Commits · e50b23e2235c241d618f38b3ca44213d1f4b5588 · Roger Ferrer / llvm-epi · GitLab

Sep 07, 2010
- Fix command line used to link these test cases. · e50b23e2
  Devang Patel authored Sep 07, 2010
  e50b23e2
- Reintroduce dbg-declare tests. · 9dc0e5be
  Devang Patel authored Sep 07, 2010
  9dc0e5be
- Remove last three tests. I need to make them independent of my setup. · 688338ee
  Devang Patel authored Sep 07, 2010
  688338ee
- Add a test case to check handling of dbg-declare during hybrid mode where we... · 55a3bab0
  Devang Patel authored Sep 07, 2010
  55a3bab0
- Add a test case to check handling of dbg-declare by selection DAG builder. · 29a775ad
  Devang Patel authored Sep 07, 2010
  29a775ad
- Add a test case to check handling of dbg-declare by fast-isel. · 184c81c3
  Devang Patel authored Sep 07, 2010
  184c81c3
Sep 05, 2010
- implement rdar://6653118 - fastisel should fold loads where possible. · eeba0c73
  Chris Lattner authored Sep 05, 2010
  eeba0c73
Sep 03, 2010
- Remove the rest of the nonexistent 64-bit AVX instructions. · 367afb5a
  Dale Johannesen authored Sep 03, 2010
  367afb5a
- Re-apply r112883: · 03f4be86
  Jim Grosbach authored Sep 03, 2010
  03f4be86
- Revert "For ARM stack frames that utilize variable sized objects and have... · 2ac3386e
  Daniel Dunbar authored Sep 03, 2010
  2ac3386e
- test/CodeGen/X86: Add explicit -mtriple=(i686|x86_64)-linux for Win32 host. · 24d039eb
  NAKAMURA Takumi authored Sep 03, 2010
  24d039eb
- AVX doesn't support mm operations neither its instrinsics. · d6634a5b
  Bruno Cardoso Lopes authored Sep 03, 2010
  d6634a5b
- Replace NEON vabdl, vaba, and vabal intrinsics with combinations of the · f65c9ef7
  Bob Wilson authored Sep 03, 2010
  f65c9ef7
- Properly emit __chkstk call instead of __alloca on non-mingw windows targets. · a5a64555
  Anton Korobeynikov authored Sep 02, 2010
  a5a64555
- For ARM stack frames that utilize variable sized objects and have either · 7fd9aea6
  Jim Grosbach authored Sep 02, 2010
  7fd9aea6
Sep 02, 2010
- Don't narrow the load and store in a load+twiddle+store sequence unless · 3c9b5f39
  Dan Gohman authored Sep 02, 2010
  3c9b5f39
- Fix an unnecessary XFAIL · 0ca17f7e
  Sandeep Patel authored Sep 02, 2010
  0ca17f7e
- Now that register allocation properly considers reserved regs, simplify the · 66c681a6
  Jim Grosbach authored Sep 02, 2010
  66c681a6
- Convert VLD1 and VLD2 instructions to use pseudo-instructions until · 75a6408f
  Bob Wilson authored Sep 02, 2010
  75a6408f
- test/loop-strength-reduce4: Add explicit triplet for Win32 host. · a224e556
  NAKAMURA Takumi authored Sep 02, 2010
  a224e556
- test/twoaddr-coalesce: Do not use @main. · 54ce5468
  NAKAMURA Takumi authored Sep 02, 2010
  54ce5468
- Remove NEON vmull, vmlal, and vmlsl intrinsics, replacing them with multiply, · 38ab35a9
  Bob Wilson authored Sep 01, 2010
  38ab35a9
- Using target specific nodes for shuffle nodes makes the mask · fea81b48
  Bruno Cardoso Lopes authored Sep 01, 2010
  fea81b48
- Teach RemoveCopyByCommutingDef to check all aliases, not just subregisters. · 4b6fd48b
  Jakob Stoklund Olesen authored Sep 01, 2010
  4b6fd48b
Sep 01, 2010
- temporarily revert r112664, it is causing a decoding conflict, and · 39eccb47
  Chris Lattner authored Sep 01, 2010
  39eccb47
- Revert 112442 and 112440 until the compile time problems introduced · 110ed64f
  Dan Gohman authored Sep 01, 2010
  110ed64f
- We have a chance for an optimization. Consider this code: · 6789f8b6
  Bill Wendling authored Aug 31, 2010
  6789f8b6
Aug 31, 2010
- Update test for 112609 · ad9b6de3
  Jim Grosbach authored Aug 31, 2010
  ad9b6de3
- Fix borken test · 3a1d87a7
  Anton Korobeynikov authored Aug 30, 2010
  3a1d87a7
Aug 30, 2010
- Remove NEON vmovn intrinsic, replacing it with vector truncate operations. · 4cd8a126
  Bob Wilson authored Aug 30, 2010
  4cd8a126
- two changes: · 34bfab0a
  Chris Lattner authored Aug 30, 2010
  34bfab0a
- Correct bogus module triple specifications. · 68c30907
  Duncan Sands authored Aug 30, 2010
  68c30907
Aug 29, 2010
- Make IVUsers iterative instead of recursive. · 3a08ed79
  Dan Gohman authored Aug 29, 2010
  3a08ed79
- Make this test less dependent on register allocation choices. · 6665550b
  Dan Gohman authored Aug 29, 2010
  6665550b
- Fix lowering of INSERT_VECTOR_ELT in SPU. · 1e616572
  Kalle Raiskila authored Aug 29, 2010
  1e616572
- Remove NEON vaddl, vaddw, vsubl, and vsubw intrinsics. Instead, use llvm · d0c05488
  Bob Wilson authored Aug 29, 2010
  d0c05488
- merge a bunch of shuffle tests into sse2.ll · c2887bc2
  Chris Lattner authored Aug 29, 2010
  c2887bc2
- add some nounwind's · b1ff9784
  Chris Lattner authored Aug 29, 2010
  b1ff9784
Aug 28, 2010
- fix the buildvector->insertp[sd] logic to not always create a redundant · 94656b1c
  Chris Lattner authored Aug 28, 2010
  94656b1c
- fix the BuildVector -> unpcklps logic to not do pointless shuffles · bcb6090a
  Chris Lattner authored Aug 28, 2010
  bcb6090a