLevel() != CodeGenOptLevel::None) {
- // We should use Altivec/VSX loads and stores when available. For unaligned
- // addresses, unaligned VSX loads are only fast starting with the P8.
+ // We should use Altivec/VSX loads and stores when available.
if (Subtarget.hasAltivec() && Op.size() >= 16) {
if (Op.isMemset() && Subtarget.hasVSX()) {
uint64_t TailSize = Op.size() % 16;
@@ -17263,7 +17262,7 @@ EVT PPCTargetLowering::getOptimalMemOpType(
}
return MVT::v4i32;
}
- if (Op.isAligned(Align(16)) || Subtarget.hasP8Vector())
+ if (Op.isAligned(Align(16)) || Subtarget.hasVSX())
return MVT::v4i32;
}
}
diff --git a/llvm/test/CodeGen/PowerPC/aix-vec-arg-spills-mir.ll b/llvm/test/CodeGen/PowerPC/aix-vec-arg-spills-mir.ll
index 7c45958a1c2ff9..d927b9edb74d1d 100644
--- a/llvm/test/CodeGen/PowerPC/aix-vec-arg-spills-mir.ll
+++ b/llvm/test/CodeGen/PowerPC/aix-vec-arg-spills-mir.ll
@@ -1,4 +1,4 @@
-; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+; NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 4
; RUN: llc -verify-machineinstrs -mcpu=pwr7 -mattr=+altivec -vec-extabi \
; RUN: -stop-after=machine-cp -mtriple powerpc-ibm-aix-xcoff < %s | \
; RUN: FileCheck %s --check-prefix=MIR32
@@ -12,21 +12,16 @@
@__const.caller.t = private unnamed_addr constant %struct.Test { double 0.000000e+00, double 1.000000e+00, double 2.000000e+00, double 3.000000e+00 }, align 8
define double @caller() {
-
; MIR32-LABEL: name: caller
; MIR32: bb.0.entry:
- ; MIR32-NEXT: renamable $r3 = LI 0
- ; MIR32-NEXT: renamable $r4 = LIS 16392
- ; MIR32-NEXT: STW killed renamable $r4, 180, $r1 :: (store (s32) into unknown-address + 24)
- ; MIR32-NEXT: renamable $r4 = LIS 16384
- ; MIR32-NEXT: STW renamable $r3, 184, $r1 :: (store (s32) into unknown-address + 28)
- ; MIR32-NEXT: STW renamable $r3, 176, $r1 :: (store (s32) into unknown-address + 20)
- ; MIR32-NEXT: STW killed renamable $r4, 172, $r1 :: (store (s32) into unknown-address + 16)
- ; MIR32-NEXT: STW renamable $r3, 168, $r1 :: (store (s32) into unknown-address + 12)
- ; MIR32-NEXT: renamable $r4 = LIS 16368
- ; MIR32-NEXT: STW killed renamable $r4, 164, $r1 :: (store (s32) into unknown-address + 8)
- ; MIR32-NEXT: STW renamable $r3, 160, $r1 :: (store (s32) into unknown-address + 4)
- ; MIR32-NEXT: STW killed renamable $r3, 156, $r1 :: (store (s32))
+ ; MIR32-NEXT: renamable $r3 = LWZtoc @__const.caller.t, $r2 :: (load (s32) from got)
+ ; MIR32-NEXT: renamable $r4 = LI 16
