maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

Remove some AddedComplexity tags that were forcing priority for AVX over SSE. Use predicates instead. 

llvm-svn: 200458
This commit is contained in:
Craig Topper 2014-01-30 06:26:25 +00:00
parent ed0d1ccc95
commit c45da1619c
1 changed files with 36 additions and 36 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

72
llvm/lib/Target/X86/X86InstrSSE.td
View File

@ -3055,24 +3055,26 @@ let isCodeGenOnly = 1 in {
// we now generate:
// addss %xmm1, %xmm0
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst), (v4f32 (scalar_to_vector (fadd
(f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
FR32:$src))))),
(ADDSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst), (v4f32 (scalar_to_vector (fsub
(f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
FR32:$src))))),
(SUBSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst), (v4f32 (scalar_to_vector (fmul
(f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
FR32:$src))))),
(MULSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst), (v4f32 (scalar_to_vector (fdiv
(f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
FR32:$src))))),
(DIVSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
let Predicates = [UseSSE1] in {
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst), (v4f32 (scalar_to_vector (fadd
(f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
FR32:$src))))),
(ADDSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst), (v4f32 (scalar_to_vector (fsub
(f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
FR32:$src))))),
(SUBSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst), (v4f32 (scalar_to_vector (fmul
(f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
FR32:$src))))),
(MULSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst), (v4f32 (scalar_to_vector (fdiv
(f32 (vector_extract (v4f32 VR128:$dst), (iPTR 0))),
FR32:$src))))),
(DIVSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
}
let Predicates = [HasSSE2] in {
let Predicates = [UseSSE2] in {
// SSE2 patterns to select scalar double-precision fp arithmetic instructions
def : Pat<(v2f64 (X86Movsd (v2f64 VR128:$dst), (v2f64 (scalar_to_vector (fadd
@ -3117,11 +3119,9 @@ let Predicates = [UseSSE41] in {
(DIVSSrr_Int v4f32:$dst, (COPY_TO_REGCLASS FR32:$src, VR128))>;
}
let AddedComplexity = 20, Predicates = [HasAVX] in {
let Predicates = [HasAVX] in {
// The following patterns select AVX Scalar single/double precision fp
// arithmetic instructions.
// The 'AddedComplexity' is required to give them higher priority over
// the equivalent SSE/SSE2 patterns.
def : Pat<(v2f64 (X86Movsd (v2f64 VR128:$dst), (v2f64 (scalar_to_vector (fadd
(f64 (vector_extract (v2f64 VR128:$dst), (iPTR 0))),
@ -3176,20 +3176,22 @@ let AddedComplexity = 20, Predicates = [HasAVX] in {
// we now generate:
// addss %xmm1, %xmm0
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fadd (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
(ADDSSrr_Int v4f32:$dst, v4f32:$src)>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fsub (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
(SUBSSrr_Int v4f32:$dst, v4f32:$src)>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fmul (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
(MULSSrr_Int v4f32:$dst, v4f32:$src)>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fdiv (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
(DIVSSrr_Int v4f32:$dst, v4f32:$src)>;
let Predicates = [UseSSE1] in {
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fadd (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
(ADDSSrr_Int v4f32:$dst, v4f32:$src)>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fsub (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
(SUBSSrr_Int v4f32:$dst, v4f32:$src)>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fmul (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
(MULSSrr_Int v4f32:$dst, v4f32:$src)>;
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fdiv (v4f32 VR128:$dst), (v4f32 VR128:$src)))),
(DIVSSrr_Int v4f32:$dst, v4f32:$src)>;
}
let Predicates = [HasSSE2] in {
let Predicates = [UseSSE2] in {
// SSE2 patterns to select scalar double-precision fp arithmetic instructions
// from a packed double-precision fp instruction plus movsd.
@ -3207,12 +3209,10 @@ let Predicates = [HasSSE2] in {
(DIVSDrr_Int v2f64:$dst, v2f64:$src)>;
}
let AddedComplexity = 20, Predicates = [HasAVX] in {
let Predicates = [HasAVX] in {
// The following patterns select AVX Scalar single/double precision fp
// arithmetic instructions from a packed single precision fp instruction
// plus movss/movsd.
// The 'AddedComplexity' is required to give them higher priority over
// the equivalent SSE/SSE2 patterns.
def : Pat<(v4f32 (X86Movss (v4f32 VR128:$dst),
(fadd (v4f32 VR128:$dst), (v4f32 VR128:$src)))),