Commits · 4da0ab1d6873333bc5a73b62a7f19a62770703e3 · Lorenzo Albano / LLVM bpEVL · GitLab

Apr 12, 2013
- Add debug prints. · 4da0ab1d
  Nadav Rotem authored Apr 12, 2013
  4da0ab1d
- LoopVectorizer: integer division is not a reduction operation · f9cea17f
  Arnold Schwaighofer authored Apr 12, 2013
  f9cea17f
Apr 11, 2013
- Rename the C function to create a SLPVectorizerPass to something sane and... · c86fdf12
  Benjamin Kramer authored Apr 11, 2013
  c86fdf12
Apr 10, 2013
- Make the SLP store-merger less paranoid about function calls. We check for... · 73dffa41
  Nadav Rotem authored Apr 10, 2013
  73dffa41
- We require DataLayout for analyzing the size of stores. · 88dd5f7a
  Nadav Rotem authored Apr 10, 2013
  88dd5f7a
Apr 09, 2013
- Add support for bottom-up SLP vectorization infrastructure. · 2d9dec32
  Nadav Rotem authored Apr 09, 2013
  2d9dec32
Apr 05, 2013
- LoopVectorizer: Pass OperandValueKind information to the cost model · df6f67ed
  Arnold Schwaighofer authored Apr 04, 2013
  df6f67ed
Mar 14, 2013
- LoopVectorize: Invert case when we use a vector cmp value to query select cost · c63cf3a0
  Arnold Schwaighofer authored Mar 14, 2013
  c63cf3a0
Mar 10, 2013
- BBVectorize: Fixup debugging statements · f610be9f
  Hal Finkel authored Mar 10, 2013
  f610be9f
Mar 09, 2013
- Remove a source of nondeterminism from the LoopVectorizer. · 6eda79f6
  Benjamin Kramer authored Mar 09, 2013
  6eda79f6
- LoopVectorizer: Ignore all dbg intrinisic · 8b3dc094
  Arnold Schwaighofer authored Mar 09, 2013
  8b3dc094
- LoopVectorizer: Ignore dbg.value instructions · 4090b61a
  Arnold Schwaighofer authored Mar 09, 2013
  4090b61a
Mar 08, 2013
- Insert the reduction start value into the first bypass block to preserve domination. · 37c2d65c
  Benjamin Kramer authored Mar 08, 2013
  37c2d65c
Mar 02, 2013
- PR14448 - prevent the loop vectorizer from vectorizing the same loop twice. · 739e37a0
  Nadav Rotem authored Mar 02, 2013
  739e37a0
Mar 01, 2013
- LoopVectorize: Don't hang forever if a PHI only has skipped PHI uses. · 12f98fae
  Benjamin Kramer authored Mar 01, 2013
  12f98fae
Feb 27, 2013
- LoopVectorize: Vectorize math builtin calls. · dc145816
  Benjamin Kramer authored Feb 27, 2013
  dc145816
Feb 21, 2013
- Allow GlobalValues to vectorize with AliasAnalysis · cf928cb5
  Renato Golin authored Feb 21, 2013
  cf928cb5
Feb 17, 2013
- BBVectorize: Fix an invalid reference bug · 76e65e45
  Hal Finkel authored Feb 17, 2013
  76e65e45
Feb 15, 2013
- BBVectorize: Call a DAG and DAG instead of a tree · 89909397
  Hal Finkel authored Feb 15, 2013
  89909397
- BBVectorize: Cap the number of candidate pairs in each instruction group · 283f4f0e
  Hal Finkel authored Feb 15, 2013
  283f4f0e
Feb 14, 2013
- BBVectorize: Remove the remaining instances of std::multimap · e7a1ef42
  Hal Finkel authored Feb 14, 2013
  e7a1ef42
- BBVectorize: Don't store candidate pairs in a std::multimap · c3a4425c
  Hal Finkel authored Feb 14, 2013
  c3a4425c
Feb 13, 2013
- LoopVectorize: Simplify code for clarity. · 0aa2ad61
  Benjamin Kramer authored Feb 13, 2013
  0aa2ad61
- Metadata for annotating loops as parallel. The first consumer for this · 0d23725a
  Pekka Jaaskelainen authored Feb 13, 2013
  0d23725a
Feb 12, 2013
- BBVectorize: Don't over-search when building the dependency map · 6ae564b4
  Hal Finkel authored Feb 11, 2013
  6ae564b4
- BBVectorize: Omit unnecessary entries in PairableInstUsers · 39a95032
  Hal Finkel authored Feb 11, 2013
  39a95032
Feb 11, 2013
- BBVectorize: Eliminate one more restricted linear search · 0b8ae895
  Hal Finkel authored Feb 11, 2013
  0b8ae895
- BBVectorize: Remove the linear searches from pair connection searching · cb268f79
  Hal Finkel authored Feb 11, 2013
  cb268f79
- BBVectorize: Avoid linear searches within the load-move set · fee38f97
  Hal Finkel authored Feb 11, 2013
  fee38f97
- BBVectorize: isa/cast cleanup in getInstructionTypes · dd4bc665
  Hal Finkel authored Feb 11, 2013
  dd4bc665
- BBVectorize: Make the bookkeeping to support full cycle checking less expensive · c1cc1669
  Hal Finkel authored Feb 11, 2013
  c1cc1669
Feb 08, 2013
- BBVectorize: Use TTI->getAddressComputationCost · dd272184
  Hal Finkel authored Feb 08, 2013
  dd272184
- Typos. · 479e5a93
  Jakob Stoklund Olesen authored Feb 08, 2013
  479e5a93
- ARM cost model: Address computation in vector mem ops not free · 594fa2dc
  Arnold Schwaighofer authored Feb 08, 2013
  594fa2dc
- Test Commit · f63b77be
  Michael Kuperstein authored Feb 08, 2013
  f63b77be
Feb 07, 2013
- fix 80-col violation and fix the docs. · a9100f36
  Nadav Rotem authored Feb 07, 2013
  a9100f36
- Loop Vectorizer: Refactor Memory Cost Computation · 3476fc8c
  Arnold Schwaighofer authored Feb 07, 2013
  3476fc8c
Feb 05, 2013
- Loop Vectorizer: Refactor code to compute vectorized memory instruction cost · 3be40b56
  Arnold Schwaighofer authored Feb 05, 2013
  3be40b56
- Loop Vectorizer: Handle pointer stores/loads in getWidestType() · 22174f5d
  Arnold Schwaighofer authored Feb 05, 2013
  22174f5d
Jan 29, 2013
- LoopVectorize: convert TinyTripCountVectorThreshold constant · f50ab84b
  Pekka Jaaskelainen authored Jan 29, 2013
  f50ab84b