From 7ba2725f5d3feb99b705de5e89b7c49b6687e5ed Mon Sep 17 00:00:00 2001 From: Craig Topper Date: Fri, 30 Dec 2011 07:16:00 +0000 Subject: [PATCH] Make FMA4 imply AVX so that YMM registers would be available. Necessitates removing from Bulldozer CPU types since it would enable AVX code generation implicitly. Also make SSE4A imply SSE3. Without some level of SSE implied, XMM registers wouldn't be legal. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@147369 91177308-0d34-0410-b5e6-96231b3b80d8 --- lib/Target/X86/X86.td | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/lib/Target/X86/X86.td b/lib/Target/X86/X86.td index d053c76de78..66779361da1 100644 --- a/lib/Target/X86/X86.td +++ b/lib/Target/X86/X86.td @@ -77,7 +77,8 @@ def FeatureFastUAMem : SubtargetFeature<"fast-unaligned-mem", "IsUAMemFast", "true", "Fast unaligned memory access">; def FeatureSSE4A : SubtargetFeature<"sse4a", "HasSSE4A", "true", - "Support SSE 4a instructions">; + "Support SSE 4a instructions", + [FeatureSSE3]>; def FeatureAVX : SubtargetFeature<"avx", "HasAVX", "true", "Enable AVX instructions">; @@ -90,8 +91,9 @@ def FeatureFMA3 : SubtargetFeature<"fma3", "HasFMA3", "true", "Enable three-operand fused multiple-add", [FeatureAVX]>; def FeatureFMA4 : SubtargetFeature<"fma4", "HasFMA4", "true", - "Enable four-operand fused multiple-add">; -def FeatureXOP : SubtargetFeature<"xop", "HasXOP", "true", + "Enable four-operand fused multiple-add", + [FeatureAVX]>; +def FeatureXOP : SubtargetFeature<"xop", "HasXOP", "true", "Enable XOP instructions">; def FeatureVectorUAMem : SubtargetFeature<"vector-unaligned-mem", "HasVectorUAMem", "true", @@ -201,12 +203,12 @@ def : Proc<"athlon64-sse3", [FeatureSSE3, Feature3DNowA, FeatureCMPXCHG16B, def : Proc<"amdfam10", [FeatureSSE3, FeatureSSE4A, Feature3DNowA, FeatureCMPXCHG16B, FeatureLZCNT, FeaturePOPCNT, FeatureSlowBTMem]>; -// FIXME: Disabling AVX for now since it's not ready. +// FIXME: Disabling AVX/FMA4 for now since it's not ready. def : Proc<"bdver1", [FeatureSSE42, FeatureSSE4A, FeatureCMPXCHG16B, - FeatureAES, FeatureCLMUL, FeatureFMA4, + FeatureAES, FeatureCLMUL, FeatureXOP, FeatureLZCNT, FeaturePOPCNT]>; def : Proc<"bdver2", [FeatureSSE42, FeatureSSE4A, FeatureCMPXCHG16B, - FeatureAES, FeatureCLMUL, FeatureFMA4, + FeatureAES, FeatureCLMUL, FeatureXOP, FeatureF16C, FeatureLZCNT, FeaturePOPCNT, FeatureBMI]>; -- 2.34.1