replica
/
hanchenye-llvm-project
2
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Clean up AddedComplexity further after adding UseSSEx 

llvm-svn: 162973
This commit is contained in:
Michael Liao 2012-08-31 03:01:35 +00:00
parent 97e44a0cc7
commit 969f3913dd
1 changed files with 13 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
llvm/lib/Target/X86/X86InstrSSE.td
View File

@ -1445,7 +1445,7 @@ def : InstAlias<"vcvtsi2sd{l}\t{$src, $src1, $dst|$dst, $src1, $src}",
def : InstAlias<"vcvtsi2sd{l}\t{$src, $src1, $dst|$dst, $src1, $src}",
(VCVTSI2SDrm FR64:$dst, FR64:$src1, i32mem:$src)>;
let Predicates = [HasAVX], AddedComplexity = 1 in {
let Predicates = [HasAVX] in {
def : Pat<(f32 (sint_to_fp (loadi32 addr:$src))),
(VCVTSI2SSrm (f32 (IMPLICIT_DEF)), addr:$src)>;
def : Pat<(f32 (sint_to_fp (loadi64 addr:$src))),
@ -1698,19 +1698,17 @@ def VCVTSS2SDrm : I<0x5A, MRMSrcMem, (outs FR64:$dst),
XS, VEX_4V, VEX_LIG, Requires<[HasAVX, OptForSize]>;
}
let AddedComplexity = 1 in { // give AVX priority
def : Pat<(f64 (fextend FR32:$src)),
(VCVTSS2SDrr FR32:$src, FR32:$src)>, Requires<[HasAVX]>;
def : Pat<(fextend (loadf32 addr:$src)),
(VCVTSS2SDrm (f32 (IMPLICIT_DEF)), addr:$src)>, Requires<[HasAVX]>;
def : Pat<(f64 (fextend FR32:$src)),
(VCVTSS2SDrr FR32:$src, FR32:$src)>, Requires<[HasAVX]>;
def : Pat<(fextend (loadf32 addr:$src)),
(VCVTSS2SDrm (f32 (IMPLICIT_DEF)), addr:$src)>, Requires<[HasAVX]>;
def : Pat<(extloadf32 addr:$src),
(VCVTSS2SDrm (f32 (IMPLICIT_DEF)), addr:$src)>,
Requires<[HasAVX, OptForSize]>;
def : Pat<(extloadf32 addr:$src),
(VCVTSS2SDrr (f32 (IMPLICIT_DEF)), (VMOVSSrm addr:$src))>,
Requires<[HasAVX, OptForSpeed]>;
} // AddedComplexity = 1
def : Pat<(extloadf32 addr:$src),
(VCVTSS2SDrm (f32 (IMPLICIT_DEF)), addr:$src)>,
Requires<[HasAVX, OptForSize]>;
def : Pat<(extloadf32 addr:$src),
(VCVTSS2SDrr (f32 (IMPLICIT_DEF)), (VMOVSSrm addr:$src))>,
Requires<[HasAVX, OptForSpeed]>;
def CVTSS2SDrr : I<0x5A, MRMSrcReg, (outs FR64:$dst), (ins FR32:$src),
"cvtss2sd\t{$src, $dst|$dst, $src}",
@ -2509,7 +2507,7 @@ let Predicates = [HasAVX1Only] in {
(VUNPCKHPDYrr VR256:$src1, VR256:$src2)>;
}
let Predicates = [HasAVX], AddedComplexity = 1 in {
let Predicates = [HasAVX] in {
// FIXME: Instead of X86Movddup, there should be a X86Unpckl here, the
// problem is during lowering, where it's not possible to recognize the load
// fold cause it has two uses through a bitcast. One use disappears at isel
@ -3180,7 +3178,6 @@ let Predicates = [HasAVX] in {
SSE_RCPP>, VEX;
}
let AddedComplexity = 1 in {
def : Pat<(f32 (fsqrt FR32:$src)),
(VSQRTSSr (f32 (IMPLICIT_DEF)), FR32:$src)>, Requires<[HasAVX]>;
def : Pat<(f32 (fsqrt (load addr:$src))),
@ -3203,9 +3200,8 @@ def : Pat<(f32 (X86frcp FR32:$src)),
def : Pat<(f32 (X86frcp (load addr:$src))),
(VRCPSSm (f32 (IMPLICIT_DEF)), addr:$src)>,
Requires<[HasAVX, OptForSize]>;
}
let Predicates = [HasAVX], AddedComplexity = 1 in {
let Predicates = [HasAVX] in {
def : Pat<(int_x86_sse_sqrt_ss VR128:$src),
(COPY_TO_REGCLASS (VSQRTSSr (f32 (IMPLICIT_DEF)),
(COPY_TO_REGCLASS VR128:$src, FR32)),