+ ; MIR32-NEXT: renamable $vsl0 = LXVW4X renamable $r3, killed renamable $r4 :: (load (s128) from unknown-address + 16, align 8)
+ ; MIR32-NEXT: renamable $r4 = LI 172
+ ; MIR32-NEXT: STXVW4X killed renamable $vsl0, $r1, killed renamable $r4 :: (store (s128) into unknown-address + 16, align 4)
+ ; MIR32-NEXT: renamable $vsl0 = LXVW4X $zero, killed renamable $r3 :: (load (s128), align 8)
+ ; MIR32-NEXT: renamable $r3 = LI 156
+ ; MIR32-NEXT: STXVW4X killed renamable $vsl0, $r1, killed renamable $r3 :: (store (s128), align 4)
; MIR32-NEXT: ADJCALLSTACKDOWN 188, 0, implicit-def dead $r1, implicit $r1
; MIR32-NEXT: renamable $vsl0 = XXLXORz
; MIR32-NEXT: renamable $r3 = LI 136
@@ -78,32 +73,30 @@ define double @caller() {
;
; MIR64-LABEL: name: caller
; MIR64: bb.0.entry:
- ; MIR64-NEXT: renamable $x3 = LI8 2049
- ; MIR64-NEXT: renamable $x4 = LI8 1
- ; MIR64-NEXT: renamable $x3 = RLDIC killed renamable $x3, 51, 1
- ; MIR64-NEXT: STD killed renamable $x3, 216, $x1 :: (store (s64) into unknown-address + 24, align 4)
- ; MIR64-NEXT: renamable $x3 = LI8 1023
- ; MIR64-NEXT: renamable $x4 = RLDIC killed renamable $x4, 62, 1
- ; MIR64-NEXT: STD killed renamable $x4, 208, $x1 :: (store (s64) into unknown-address + 16, align 4)
- ; MIR64-NEXT: renamable $x4 = LI8 0
- ; MIR64-NEXT: STD renamable $x4, 192, $x1 :: (store (s64), align 4)
- ; MIR64-NEXT: renamable $x3 = RLDIC killed renamable $x3, 52, 2
- ; MIR64-NEXT: STD killed renamable $x3, 200, $x1 :: (store (s64) into unknown-address + 8, align 4)
+ ; MIR64-NEXT: renamable $x3 = LDtoc @__const.caller.t, $x2 :: (load (s64) from got)
+ ; MIR64-NEXT: renamable $x4 = LI8 16
+ ; MIR64-NEXT: renamable $vsl0 = LXVW4X renamable $x3, killed renamable $x4 :: (load (s128) from unknown-address + 16, align 8)
+ ; MIR64-NEXT: renamable $x4 = LI8 208
+ ; MIR64-NEXT: STXVW4X killed renamable $vsl0, $x1, killed renamable $x4 :: (store (s128) into unknown-address + 16, align 4)
+ ; MIR64-NEXT: renamable $vsl0 = LXVW4X $zero8, killed renamable $x3 :: (load (s128), align 8)
+ ; MIR64-NEXT: renamable $x3 = LI8 192
+ ; MIR64-NEXT: STXVW4X killed renamable $vsl0, $x1, killed renamable $x3 :: (store (s128), align 4)
; MIR64-NEXT: ADJCALLSTACKDOWN 224, 0, implicit-def dead $r1, implicit $r1
; MIR64-NEXT: renamable $vsl0 = XXLXORz
; MIR64-NEXT: renamable $x3 = LI8 160
+ ; MIR64-NEXT: renamable $x4 = LI8 144
; MIR64-NEXT: STXVW4X renamable $vsl0, $x1, killed renamable $x3 :: (store (s128), align 8)
- ; MIR64-NEXT: renamable $x3 = LI8 144
