| 2010-03-01 |
Chris Lattner | remove a little hack I did for the old isel, not needed
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | resolve some fixmes
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | resolve a fixme and simplify code by moving insertion...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | resolve a fixme by having the .td file parser reject...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | remove all but one version of SelectionDAG::MorphNodeTo
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | remove dead code, simplify.
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | Fix PR2590 by making PatternSortingPredicate actually be
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | tolerate factoring the *last* node for CellSPU.
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | optimize tblgen compile time by eliminating the old...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | remove a terrible hack that disabled assertions from...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | stop using anders-aa
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | stop using anders-aa
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | remove andersen's tests.
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | don't emit the old sdnodexform stuff for the new isel.
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | stop using generated sdnodexforms.
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | note that andersaa was removed in 2.7
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | remove anders-aa from mainline, it isn't maintained...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | Turn on the new isel by default. Here are some fun...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | Accelerate isel dispatch for tables that start with...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | Emit type checks late instead of early, this encourages
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | some trivial microoptimizations.
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | Emit a redundant check for immediates at root context...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | eliminate the CheckMultiOpcodeMatcher code and have...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | add a new OPC_SwitchOpcode which is semantically equivalent
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | pull MarkFlagResult out from between an EmitNode/CompleteMatch
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | enhance RecordNode and RecordChild comments to indicate...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | Emit redundant opcode checks for andimm and orimm tests...
|
commit | commitdiff | tree |
| 2010-03-01 |
Chris Lattner | inline the node transforms and node predicates into...
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | simplify some code now that chain/flag results are...
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | don't emit useless functions. These were producing
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | eliminate GetInt1/2
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | hoist the new isel interpreter out of DAGISelHeader.h
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | change a few opcodes to use VBRs instead of embedding
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | move all the opcodes out of DAGISelHeader.h (which...
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | resolve a fixme by using hte new flag.
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | enhance the EmitNode/MorphNodeTo operands to take a...
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | enhance the new isel to handle the 'node already exists'
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | use MorphNodeTo instead of SelectNodeTo. SelectNodeTo
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | enhance the new isel to use SelectNodeTo for most patterns,
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | add some random nounwinds.
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | I'm removing andersaa after the branch is made.
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | simplify this code, return only ever has zero or one...
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | remove redundant instruction.
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | enhance EmitNodeMatcher to keep track of the recorded...
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | add infrastructure to support forming selectnodeto...
|
commit | commitdiff | tree |
| 2010-02-28 |
Chris Lattner | Generalize my hack to use SDNodeInfo to find out when a
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | change CheckOpcodeMatcher to hold the SDNodeInfo instead of
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | remove a bogus pattern, which had the same pattern...
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | add another case from the ppc backend. This is obviously...
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | fix an incorrect (overly conservative) predicate.
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | fix logic in DEBUG.
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | teach the optimizer that opcode == ISD::STORE is contradictory
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | fix grammaro's pointed out by daniel
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | Teach the grouper some simple tricks about looking...
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | use DEBUG instead of DebugFlag directly so that this
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | implement a new optimization to sink pattern predicates...
|
commit | commitdiff | tree |
| 2010-02-27 |
Chris Lattner | fix PR6414, a nondeterminism issue in IPSCCP which...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | Fix rdar://7694996 a miscompile of 183.equake from...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | remove dead code, by this point all uses of CI are...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | don't build edis if the x86 target isn't enabld.
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | pass in more section kinds, enough to get the .align...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | fix PR6435 another bug from the MallocInst elimination...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | add some helpful comments to the emitter
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | switch from my nice hashtable based merging solution...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | fix same bug in CheckChainCompatibleMatcher::isEqualImpl
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | fix a nasty bug in CheckTypeMatcher::isEqualImpl
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | finish off the factoring optimization along the lines...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | add a new setNumChildren method for resizing scopes...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | fix the matcher in the presence of multiple scopes...
|
commit | commitdiff | tree |
| 2010-02-26 |
Chris Lattner | this file lacks a run line!
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | rewrite OptimizeGlobalAddressOfMalloc to fix PR6422...
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | change the scope node to include a list of children...
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | remove a dead PatLeaf, I previously changed all uses...
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | Implement the first half of redundancy factoring: efficiently
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | factor the print method better.
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | add methods to do equality checks and get hashes of...
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | formatting.
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | rename fooMatcherNode to fooMatcher.
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | add some noop code to push it out of my tree.
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | rename PushMatcherNode -> ScopeMatcherNode to more...
|
commit | commitdiff | tree |
| 2010-02-25 |
Chris Lattner | add a fixme for an experiment that defeated me for...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | fix a bug I introduced by checking the wrong node's...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | clean up various VT manipulations, patch by Micah Villmow...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | fix PR5954, patch by Roman Divacky.
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | convert cycle checker to smallptrset, add comments...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | revert david's patch which does not even build.
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | contract movechild+checktype into a new checkchild...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | split the movechild/record/moveparent -> recordchild...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | emit a histogram of the opcodes in comments.
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | lets not break the old isel.
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | Since the new instruction selector now works, I don...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | implement a simple proof-of-concept optimization for
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | The new isel passes all tests, time to start making...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | Change the scheduler from adding nodes in allnodes...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | The new isel was not properly handling patterns that...
|
commit | commitdiff | tree |
| 2010-02-24 |
Chris Lattner | add node #'s to debug dumps.
|
commit | commitdiff | tree |
| 2010-02-23 |
Chris Lattner | fix X86/uint_to_fp-2.ll, only fold loads when they...
|
commit | commitdiff | tree |
| 2010-02-23 |
Chris Lattner | make selectnodeto set the nodeid to -1. This makes...
|
commit | commitdiff | tree |
| 2010-02-23 |
Chris Lattner | no need to override IsLegalToFold, the base implementation
|
commit | commitdiff | tree |
| 2010-02-23 |
Chris Lattner | fix a bug in findNonImmUse (used by IsLegalToFold)...
|
commit | commitdiff | tree |
| next |
projects / oota-llvm.git / search