AVX-512: Removed "z" suffix from AVX-512 instructions, since it is in…

…compatible with GCC. I moved a test from avx512-vbroadcast-crash.ll to avx512-vbroadcast.ll I defined HasAVX512 predicate as AssemblerPredicate. It means that you should invoke llvm-mc with "-mcpu=knl" to get encoding for AVX-512 instructions. I need this to let AsmMatcher to set different encoding for AVX and AVX-512 instructions that have the same mnemonic and operands (all scalar instructions). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@197041 91177308-0d34-0410-b5e6-96231b3b80d8
tzlaine · Dec 11, 2013 · 3bf51cf · 3bf51cf
1 parent e54c106
commit 3bf51cf
Show file tree

Hide file tree

Showing 11 changed files with 174 additions and 171 deletions.
diff --git a/lib/Target/X86/X86InstrAVX512.td b/lib/Target/X86/X86InstrAVX512.td
diff --git a/lib/Target/X86/X86InstrInfo.td b/lib/Target/X86/X86InstrInfo.td
@@ -657,7 +657,8 @@ def HasSSE4A     : Predicate<"Subtarget->hasSSE4A()">;
 def HasAVX       : Predicate<"Subtarget->hasAVX()">;
 def HasAVX2      : Predicate<"Subtarget->hasAVX2()">;
 def HasAVX1Only  : Predicate<"Subtarget->hasAVX() && !Subtarget->hasAVX2()">;
-def HasAVX512    : Predicate<"Subtarget->hasAVX512()">;
+def HasAVX512    : Predicate<"Subtarget->hasAVX512()">,
+                     AssemblerPredicate<"FeatureAVX512", "AVX-512 ISA">;
 def UseAVX       : Predicate<"Subtarget->hasAVX() && !Subtarget->hasAVX512()">;
 def UseAVX2      : Predicate<"Subtarget->hasAVX2() && !Subtarget->hasAVX512()">;
 def NoAVX512       : Predicate<"!Subtarget->hasAVX512()">;

diff --git a/test/CodeGen/X86/avx512-arith.ll b/test/CodeGen/X86/avx512-arith.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl --show-mc-encoding| FileCheck %s
 
 ; CHECK-LABEL: addpd512
 ; CHECK: vaddpd
@@ -196,7 +196,7 @@ define <16 x i32> @vpmulld_test(<16 x i32> %i, <16 x i32> %j) {
 }
 
 ; CHECK-LABEL: sqrtA
-; CHECK: vsqrtssz
+; CHECK: vsqrtss {{.*}} encoding: [0x62
 ; CHECK: ret
 declare float @sqrtf(float) readnone
 define float @sqrtA(float %a) nounwind uwtable readnone ssp {
@@ -206,7 +206,7 @@ entry:
 }
 
 ; CHECK-LABEL: sqrtB
-; CHECK: vsqrtsdz
+; CHECK: vsqrtsd {{.*}}## encoding: [0x62
 ; CHECK: ret
 declare double @sqrt(double) readnone
 define double @sqrtB(double %a) nounwind uwtable readnone ssp {
@@ -216,7 +216,7 @@ entry:
 }
 
 ; CHECK-LABEL: sqrtC
-; CHECK: vsqrtssz
+; CHECK: vsqrtss {{.*}}## encoding: [0x62
 ; CHECK: ret
 declare float @llvm.sqrt.f32(float)
 define float @sqrtC(float %a) nounwind {

diff --git a/test/CodeGen/X86/avx512-cmp.ll b/test/CodeGen/X86/avx512-cmp.ll
@@ -1,6 +1,6 @@
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl --show-mc-encoding | FileCheck %s
 
-; CHECK: vucomisdz
+; CHECK: vucomisd {{.*}}encoding: [0x62
 define double @test1(double %a, double %b) nounwind {
   %tobool = fcmp une double %a, %b
   br i1 %tobool, label %l1, label %l2
@@ -13,7 +13,7 @@ l2:
   ret double %c1
 }
 
