Commits · 2216f73676b8112fa9d29c866fae3fa617dc401e · Roger Ferrer / llvm-epi-0.8

Nov 07, 2011
- Various Mips64 floating point instruction patterns. · 2216f736
  Akira Hatanaka authored Nov 07, 2011
```
llvm-svn: 144019
```
  2216f736
- Add definition of the base class for floating point comparison instructions · b2d37760
  Akira Hatanaka authored Nov 07, 2011
```
and add Mips64's version too.

llvm-svn: 144018
```
  b2d37760
- Add code needed for copying between 64-bit integer and floating pointer · 81c14002
  Akira Hatanaka authored Nov 07, 2011
```
registers.

llvm-svn: 144017
```
  81c14002
- Add definitions of 64-bit instructions which move data between integer and · 1537e297
  Akira Hatanaka authored Nov 07, 2011
```
floating pointer registers.

llvm-svn: 144016
```
  1537e297
- Simplify some uses of utohexstr. · 69d57cf9
  Benjamin Kramer authored Nov 07, 2011
```
As a side effect hex is printed lowercase instead of uppercase now.

llvm-svn: 144013
```
  69d57cf9
- Simplify code. No functionality change. · 03d73e47
  Benjamin Kramer authored Nov 07, 2011
```
llvm-svn: 144012
```
  03d73e47
- Expand V_SET0 to xorps by default. · 02413089
  Jakob Stoklund Olesen authored Nov 07, 2011
```
The xorps instruction is smaller than pxor, so prefer that encoding.

The ExecutionDepsFix pass will switch the encoding to pxor and xorpd
when appropriate.

llvm-svn: 143996
```
  02413089
- Add definition of 64-bit load upper immediate. · 2b8d1f16
  Akira Hatanaka authored Nov 07, 2011
```
llvm-svn: 143994
```
  2b8d1f16
- Include RegSaveAreaSize in the computation of stack size. · 2f448004
  Akira Hatanaka authored Nov 07, 2011
```
llvm-svn: 143993
```
  2f448004
- Define functions that get or set the size of area on callee's stack frame which · 7bcecd48
  Akira Hatanaka authored Nov 07, 2011
```
is used to save va_arg or byval arguments passed in registers.

llvm-svn: 143992
```
  7bcecd48
- Use array_lengthof to compute the number of iterations of a loop. · d9c2e46c
  Akira Hatanaka authored Nov 07, 2011
```
llvm-svn: 143991
```
  d9c2e46c
- Fix patterns for unaligned 32-bit load. DSLL32 or DSRL32 should be emitted · cf7e5b09
  Akira Hatanaka authored Nov 07, 2011
```
when shift amount is larger than 32.

llvm-svn: 143990
```
  cf7e5b09
- Make the type of shift amount i32 in order to reduce the number of shift · 770f0646
  Akira Hatanaka authored Nov 07, 2011
```
instruction definitions.

llvm-svn: 143989
```
  770f0646
- Add 64-bit to 32-bit trunc pattern. · d5c13290
  Akira Hatanaka authored Nov 07, 2011
```
llvm-svn: 143988
```
  d5c13290
- Add AVX2 variable shift instructions and intrinsics. · a6d409d5
  Craig Topper authored Nov 07, 2011
```
llvm-svn: 143915
```
  a6d409d5
- Add AVX2 VPMOVMASK instructions and intrinsics. · ff39be0a
  Craig Topper authored Nov 07, 2011
```
llvm-svn: 143904
```
  ff39be0a
- Add AVX2 VEXTRACTI128 and VINSERTI128 instructions. Fix VPERM2I128 to be... · e122dcbf
  Craig Topper authored Nov 07, 2011
```
Add AVX2 VEXTRACTI128 and VINSERTI128 instructions. Fix VPERM2I128 to be qualified with HasAVX2 instead of HasAVX. Mark VINSERTF128 and VEXTRACTF128 as never having side effects.

llvm-svn: 143902
```
  e122dcbf
- More AVX2 instructions and their intrinsics. · f01f1b5c
  Craig Topper authored Nov 06, 2011
```
llvm-svn: 143895
```
  f01f1b5c
Nov 06, 2011
- Replace (Lower|Upper)caseString in favor of StringRef's newest methods. · 20baffb2
  Benjamin Kramer authored Nov 06, 2011
```
llvm-svn: 143891
```
  20baffb2
- Add more AVX2 instructions and intrinsics. · 05d1cb98
  Craig Topper authored Nov 06, 2011
```
llvm-svn: 143861
```
  05d1cb98
Nov 05, 2011
- Add support for passing i1, i8, and i16 call parameters. Also, be sure to · d0191a53
  Chad Rosier authored Nov 05, 2011
```
zero-extend the constant integer encoding.  Test case provides testing for
both call parameters and materialization of i1, i8, and i16 types.