- ; MIR64-NEXT: STXVW4X renamable $vsl0, $x1, killed renamable $x3 :: (store (s128), align 8)
+ ; MIR64-NEXT: STXVW4X renamable $vsl0, $x1, killed renamable $x4 :: (store (s128), align 8)
; MIR64-NEXT: renamable $x3 = LI8 128
+ ; MIR64-NEXT: renamable $x4 = LDtocCPT %const.0, $x2 :: (load (s64) from got)
; MIR64-NEXT: STXVW4X killed renamable $vsl0, $x1, killed renamable $x3 :: (store (s128), align 8)
- ; MIR64-NEXT: renamable $x3 = LDtocCPT %const.0, $x2 :: (load (s64) from got)
- ; MIR64-NEXT: renamable $vsl0 = LXVD2X $zero8, killed renamable $x3 :: (load (s128) from constant-pool)
; MIR64-NEXT: renamable $x3 = LI8 80
+ ; MIR64-NEXT: renamable $vsl0 = LXVD2X $zero8, killed renamable $x4 :: (load (s128) from constant-pool)
+ ; MIR64-NEXT: renamable $x4 = LI8 512
; MIR64-NEXT: STXVD2X killed renamable $vsl0, $x1, killed renamable $x3 :: (store (s128))
- ; MIR64-NEXT: renamable $x3 = LI8 512
- ; MIR64-NEXT: STD killed renamable $x3, 184, $x1 :: (store (s64))
- ; MIR64-NEXT: STD killed renamable $x4, 176, $x1 :: (store (s64))
+ ; MIR64-NEXT: renamable $x3 = LI8 0
+ ; MIR64-NEXT: STD killed renamable $x4, 184, $x1 :: (store (s64))
+ ; MIR64-NEXT: STD killed renamable $x3, 176, $x1 :: (store (s64))
; MIR64-NEXT: $f1 = XXLXORdpz
; MIR64-NEXT: $f2 = XXLXORdpz
; MIR64-NEXT: $v2 = XXLXORz
@@ -112,8 +105,8 @@ define double @caller() {
; MIR64-NEXT: $v5 = XXLXORz
; MIR64-NEXT: $v6 = XXLXORz
; MIR64-NEXT: $x3 = LI8 128
- ; MIR64-NEXT: $x4 = LI8 256
; MIR64-NEXT: $v7 = XXLXORz
+ ; MIR64-NEXT: $x4 = LI8 256
; MIR64-NEXT: $v8 = XXLXORz
; MIR64-NEXT: $v9 = XXLXORz
; MIR64-NEXT: $v10 = XXLXORz
@@ -136,7 +129,7 @@ define double @caller() {
; MIR64-NEXT: BLR8 implicit $lr8, implicit $rm, implicit $f1
entry:
%call = tail call double @callee(i32 signext 128, i32 signext 256, double 0.000000e+00, double 0.000000e+00, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 0.000000e+00, double 0.000000e+00>, <2 x double> <double 2.400000e+01, double 2.500000e+01>, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, double 0.000000e+00, i32 signext 512, ptr nonnull byval(%struct.Test) align 4 @__const.caller.t)
- ret double %call
+ ret double %call
}
declare double @callee(i32 signext, i32 signext, double, double, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, <2 x double>, double, double, double, double, double, double, double, double, double, double, double, i32 signext, ptr byval(%struct.Test) align 8)
diff --git a/llvm/test/CodeGen/PowerPC/aix-vec-arg-spills.ll b/llvm/test/CodeGen/PowerPC/aix-vec-arg-spills.ll