-; CHECK: vucomissz
+; CHECK: vucomiss {{.*}}encoding: [0x62
 define float @test2(float %a, float %b) nounwind {
   %tobool = fcmp olt float %a, %b
   br i1 %tobool, label %l1, label %l2

diff --git a/test/CodeGen/X86/avx512-cvt.ll b/test/CodeGen/X86/avx512-cvt.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl --show-mc-encoding | FileCheck %s
 
 ; CHECK-LABEL: sitof32
 ; CHECK: vcvtdq2ps %zmm
@@ -67,7 +67,7 @@ define <8 x double> @fpext00(<8 x float> %b) nounwind {
 }
 
 ; CHECK-LABEL: funcA
-; CHECK: vcvtsi2sdqz (%
+; CHECK: vcvtsi2sdq (%rdi){{.*}} encoding: [0x62
 ; CHECK: ret
 define double @funcA(i64* nocapture %e) {
 entry:
@@ -77,7 +77,7 @@ entry:
 }
 
 ; CHECK-LABEL: funcB
-; CHECK: vcvtsi2sdlz (%
+; CHECK: vcvtsi2sdl (%{{.*}} encoding: [0x62
 ; CHECK: ret
 define double @funcB(i32* %e) {
 entry:
@@ -87,7 +87,7 @@ entry:
 }
 
 ; CHECK-LABEL: funcC
-; CHECK: vcvtsi2sslz (%
+; CHECK: vcvtsi2ssl (%{{.*}} encoding: [0x62
 ; CHECK: ret
 define float @funcC(i32* %e) {
 entry:
@@ -97,7 +97,7 @@ entry:
 }
 
 ; CHECK-LABEL: i64tof32
-; CHECK: vcvtsi2ssqz  (%
+; CHECK: vcvtsi2ssq  (%{{.*}} encoding: [0x62
 ; CHECK: ret
 define float @i64tof32(i64* %e) {
 entry:
@@ -107,7 +107,7 @@ entry:
 }
 
 ; CHECK-LABEL: fpext
-; CHECK: vcvtss2sdz
+; CHECK: vcvtss2sd {{.*}} encoding: [0x62
 ; CHECK: ret
 define void @fpext() {
 entry:
@@ -120,9 +120,9 @@ entry:
 }
 
 ; CHECK-LABEL: fpround_scalar
-; CHECK: vmovsdz
-; CHECK: vcvtsd2ssz
-; CHECK: vmovssz
+; CHECK: vmovsd {{.*}} encoding: [0x62
+; CHECK: vcvtsd2ss {{.*}} encoding: [0x62
+; CHECK: vmovss {{.*}} encoding: [0x62
 ; CHECK: ret
 define void @fpround_scalar() nounwind uwtable {
 entry:
@@ -135,31 +135,31 @@ entry:
 }
 
 ; CHECK-LABEL: long_to_double
-; CHECK: vmovqz
+; CHECK: vmovq {{.*}} encoding: [0x62
 ; CHECK: ret
 define double @long_to_double(i64 %x) {
    %res = bitcast i64 %x to double
    ret double %res
 }
 
 ; CHECK-LABEL: double_to_long
-; CHECK: vmovqz
+; CHECK: vmovq {{.*}} encoding: [0x62
 ; CHECK: ret
 define i64 @double_to_long(double %x) {
    %res = bitcast double %x to i64
    ret i64 %res
 }
 
 ; CHECK-LABEL: int_to_float
-; CHECK: vmovdz
+; CHECK: vmovd {{.*}} encoding: [0x62
 ; CHECK: ret
 define float @int_to_float(i32 %x) {
    %res = bitcast i32 %x to float
    ret float %res
 }
 
 ; CHECK-LABEL: float_to_int
-; CHECK: vmovdz
+; CHECK: vmovd {{.*}} encoding: [0x62
 ; CHECK: ret
 define i32 @float_to_int(float %x) {
    %res = bitcast float %x to i32
@@ -185,15 +185,15 @@ define <16 x float> @uitof32(<16 x i32> %a) nounwind {
 }
 
 ; CHECK-LABEL: @fptosi02
-; CHECK vcvttss2siz
+; CHECK vcvttss2si {{.*}} encoding: [0x62
 ; CHECK: ret
 define i32 @fptosi02(float %a) nounwind {
   %b = fptosi float %a to i32
   ret i32 %b
 }
 
