- Apr 28, 2006
-
-
Chris Lattner authored
llvm-svn: 28007
-
Evan Cheng authored
llvm-svn: 28005
-
Evan Cheng authored
node can be folded. llvm-svn: 28003
-
Chris Lattner authored
and is already available, instead of falling back to emitting a load, fall back to emitting a reg-reg copy. This generates significantly better code for some SSE testcases, as SSE has lots of two-address instructions and none of them are read/modify/write. As one example, this change does: pshufd %XMM5, XMMWORD PTR [%ESP + 84], 255 xorps %XMM2, %XMM5 cmpltps %XMM1, %XMM0 - movaps XMMWORD PTR [%ESP + 52], %XMM0 - movapd %XMM6, XMMWORD PTR [%ESP + 52] + movaps %XMM6, %XMM0 cmpltps %XMM6, XMMWORD PTR [%ESP + 68] movapd XMMWORD PTR [%ESP + 52], %XMM6 movaps %XMM6, %XMM0 cmpltps %XMM6, XMMWORD PTR [%ESP + 36] cmpltps %XMM3, %XMM0 - movaps XMMWORD PTR [%ESP + 20], %XMM0 - movapd %XMM7, XMMWORD PTR [%ESP + 20] + movaps %XMM7, %XMM0 cmpltps %XMM7, XMMWORD PTR [%ESP + 4] movapd XMMWORD PTR [%ESP + 20], %XMM7 cmpltps %XMM4, %XMM0 ... which is far better than a store followed by a load! llvm-svn: 28001
-
Chris Lattner authored
llvm-svn: 27999
-
- Apr 27, 2006
-
-
Chris Lattner authored
llvm-svn: 27998
-
Chris Lattner authored
Transforms/InstCombine/vec_insert_to_shuffle.ll llvm-svn: 27997
-
Evan Cheng authored
for now. Patch contributed by Alexander Friedman. llvm-svn: 27994
-
Evan Cheng authored
llvm-svn: 27993
-
Evan Cheng authored
llvm-svn: 27992
-
Evan Cheng authored
(VAND, VADD, etc.). Legalizer will assert otherwise. llvm-svn: 27991
-
Evan Cheng authored
llvm-svn: 27989
-
Evan Cheng authored
llvm-svn: 27988
-
Chris Lattner authored
PR748. llvm-svn: 27987
-
Evan Cheng authored
- Fixed vararg support. llvm-svn: 27985
-
Chris Lattner authored
nondeterminism being bad) could cause some trivial missed optimizations (dead phi nodes being left around for later passes to clean up). With this, llvm-gcc4 now bootstraps and correctly compares. I don't know why I never tried to do it before... :) llvm-svn: 27984
-
- Apr 26, 2006
-
-
Chris Lattner authored
recognize some simple affine IV's. llvm-svn: 27982
-
Evan Cheng authored
llvm-svn: 27980
-
Evan Cheng authored
llvm-svn: 27975
-
Evan Cheng authored
llvm-svn: 27974
-
- Apr 25, 2006
-
-
Nate Begeman authored
failres. llvm-svn: 27973
-
Evan Cheng authored
llvm-svn: 27972
-
Andrew Lenharth authored
llvm-svn: 27971
-
Andrew Lenharth authored
llvm-svn: 27970
-
Evan Cheng authored
llvm-svn: 27968
-
Nate Begeman authored
llvm-svn: 27967
-
Nate Begeman authored
llvm-svn: 27966
-
Evan Cheng authored
result). Otherwise tblgen will pick the default (v16i8 for 128-bit vector). llvm-svn: 27965
-
Evan Cheng authored
a temporary workaround for the 2-wide vector_shuffle problem (i.e. its mask would have type v2i32 which is not legal). llvm-svn: 27964
-
Evan Cheng authored
llvm-svn: 27963
-
Evan Cheng authored
llvm-svn: 27961
-
- Apr 24, 2006
-
-
Evan Cheng authored
llvm-svn: 27960
-
Evan Cheng authored
llvm-svn: 27959
-
Evan Cheng authored
llvm-svn: 27958
-
Evan Cheng authored
llvm-svn: 27955
-
- Apr 23, 2006
-
-
Chris Lattner authored
llvm-svn: 27954
-
Evan Cheng authored
llvm-svn: 27953
-
Nate Begeman authored
the jump table's range check block. This re-enables 100% dense jump tables by default on PPC & x86 llvm-svn: 27952
-
Nate Begeman authored
these. llvm-svn: 27950
-
Nate Begeman authored
updating the machine CFG. llvm-svn: 27949
-