index 66f88b4e3d5ab3..91e7d4094fc344 100644
--- a/llvm/test/CodeGen/PowerPC/aix-vec-arg-spills.ll
+++ b/llvm/test/CodeGen/PowerPC/aix-vec-arg-spills.ll
@@ -15,56 +15,52 @@ define double @caller() {
; 32BIT: # %bb.0: # %entry
; 32BIT-NEXT: mflr 0
; 32BIT-NEXT: stwu 1, -192(1)
-; 32BIT-NEXT: lis 4, 16392
+; 32BIT-NEXT: lwz 3, L..C0(2) # @__const.caller.t
+; 32BIT-NEXT: li 4, 16
; 32BIT-NEXT: stw 0, 200(1)
-; 32BIT-NEXT: li 3, 0
-; 32BIT-NEXT: xxlxor 0, 0, 0
; 32BIT-NEXT: xxlxor 1, 1, 1
-; 32BIT-NEXT: stw 4, 180(1)
-; 32BIT-NEXT: lis 4, 16384
-; 32BIT-NEXT: stw 3, 184(1)
-; 32BIT-NEXT: stw 3, 176(1)
-; 32BIT-NEXT: stw 4, 172(1)
-; 32BIT-NEXT: lis 4, 16368
-; 32BIT-NEXT: stw 3, 168(1)
-; 32BIT-NEXT: stw 3, 160(1)
-; 32BIT-NEXT: stw 4, 164(1)
-; 32BIT-NEXT: stw 3, 156(1)
-; 32BIT-NEXT: li 3, 136
-; 32BIT-NEXT: li 4, 120
; 32BIT-NEXT: xxlxor 2, 2, 2
-; 32BIT-NEXT: stxvw4x 0, 1, 3
-; 32BIT-NEXT: li 3, 104
-; 32BIT-NEXT: stxvw4x 0, 1, 4
-; 32BIT-NEXT: li 4, 88
-; 32BIT-NEXT: stxvw4x 0, 1, 3
-; 32BIT-NEXT: stxvw4x 0, 1, 4
-; 32BIT-NEXT: lwz 4, L..C0(2) # %const.0
-; 32BIT-NEXT: li 3, 72
-; 32BIT-NEXT: stxvw4x 0, 1, 3
-; 32BIT-NEXT: li 3, 48
+; 32BIT-NEXT: lxvw4x 0, 3, 4
+; 32BIT-NEXT: li 4, 172
; 32BIT-NEXT: xxlxor 34, 34, 34
; 32BIT-NEXT: xxlxor 35, 35, 35
-; 32BIT-NEXT: lxvd2x 0, 0, 4
-; 32BIT-NEXT: li 4, 512
+; 32BIT-NEXT: stxvw4x 0, 1, 4
+; 32BIT-NEXT: li 4, 120
; 32BIT-NEXT: xxlxor 36, 36, 36
; 32BIT-NEXT: xxlxor 37, 37, 37
; 32BIT-NEXT: xxlxor 38, 38, 38
+; 32BIT-NEXT: lxvw4x 0, 0, 3
+; 32BIT-NEXT: li 3, 156
; 32BIT-NEXT: xxlxor 39, 39, 39
+; 32BIT-NEXT: stxvw4x 0, 1, 3
+; 32BIT-NEXT: xxlxor 0, 0, 0
+; 32BIT-NEXT: li 3, 136
; 32BIT-NEXT: xxlxor 40, 40, 40
; 32BIT-NEXT: xxlxor 41, 41, 41
+; 32BIT-NEXT: stxvw4x 0, 1, 3
+; 32BIT-NEXT: li 3, 104
+; 32BIT-NEXT: stxvw4x 0, 1, 4
+; 32BIT-NEXT: li 4, 88
; 32BIT-NEXT: xxlxor 42, 42, 42
-; 32BIT-NEXT: stxvd2x 0, 1, 3
-; 32BIT-NEXT: stw 4, 152(1)
-; 32BIT-NEXT: li 3, 128
-; 32BIT-NEXT: li 4, 256
; 32BIT-NEXT: xxlxor 43, 43, 43
; 32BIT-NEXT: xxlxor 44, 44, 44
+; 32BIT-NEXT: stxvw4x 0, 1, 3
+; 32BIT-NEXT: stxvw4x 0, 1, 4
+; 32BIT-NEXT: lwz 4, L..C1(2) # %const.0
+; 32BIT-NEXT: li 3, 72
; 32BIT-NEXT: xxlxor 45, 45, 45
+; 32BIT-NEXT: stxvw4x 0, 1, 3
+; 32BIT-NEXT: li 3, 48
; 32BIT-NEXT: xxlxor 3, 3, 3
; 32BIT-NEXT: xxlxor 4, 4, 4
+; 32BIT-NEXT: lxvd2x 0, 0, 4
+; 32BIT-NEXT: li 4, 512
; 32BIT-NEXT: xxlxor 5, 5, 5
; 32BIT-NEXT: xxlxor 6, 6, 6