 ; CHECK-LABEL: @fptoui02
-; CHECK vcvttss2usiz
+; CHECK vcvttss2usi {{.*}} encoding: [0x62
 ; CHECK: ret
 define i32 @fptoui02(float %a) nounwind {
   %b = fptoui float %a to i32

diff --git a/test/CodeGen/X86/avx512-insert-extract.ll b/test/CodeGen/X86/avx512-insert-extract.ll
@@ -44,7 +44,7 @@ define <8 x i64> @test4(<8 x i64> %x) nounwind {
 }
 
 ;CHECK-LABEL: test5:
-;CHECK: vextractpsz
+;CHECK: vextractps
 ;CHECK: ret
 define i32 @test5(<4 x float> %x) nounwind {
   %ef = extractelement <4 x float> %x, i32 3
@@ -53,7 +53,7 @@ define i32 @test5(<4 x float> %x) nounwind {
 }
 
 ;CHECK-LABEL: test6:
-;CHECK: vextractpsz {{.*}}, (%rdi)
+;CHECK: vextractps {{.*}}, (%rdi)
 ;CHECK: ret
 define void @test6(<4 x float> %x, float* %out) nounwind {
   %ef = extractelement <4 x float> %x, i32 3
@@ -62,7 +62,7 @@ define void @test6(<4 x float> %x, float* %out) nounwind {
 }
 
 ;CHECK-LABEL: test7
-;CHECK: vmovdz
+;CHECK: vmovd
 ;CHECK: vpermps %zmm
 ;CHECK: ret
 define float @test7(<16 x float> %x, i32 %ind) nounwind {
@@ -71,7 +71,7 @@ define float @test7(<16 x float> %x, i32 %ind) nounwind {
 }
 
 ;CHECK-LABEL: test8
-;CHECK: vmovqz
+;CHECK: vmovq
 ;CHECK: vpermpd %zmm
 ;CHECK: ret
 define double @test8(<8 x double> %x, i32 %ind) nounwind {
@@ -89,7 +89,7 @@ define float @test9(<8 x float> %x, i32 %ind) nounwind {
 }
 
 ;CHECK-LABEL: test10
-;CHECK: vmovdz
+;CHECK: vmovd
 ;CHECK: vpermd %zmm
 ;CHEKK: vmovdz  %xmm0, %eax
 ;CHECK: ret

diff --git a/test/CodeGen/X86/avx512-intrinsics.ll b/test/CodeGen/X86/avx512-intrinsics.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl --show-mc-encoding| FileCheck %s
 
 declare i32 @llvm.x86.avx512.kortestz.w(i16, i16) nounwind readnone
 ; CHECK-LABEL: test_kortestz
@@ -147,73 +147,73 @@ define <16 x float> @test_sqrt_ps_512(<16 x float> %a0) {
 declare <16 x float> @llvm.x86.avx512.sqrt.ps.512(<16 x float>) nounwind readnone
 
 define <4 x float> @test_sqrt_ss(<4 x float> %a0, <4 x float> %a1) {
-  ; CHECK: vsqrtssz
+  ; CHECK: vsqrtss {{.*}}encoding: [0x62
   %res = call <4 x float> @llvm.x86.avx512.sqrt.ss(<4 x float> %a0, <4 x float> %a1) ; <<4 x float>> [#uses=1]
   ret <4 x float> %res
 }
 declare <4 x float> @llvm.x86.avx512.sqrt.ss(<4 x float>, <4 x float>) nounwind readnone
 
 define <2 x double> @test_sqrt_sd(<2 x double> %a0, <2 x double> %a1) {
-  ; CHECK: vsqrtsdz
+  ; CHECK: vsqrtsd {{.*}}encoding: [0x62
   %res = call <2 x double> @llvm.x86.avx512.sqrt.sd(<2 x double> %a0, <2 x double> %a1) ; <<2 x double>> [#uses=1]
   ret <2 x double> %res
 }
 declare <2 x double> @llvm.x86.avx512.sqrt.sd(<2 x double>, <2 x double>) nounwind readnone
 