llvm-svn: 143821
```
  d0191a53
- Add more PRI.64 macros for MSVC and use them throughout the codebase. · f3da5290
  Benjamin Kramer authored Nov 05, 2011
```
llvm-svn: 143799
```
  f3da5290
- Allow i1 to be promoted to i32 for ARM APCS calling convention. · f0055f61
  Chad Rosier authored Nov 05, 2011
```
llvm-svn: 143755
```
  f0055f61
- Enhanced vzeroupper insertion pass that avoids inserting vzeroupper where it... · 8f249600
  Eli Friedman authored Nov 04, 2011
```
Enhanced vzeroupper insertion pass that avoids inserting vzeroupper where it is unnecessary through local analysis.  Patch from Bruno Cardoso Lopes, with some additional changes.

I'm going to wait for any review comments and perform some additional testing before turning this on by default.

llvm-svn: 143750
```
  8f249600
- Cannot create a result register for non-legal types. · 5b8fdd7b
  Chad Rosier authored Nov 04, 2011
```
llvm-svn: 143749
```
  5b8fdd7b
- When materializing an i32, SExt vs ZExt doesn't matter when we're trying to fit · e8b8b773
  Chad Rosier authored Nov 04, 2011
```
in a 16-bit immediate.  However, for the shorter non-legal types (i.e., i1, i8,
i16) we should not sign-extend.  This prevents us from materializing things
such as 'true' (i.e., i1 1).

llvm-svn: 143743
```
  e8b8b773
Nov 04, 2011
- Enable support for materializing i1, i8, and i16 integers via move immediate. · 67f96887
  Chad Rosier authored Nov 04, 2011
```
llvm-svn: 143739
```
  67f96887
- build/cmake: Coalesce the configuration time header include fragment generation · 4a2eab0d
  Daniel Dunbar authored Nov 04, 2011
```
for target definitions.

llvm-svn: 143731
```
  4a2eab0d
- build/cmake: Use tblgen macro directly instead of llvm_tablegen, which just · 4a9c6426
  Daniel Dunbar authored Nov 04, 2011
```
added a layer of indirection with no value (not even conciseness).

llvm-svn: 143727
```
  4a9c6426
- Add missing argument for atomic instructions in c++ backend. PR11268, part 2. · 5b693c2f
  Eli Friedman authored Nov 04, 2011
```
llvm-svn: 143712
```
  5b693c2f
- Add intrinsics for X86 vcvtps2ph and vcvtph2ps instructions · caba032f
  Craig Topper authored Nov 04, 2011
```
llvm-svn: 143683
```
  caba032f
- Fix some minor scheduling itinerary bug. It's not expected to actually affect codegen. · 1ddeb167
  Evan Cheng authored Nov 04, 2011
```
llvm-svn: 143675
```
  1ddeb167
- Indentation. · 8a98ec4d
  Chad Rosier authored Nov 04, 2011
```
llvm-svn: 143670
```
  8a98ec4d
- Add fast-isel support for returning i1, i8, and i16. · f3e73ad5
  Chad Rosier authored Nov 04, 2011
```
llvm-svn: 143669
```
  f3e73ad5
Nov 03, 2011
- Reapply r143206, with fixes. Disallow physical register lifetimes · 198b7ffc
  Dan Gohman authored Nov 03, 2011
```
across calls, and only check for nested dependences on the special
call-sequence-resource register.

llvm-svn: 143660
```
  198b7ffc
- fixed global array handling for ptx to use the correct bit widths · b68515c2
  Dan Bailey authored Nov 03, 2011
```
llvm-svn: 143640
```
  b68515c2
- build: Add initial cut at LLVMBuild.txt files. · bf9bba47
  Daniel Dunbar authored Nov 03, 2011
```
llvm-svn: 143634
```
  bf9bba47
- Add new X86 AVX2 VBROADCAST instructions. · 0e7cbbab
  Craig Topper authored Nov 03, 2011
```
llvm-svn: 143612
```
  0e7cbbab
- Add support for sign-extending non-legal types in SelectSIToFP(). · bf5f4bec
  Chad Rosier authored Nov 03, 2011
```
llvm-svn: 143603
```
  bf5f4bec
- Fixed parameter name. · 1f4603d4
  Lang Hames authored Nov 02, 2011
```
llvm-svn: 143594
```
  1f4603d4