+; 32BIT-NEXT: stxvd2x 0, 1, 3
+; 32BIT-NEXT: stw 4, 152(1)
+; 32BIT-NEXT: li 3, 128
+; 32BIT-NEXT: li 4, 256
; 32BIT-NEXT: xxlxor 7, 7, 7
; 32BIT-NEXT: xxlxor 8, 8, 8
; 32BIT-NEXT: xxlxor 9, 9, 9
@@ -83,54 +79,52 @@ define double @caller() {
; 64BIT: # %bb.0: # %entry
; 64BIT-NEXT: mflr 0
; 64BIT-NEXT: stdu 1, -224(1)
-; 64BIT-NEXT: li 3, 2049
+; 64BIT-NEXT: ld 3, L..C0(2) # @__const.caller.t
+; 64BIT-NEXT: li 4, 16
; 64BIT-NEXT: std 0, 240(1)
-; 64BIT-NEXT: li 4, 1
-; 64BIT-NEXT: xxlxor 0, 0, 0
; 64BIT-NEXT: xxlxor 1, 1, 1
-; 64BIT-NEXT: rldic 3, 3, 51, 1
-; 64BIT-NEXT: rldic 4, 4, 62, 1
; 64BIT-NEXT: xxlxor 2, 2, 2
+; 64BIT-NEXT: lxvw4x 0, 3, 4
+; 64BIT-NEXT: li 4, 208
; 64BIT-NEXT: xxlxor 34, 34, 34
-; 64BIT-NEXT: std 3, 216(1)
-; 64BIT-NEXT: li 3, 1023
-; 64BIT-NEXT: std 4, 208(1)
-; 64BIT-NEXT: li 4, 0
; 64BIT-NEXT: xxlxor 35, 35, 35
+; 64BIT-NEXT: stxvw4x 0, 1, 4
+; 64BIT-NEXT: li 4, 144
; 64BIT-NEXT: xxlxor 36, 36, 36
-; 64BIT-NEXT: rldic 3, 3, 52, 2
-; 64BIT-NEXT: std 4, 192(1)
; 64BIT-NEXT: xxlxor 37, 37, 37
; 64BIT-NEXT: xxlxor 38, 38, 38
+; 64BIT-NEXT: lxvw4x 0, 0, 3
+; 64BIT-NEXT: li 3, 192
; 64BIT-NEXT: xxlxor 39, 39, 39
-; 64BIT-NEXT: std 3, 200(1)
+; 64BIT-NEXT: stxvw4x 0, 1, 3
+; 64BIT-NEXT: xxlxor 0, 0, 0
; 64BIT-NEXT: li 3, 160
; 64BIT-NEXT: xxlxor 40, 40, 40
-; 64BIT-NEXT: stxvw4x 0, 1, 3
-; 64BIT-NEXT: li 3, 144
; 64BIT-NEXT: xxlxor 41, 41, 41
-; 64BIT-NEXT: xxlxor 42, 42, 42
; 64BIT-NEXT: stxvw4x 0, 1, 3
+; 64BIT-NEXT: stxvw4x 0, 1, 4
+; 64BIT-NEXT: ld 4, L..C1(2) # %const.0
; 64BIT-NEXT: li 3, 128
+; 64BIT-NEXT: xxlxor 42, 42, 42
; 64BIT-NEXT: xxlxor 43, 43, 43
; 64BIT-NEXT: stxvw4x 0, 1, 3
-; 64BIT-NEXT: ld 3, L..C0(2) # %const.0
+; 64BIT-NEXT: li 3, 80
; 64BIT-NEXT: xxlxor 44, 44, 44
; 64BIT-NEXT: xxlxor 45, 45, 45
-; 64BIT-NEXT: lxvd2x 0, 0, 3
-; 64BIT-NEXT: li 3, 80
+; 64BIT-NEXT: lxvd2x 0, 0, 4
+; 64BIT-NEXT: li 4, 512
; 64BIT-NEXT: xxlxor 3, 3, 3
-; 64BIT-NEXT: xxlxor 4, 4, 4
-; 64BIT-NEXT: xxlxor 5, 5, 5
; 64BIT-NEXT: stxvd2x 0, 1, 3
-; 64BIT-NEXT: li 3, 512
-; 64BIT-NEXT: std 4, 176(1)
+; 64BIT-NEXT: li 3, 0
+; 64BIT-NEXT: std 4, 184(1)
; 64BIT-NEXT: li 4, 256
+; 64BIT-NEXT: xxlxor 4, 4, 4
+; 64BIT-NEXT: xxlxor 5, 5, 5
; 64BIT-NEXT: xxlxor 6, 6, 6
+; 64BIT-NEXT: std 3, 176(1)
+; 64BIT-NEXT: li 3, 128
; 64BIT-NEXT: xxlxor 7, 7, 7
; 64BIT-NEXT: xxlxor 8, 8, 8
-; 64BIT-NEXT: std 3, 184(1)
-; 64BIT-NEXT: li 3, 128
; 64BIT-NEXT: xxlxor 9, 9, 9
; 64BIT-NEXT: xxlxor 10, 10, 10
; 64BIT-NEXT: xxlxor 11, 11, 11
