diff options
| author | Evan Cheng <evan.cheng@apple.com> | 2008-09-25 20:50:48 +0000 |
|---|---|---|
| committer | Evan Cheng <evan.cheng@apple.com> | 2008-09-25 20:50:48 +0000 |
| commit | 0b457f0c3a7e21b1fb9ac8b9f8e404e1312b6a60 (patch) | |
| tree | 2ba33e7e3f6ad009c22409a680fc91cb58deb29b /lib/Target/X86/X86InstrSSE.td | |
| parent | c9c6da61ac027d9818652d417907f84398288b99 (diff) | |
With sse3 and when the source is a load or has multiple uses, favors movddup over shuffp*, pshufd, etc. Without sse3 or when the source is from a register, make use of movlhps
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@56620 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'lib/Target/X86/X86InstrSSE.td')
| -rw-r--r-- | lib/Target/X86/X86InstrSSE.td | 33 |
1 files changed, 25 insertions, 8 deletions
diff --git a/lib/Target/X86/X86InstrSSE.td b/lib/Target/X86/X86InstrSSE.td index d67beac3e9..f9de8e7475 100644 --- a/lib/Target/X86/X86InstrSSE.td +++ b/lib/Target/X86/X86InstrSSE.td @@ -208,6 +208,10 @@ def SSE_splat_lo_mask : PatLeaf<(build_vector), [{ return X86::isSplatLoMask(N); }]>; +def MOVDDUP_shuffle_mask : PatLeaf<(build_vector), [{ + return X86::isMOVDDUPMask(N); +}]>; + def MOVHLPS_shuffle_mask : PatLeaf<(build_vector), [{ return X86::isMOVHLPSMask(N); }]>; @@ -755,6 +759,11 @@ def MOVHLPSrr : PSI<0x12, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src1, VR128: } // AddedComplexity } // Constraints = "$src1 = $dst" +let AddedComplexity = 15 in +def : Pat<(v4f32 (vector_shuffle VR128:$src, (undef), MOVDDUP_shuffle_mask)), + (MOVLHPSrr VR128:$src, VR128:$src)>, Requires<[HasSSE1]>; + + // Arithmetic @@ -2452,16 +2461,24 @@ def MOVSLDUPrm : S3SI<0x12, MRMSrcMem, (outs VR128:$dst), (ins f128mem:$src), def MOVDDUPrr : S3DI<0x12, MRMSrcReg, (outs VR128:$dst), (ins VR128:$src), "movddup\t{$src, $dst|$dst, $src}", - [(set VR128:$dst, (v2f64 (vector_shuffle - VR128:$src, (undef), - SSE_splat_lo_mask)))]>; + [(set VR128:$dst, + (v2f64 (vector_shuffle VR128:$src, (undef), + MOVDDUP_shuffle_mask)))]>; def MOVDDUPrm : S3DI<0x12, MRMSrcMem, (outs VR128:$dst), (ins f64mem:$src), "movddup\t{$src, $dst|$dst, $src}", - [(set VR128:$dst, - (v2f64 (vector_shuffle - (scalar_to_vector (loadf64 addr:$src)), - (undef), - SSE_splat_lo_mask)))]>; + [(set VR128:$dst, + (v2f64 (vector_shuffle + (scalar_to_vector (loadf64 addr:$src)), + (undef), MOVDDUP_shuffle_mask)))]>; + +def : Pat<(vector_shuffle + (bc_v2f64 (v2i64 (scalar_to_vector (loadi64 addr:$src)))), + (undef), MOVDDUP_shuffle_mask), + (MOVDDUPrm addr:$src)>, Requires<[HasSSE3]>; +def : Pat<(vector_shuffle + (memopv2f64 addr:$src), (undef), MOVDDUP_shuffle_mask), + (MOVDDUPrm addr:$src)>, Requires<[HasSSE3]>; + // Arithmetic let Constraints = "$src1 = $dst" in { |