summaryrefslogtreecommitdiff
path: root/lib/Transforms/Vectorize
AgeCommit message (Expand)AuthorFilesLines
2013-04-19LoopVectorizer: Use matcher from PatternMatch.h for the min/max patternsArnold Schwaighofer1-104/+102
2013-04-18Fix a -Wdocumentation warningDmitri Gribenko1-1/+1
2013-04-18LoopVectorizer: Recognize min/max reductionsArnold Schwaighofer1-34/+209
2013-04-18LoopVectorize: Use a set to avoid longer cycles in the reduction chain too.Benjamin Kramer1-8/+6
2013-04-15SLPVectorizer: Make it a function pass and add code for hoisting the vector-g...Nadav Rotem3-159/+254
2013-04-15SLPVectorizer: Add support for vectorizing trees that start at compare instru...Nadav Rotem1-21/+40
2013-04-14Miscellaneous cleanups for VecUtils.hBenjamin Kramer1-9/+6
2013-04-14SLP: Document the scalarization cost method.Nadav Rotem1-3/+10
2013-04-14SLPVectorizer: Add support for trees that don't start at binary operators, an...Nadav Rotem3-7/+25
2013-04-14SLPVectorizer: add initial support for reduction variable vectorization.Nadav Rotem3-7/+95
2013-04-12SLPVectorizer: add support for vectorization of diamond shaped trees. We now ...Nadav Rotem2-46/+254
2013-04-12Add debug prints.Nadav Rotem1-1/+5
2013-04-12LoopVectorizer: integer division is not a reduction operationArnold Schwaighofer1-2/+0
2013-04-11Rename the C function to create a SLPVectorizerPass to something sane and exp...Benjamin Kramer1-2/+2
2013-04-10Make the SLP store-merger less paranoid about function calls. We check for fu...Nadav Rotem1-4/+0
2013-04-10We require DataLayout for analyzing the size of stores.Nadav Rotem2-1/+6
2013-04-09Add support for bottom-up SLP vectorization infrastructure.Nadav Rotem5-0/+707
2013-04-04LoopVectorizer: Pass OperandValueKind information to the cost modelArnold Schwaighofer1-2/+13
2013-03-14LoopVectorize: Invert case when we use a vector cmp value to query select costArnold Schwaighofer1-1/+1
2013-03-10BBVectorize: Fixup debugging statementsHal Finkel1-2/+2
2013-03-09Remove a source of nondeterminism from the LoopVectorizer.Benjamin Kramer1-1/+1
2013-03-09LoopVectorizer: Ignore all dbg intrinisicArnold Schwaighofer1-6/+6
2013-03-09LoopVectorizer: Ignore dbg.value instructionsArnold Schwaighofer1-2/+11
2013-03-08Insert the reduction start value into the first bypass block to preserve domi...Benjamin Kramer1-1/+1
2013-03-02PR14448 - prevent the loop vectorizer from vectorizing the same loop twice.Nadav Rotem1-0/+18
2013-03-01LoopVectorize: Don't hang forever if a PHI only has skipped PHI uses.Benjamin Kramer1-1/+8
2013-02-27LoopVectorize: Vectorize math builtin calls.Benjamin Kramer1-50/+137
2013-02-21Allow GlobalValues to vectorize with AliasAnalysisRenato Golin1-35/+154
2013-02-17BBVectorize: Fix an invalid reference bugHal Finkel1-4/+7
2013-02-15BBVectorize: Call a DAG and DAG instead of a treeHal Finkel1-84/+84
2013-02-15BBVectorize: Cap the number of candidate pairs in each instruction groupHal Finkel1-1/+9
2013-02-14BBVectorize: Remove the remaining instances of std::multimapHal Finkel1-231/+256
2013-02-14BBVectorize: Don't store candidate pairs in a std::multimapHal Finkel1-60/+92
2013-02-13LoopVectorize: Simplify code for clarity.Benjamin Kramer1-10/+8
2013-02-13Metadata for annotating loops as parallel. The first consumer for this Pekka Jaaskelainen1-0/+8
2013-02-11BBVectorize: Don't over-search when building the dependency mapHal Finkel1-2/+10
2013-02-11BBVectorize: Omit unnecessary entries in PairableInstUsersHal Finkel1-1/+3
2013-02-11BBVectorize: Eliminate one more restricted linear searchHal Finkel1-27/+31
2013-02-11BBVectorize: Remove the linear searches from pair connection searchingHal Finkel1-24/+11
2013-02-11BBVectorize: Avoid linear searches within the load-move setHal Finkel1-20/+30
2013-02-11BBVectorize: isa/cast cleanup in getInstructionTypesHal Finkel1-4/+4
2013-02-11BBVectorize: Make the bookkeeping to support full cycle checking less expensiveHal Finkel1-14/+25
2013-02-08BBVectorize: Use TTI->getAddressComputationCostHal Finkel1-0/+5
2013-02-08Typos.Jakob Stoklund Olesen1-4/+4
2013-02-08ARM cost model: Address computation in vector mem ops not freeArnold Schwaighofer1-8/+14
2013-02-08Test CommitMichael Kuperstein1-1/+1
2013-02-07fix 80-col violation and fix the docs.Nadav Rotem1-3/+7
2013-02-07Loop Vectorizer: Refactor Memory Cost ComputationArnold Schwaighofer1-180/+52
2013-02-05Loop Vectorizer: Refactor code to compute vectorized memory instruction costArnold Schwaighofer1-79/+178
2013-02-05Loop Vectorizer: Handle pointer stores/loads in getWidestType()Arnold Schwaighofer1-9/+31