diff --git a/llvm/test/CodeGen/PowerPC/memcpy-vec.ll b/llvm/test/CodeGen/PowerPC/memcpy-vec.ll
index d636921eea3e51..34a7af4bc45916 100644
--- a/llvm/test/CodeGen/PowerPC/memcpy-vec.ll
+++ b/llvm/test/CodeGen/PowerPC/memcpy-vec.ll
@@ -10,12 +10,8 @@ entry:
ret void
; PWR7-LABEL: @foo1
-; PWR7-NOT: bl memcpy
-; PWR7-DAG: li [[OFFSET:[0-9]+]], 16
-; PWR7-DAG: lxvd2x [[TMP0:[0-9]+]], 4, [[OFFSET]]
-; PWR7-DAG: stxvd2x [[TMP0]], 3, [[OFFSET]]
-; PWR7-DAG: lxvd2x [[TMP1:[0-9]+]], 0, 4
-; PWR7-DAG: stxvd2x [[TMP1]], 0, 3
+; PWR7: lxvw4x
+; PWR7: stxvw4x
; PWR7: blr
; PWR8-LABEL: @foo1
@@ -34,7 +30,8 @@ entry:
ret void
; PWR7-LABEL: @foo2
-; PWR7: bl memcpy
+; PWR7: lxvw4x
+; PWR7: stxvw4x
; PWR7: blr
; PWR8-LABEL: @foo2
diff --git a/llvm/test/CodeGen/PowerPC/unal-altivec-wint.ll b/llvm/test/CodeGen/PowerPC/unal-altivec-wint.ll
index d6244cd828e5a6..a590f54b5a6765 100644
--- a/llvm/test/CodeGen/PowerPC/unal-altivec-wint.ll
+++ b/llvm/test/CodeGen/PowerPC/unal-altivec-wint.ll
@@ -17,7 +17,7 @@ entry:
; CHECK-LABEL: @test1
; CHECK: li [[REG:[0-9]+]], 16
; CHECK-NOT: li {{[0-9]+}}, 15
-; CHECK-DAG: lvx {{[0-9]+}}, 0, 3
+; CHECK-DAG: lxvw4x {{[0-9]+}}, 0, 3
; CHECK-DAG: lvx {{[0-9]+}}, 3, [[REG]]
; CHECK: blr
}
@@ -36,8 +36,8 @@ entry:
; CHECK-LABEL: @test2
; CHECK: li [[REG:[0-9]+]], 16
; CHECK-NOT: li {{[0-9]+}}, 15
-; CHECK-DAG: lvx {{[0-9]+}}, 0, 3
-; CHECK-DAG: lvx {{[0-9]+}}, 3, [[REG]]
+; CHECK-DAG: stvx 2, 3, [[REG]]
+; CHECK-DAG: lxvw4x {{[0-9]+}}, 0, 3
; CHECK: blr
}
diff --git a/llvm/test/CodeGen/PowerPC/unal-altivec2.ll b/llvm/test/CodeGen/PowerPC/unal-altivec2.ll
index fafcab8468eb4d..39a82fe0a0977c 100644
--- a/llvm/test/CodeGen/PowerPC/unal-altivec2.ll
+++ b/llvm/test/CodeGen/PowerPC/unal-altivec2.ll
@@ -1,4 +1,4 @@
-; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7 < %s | FileCheck %s
+; RUN: llc -verify-machineinstrs -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr6 < %s | FileCheck %s
target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-f128:128:128-v128:128:128-n32:64"
target triple = "powerpc64-unknown-linux-gnu"
diff --git a/llvm/test/CodeGen/PowerPC/unal-vec-ldst.ll b/llvm/test/CodeGen/PowerPC/unal-vec-ldst.ll
index b0ed395fc3a190..c2e20149cb9fee 100644
--- a/llvm/test/CodeGen/PowerPC/unal-vec-ldst.ll
+++ b/llvm/test/CodeGen/PowerPC/unal-vec-ldst.ll
@@ -5,11 +5,7 @@
define <16 x i8> @test_l_v16i8(ptr %p) #0 {
; CHECK-LABEL: test_l_v16i8:
; CHECK: # %bb.0: # %entry
-; CHECK-NEXT: li 4, 15
-; CHECK-NEXT: lvsl 3, 0, 3