 define i64 @test_x86_sse2_cvtsd2si64(<2 x double> %a0) {
-  ; CHECK: vcvtsd2siz
+  ; CHECK: vcvtsd2si {{.*}}encoding: [0x62
   %res = call i64 @llvm.x86.sse2.cvtsd2si64(<2 x double> %a0) ; <i64> [#uses=1]
   ret i64 %res
 }
 declare i64 @llvm.x86.sse2.cvtsd2si64(<2 x double>) nounwind readnone
 
 define <2 x double> @test_x86_sse2_cvtsi642sd(<2 x double> %a0, i64 %a1) {
-  ; CHECK: vcvtsi2sdqz
+  ; CHECK: vcvtsi2sdq {{.*}}encoding: [0x62
   %res = call <2 x double> @llvm.x86.sse2.cvtsi642sd(<2 x double> %a0, i64 %a1) ; <<2 x double>> [#uses=1]
   ret <2 x double> %res
 }
 declare <2 x double> @llvm.x86.sse2.cvtsi642sd(<2 x double>, i64) nounwind readnone
 
 define <2 x double> @test_x86_avx512_cvtusi642sd(<2 x double> %a0, i64 %a1) {
-  ; CHECK: vcvtusi2sdqz
+  ; CHECK: vcvtusi2sdq {{.*}}encoding: [0x62
   %res = call <2 x double> @llvm.x86.avx512.cvtusi642sd(<2 x double> %a0, i64 %a1) ; <<2 x double>> [#uses=1]
   ret <2 x double> %res
 }
 declare <2 x double> @llvm.x86.avx512.cvtusi642sd(<2 x double>, i64) nounwind readnone
 
 define i64 @test_x86_sse2_cvttsd2si64(<2 x double> %a0) {
-  ; CHECK: vcvttsd2siz
+  ; CHECK: vcvttsd2si {{.*}}encoding: [0x62
   %res = call i64 @llvm.x86.sse2.cvttsd2si64(<2 x double> %a0) ; <i64> [#uses=1]
   ret i64 %res
 }
 declare i64 @llvm.x86.sse2.cvttsd2si64(<2 x double>) nounwind readnone
 
 
 define i64 @test_x86_sse_cvtss2si64(<4 x float> %a0) {
-  ; CHECK: vcvtss2siz
+  ; CHECK: vcvtss2si {{.*}}encoding: [0x62
   %res = call i64 @llvm.x86.sse.cvtss2si64(<4 x float> %a0) ; <i64> [#uses=1]
   ret i64 %res
 }
 declare i64 @llvm.x86.sse.cvtss2si64(<4 x float>) nounwind readnone
 
 
 define <4 x float> @test_x86_sse_cvtsi642ss(<4 x float> %a0, i64 %a1) {
-  ; CHECK: vcvtsi2ssqz
+  ; CHECK: vcvtsi2ssq {{.*}}encoding: [0x62
   %res = call <4 x float> @llvm.x86.sse.cvtsi642ss(<4 x float> %a0, i64 %a1) ; <<4 x float>> [#uses=1]
   ret <4 x float> %res
 }
 declare <4 x float> @llvm.x86.sse.cvtsi642ss(<4 x float>, i64) nounwind readnone
 
 
 define i64 @test_x86_sse_cvttss2si64(<4 x float> %a0) {
-  ; CHECK: vcvttss2siz
+  ; CHECK: vcvttss2si {{.*}}encoding: [0x62
   %res = call i64 @llvm.x86.sse.cvttss2si64(<4 x float> %a0) ; <i64> [#uses=1]
   ret i64 %res
 }
 declare i64 @llvm.x86.sse.cvttss2si64(<4 x float>) nounwind readnone
 
 define i64 @test_x86_avx512_cvtsd2usi64(<2 x double> %a0) {
-  ; CHECK: vcvtsd2usiz
+  ; CHECK: vcvtsd2usi {{.*}}encoding: [0x62
   %res = call i64 @llvm.x86.avx512.cvtsd2usi64(<2 x double> %a0) ; <i64> [#uses=1]
   ret i64 %res
 }