Commits · 3a554ac06501cf7175ee8d9f94aaa6140e9dd46e · Roger Ferrer / llvm-epi · GitLab

Dec 30, 2012
- Add a check to the test Analysis/ScalarEvolution/2010-09-03-RequiredTransitive.ll · 10c4b4d2
  Dmitri Gribenko authored Dec 30, 2012
  10c4b4d2
- Tests: rewrite 'opt ... %s' to 'opt ... < %s' so that opt does not emit a ModuleID · b137c9e5
  Dmitri Gribenko authored Dec 30, 2012
  b137c9e5
- llvm/test/Transforms/GVN/null-aliases-nothing.ll: Fix a RUN line not to emit ModuleID. · 5a495a5c
  NAKAMURA Takumi authored Dec 30, 2012
  5a495a5c
Dec 28, 2012
- Fix a stunning oversight in the inline cost analysis. It was never · 86ed5308
  Chandler Carruth authored Dec 28, 2012
  86ed5308
- Teach the inline cost analysis about calls that can be simplified and · 753e21d0
  Chandler Carruth authored Dec 28, 2012
  753e21d0
- Teach instsimplify to use the constant folder where appropriate for · f6182155
  Chandler Carruth authored Dec 28, 2012
  f6182155
- AVX: Move the ZEXT/ANYEXT DAGCo optimizations to the lowering of these... · 3da9ac72
  Nadav Rotem authored Dec 28, 2012
  3da9ac72
Dec 27, 2012
- [ASan] Fix lifetime intrinsics handling. Now for each intrinsic we check if it... · 29dd7f20
  Alexey Samsonov authored Dec 27, 2012
  29dd7f20
- On AVX/AVX2 the type v8i1 is legalized to v8i16, which is an XMM sized · 2a054b44
  Nadav Rotem authored Dec 27, 2012
  2a054b44
- For the dwarf5 split debug info code split out the string section · 3bf29fda
  Eric Christopher authored Dec 27, 2012
  3bf29fda
- FileCheck-ize. · c8a88ee6
  Eric Christopher authored Dec 27, 2012
  c8a88ee6
- FileCheck-ize. · d6152aab
  Eric Christopher authored Dec 27, 2012
  d6152aab
- Right now all of the relocations are 32-bit dwarf, and the relocation · 5a6acfa4
  Eric Christopher authored Dec 27, 2012
  5a6acfa4
- If all of the write objects are identified then we can vectorize the loop even... · 5350cd31
  Nadav Rotem authored Dec 26, 2012
  5350cd31
Dec 26, 2012
- LoopVectorizer: Optimize the vectorization of consecutive memory access when... · 3f7c4f36
  Nadav Rotem authored Dec 26, 2012
  3f7c4f36
- [msan] Raise alignment of origin stores/loads when possible. · 5eb5bf8b
  Evgeniy Stepanov authored Dec 26, 2012
  5eb5bf8b
- llvm/test/CodeGen/X86: FileCheck-ize two tests in r171083. · 40aa3285
  NAKAMURA Takumi authored Dec 26, 2012
  40aa3285
- llvm/test/CodeGen/X86: Disable avx in two tests corresponding to r171082. · 334f6853
  NAKAMURA Takumi authored Dec 26, 2012
  334f6853
- BBVectorize: Use VTTI to compute costs for intrinsics vectorization · 30e95a8e
  Hal Finkel authored Dec 26, 2012
  30e95a8e
- LoopVectorize: Enable vectorization of the fmuladd intrinsic · b44f8901
  Hal Finkel authored Dec 25, 2012
  b44f8901
Dec 25, 2012
- BBVectorize: Enable vectorization of the fmuladd intrinsic · 2a456112
  Hal Finkel authored Dec 25, 2012
  2a456112
- Loosen scheduling restrictions on the PPC dcbt intrinsic · 2ebe6d08
  Hal Finkel authored Dec 25, 2012
  2ebe6d08
- Expand PPC64 atomic load and store · 1b5ff08d
  Hal Finkel authored Dec 25, 2012
  1b5ff08d
- [msan] Fix handling of vectors of pointers. · f19c086d
  Evgeniy Stepanov authored Dec 25, 2012
  f19c086d
- [msan] Fix handling of select with vector condition. · ec837128
  Evgeniy Stepanov authored Dec 25, 2012
  ec837128
- Harden test so it's not affected by changes to compare lowering. · a9f265ee
  Benjamin Kramer authored Dec 25, 2012
  a9f265ee
- X86: Shave off one shuffle from the pcmpeqq sequence for SSE2 by making use of and commutativity. · 81b5a8fd
  Benjamin Kramer authored Dec 25, 2012
  81b5a8fd
- X86: Custom lower <2 x i64> eq and ne when SSE41 is not available. · df4af41b
  Benjamin Kramer authored Dec 25, 2012
  df4af41b
Dec 24, 2012
- Fix typo "Makre" -> "Make". · fb432580
  Nick Lewycky authored Dec 24, 2012
  fb432580
- llvm/test/CodeGen/X86/fold-vex.ll: Add explicit triple. · 1b18db7e
  NAKAMURA Takumi authored Dec 24, 2012
  1b18db7e
- Some x86 instructions can load/store one of the operands to memory. On SSE,... · dc0ad92b
  Nadav Rotem authored Dec 24, 2012
  dc0ad92b
- LoopVectorizer: When checking for vectorizable types, also check · 5f7c12cf
  Nadav Rotem authored Dec 24, 2012
  5f7c12cf
- LoopVectorizer: Fix an endless loop in the code that looks for reductions. · bd5d1d83
  Nadav Rotem authored Dec 24, 2012
  bd5d1d83
Dec 23, 2012
- CostModel: Change the default target-independent implementation for finding · cf9999d9
  Nadav Rotem authored Dec 23, 2012
  cf9999d9
- We are not ready to estimate the cost of integer expansions based on the... · aa92ea4f
  Nadav Rotem authored Dec 23, 2012
  aa92ea4f
- Loop Vectorizer: Update the cost model of scatter/gather operations and make · 2cade680
  Nadav Rotem authored Dec 23, 2012
  2cade680
Dec 22, 2012
- X86: Turn mul of <4 x i32> into pmuludq when no SSE4.1 is available. · 76268ac6
  Benjamin Kramer authored Dec 22, 2012
  76268ac6
- X86: Emit vector sext as shuffle + sra if vpmovsx is not available. · b2f0a2bd
  Benjamin Kramer authored Dec 22, 2012
  b2f0a2bd
- In some cases, due to scheduling constraints we copy the EFLAGS. · d5aae980
  Nadav Rotem authored Dec 21, 2012
  d5aae980
- [mips] Fix encoding of BAL instruction. Also, fix assembler test case which · d6b694f0
  Akira Hatanaka authored Dec 21, 2012
  d6b694f0