-; CHECK-NEXT: lvx 2, 3, 4
-; CHECK-NEXT: lvx 4, 0, 3
-; CHECK-NEXT: vperm 2, 4, 2, 3
+; CHECK-NEXT: lxvw4x 34, 0, 3
; CHECK-NEXT: blr
entry:
%r = load <16 x i8>, ptr %p, align 1
@@ -20,14 +16,9 @@ entry:
define <32 x i8> @test_l_v32i8(ptr %p) #0 {
; CHECK-LABEL: test_l_v32i8:
; CHECK: # %bb.0: # %entry
-; CHECK-NEXT: li 4, 31
-; CHECK-NEXT: lvsl 4, 0, 3
-; CHECK-NEXT: lvx 2, 3, 4
; CHECK-NEXT: li 4, 16
-; CHECK-NEXT: lvx 5, 3, 4
-; CHECK-NEXT: vperm 3, 5, 2, 4
-; CHECK-NEXT: lvx 2, 0, 3
-; CHECK-NEXT: vperm 2, 2, 5, 4
+; CHECK-NEXT: lxvw4x 34, 0, 3
+; CHECK-NEXT: lxvw4x 35, 3, 4
; CHECK-NEXT: blr
entry:
%r = load <32 x i8>, ptr %p, align 1
@@ -38,11 +29,7 @@ entry:
define <8 x i16> @test_l_v8i16(ptr %p) #0 {
; CHECK-LABEL: test_l_v8i16:
; CHECK: # %bb.0: # %entry
-; CHECK-NEXT: li 4, 15
-; CHECK-NEXT: lvsl 3, 0, 3
-; CHECK-NEXT: lvx 2, 3, 4
-; CHECK-NEXT: lvx 4, 0, 3
-; CHECK-NEXT: vperm 2, 4, 2, 3
+; CHECK-NEXT: lxvw4x 34, 0, 3
; CHECK-NEXT: blr
entry:
%r = load <8 x i16>, ptr %p, align 2
@@ -53,14 +40,9 @@ entry:
define <16 x i16> @test_l_v16i16(ptr %p) #0 {
; CHECK-LABEL: test_l_v16i16:
; CHECK: # %bb.0: # %entry
-; CHECK-NEXT: li 4, 31
-; CHECK-NEXT: lvsl 4, 0, 3
-; CHECK-NEXT: lvx 2, 3, 4
; CHECK-NEXT: li 4, 16
-; CHECK-NEXT: lvx 5, 3, 4
-; CHECK-NEXT: vperm 3, 5, 2, 4
-; CHECK-NEXT: lvx 2, 0, 3
-; CHECK-NEXT: vperm 2, 2, 5, 4
+; CHECK-NEXT: lxvw4x 34, 0, 3
+; CHECK-NEXT: lxvw4x 35, 3, 4
; CHECK-NEXT: blr
entry:
%r = load <16 x i16>, ptr %p, align 2
@@ -71,11 +53,7 @@ entry:
define <4 x i32> @test_l_v4i32(ptr %p) #0 {
; CHECK-LABEL: test_l_v4i32:
; CHECK: # %bb.0: # %entry
-; CHECK-NEXT: li 4, 15
-; CHECK-NEXT: lvsl 3, 0, 3
-; CHECK-NEXT: lvx 2, 3, 4
-; CHECK-NEXT: lvx 4, 0, 3
-; CHECK-NEXT: vperm 2, 4, 2, 3
+; CHECK-NEXT: lxvw4x 34, 0, 3
; CHECK-NEXT: blr
entry:
%r = load <4 x i32>, ptr %p, align 4
@@ -86,14 +64,9 @@ entry:
define <8 x i32> @test_l_v8i32(ptr %p) #0 {
; CHECK-LABEL: test_l_v8i32:
; CHECK: # %bb.0: # %entry
-; CHECK-NEXT: li 4, 31
-; CHECK-NEXT: lvsl 4, 0, 3
-; CHECK-NEXT: lvx 2, 3, 4
; CHECK-NEXT: li 4, 16
-; CHECK-NEXT: lvx 5, 3, 4
-; CHECK-NEXT: vperm 3, 5, 2, 4
-; CHECK-NEXT: lvx 2, 0, 3
-; CHECK-NEXT: vperm 2, 2, 5, 4
+; CHECK-NEXT: lxvw4x 34, 0, 3
+; CHECK-NEXT: lxvw4x 35, 3, 4
; CHECK-NEXT: blr
entry:
%r = load <8 x i32>, ptr %p, align 4
@@ -128,11 +101,7 @@ entry:
def...
[truncated]
``````````
</details>
https://github.com/llvm/llvm-project/pull/87049
More information about the llvm-commits
mailing list