summaryrefslogtreecommitdiffstats
path: root/llvm
Commit message (Collapse)AuthorAgeFilesLines
* Fix a bug in a recent refactoring that broke a bunch of stuff.Chris Lattner2006-08-121-1/+1
| | | | llvm-svn: 29649
* Fix an obvious bug, noticed by inspection. No current targets trigger this.Chris Lattner2006-08-121-3/+3
| | | | llvm-svn: 29648
* Don't attempt to split subloops out of a loop with a huge number of backedges.Chris Lattner2006-08-121-8/+19
| | | | | | | | | Not only will this take huge amounts of compile time, the resultant loop nests won't be useful for optimization. This reduces loopsimplify time on Transforms/LoopSimplify/2006-08-11-LoopSimplifyLongTime.ll from ~32s to ~0.4s with a debug build of llvm on a 2.7Ghz G5. llvm-svn: 29647
* New testcase, this used to take hours to loopsimplify.Chris Lattner2006-08-122-0/+5
| | | | llvm-svn: 29646
* Make Loop::getExitBlocks significantly faster for large loops. Instead ofChris Lattner2006-08-121-3/+9
| | | | | | | | pounding on Loop::contains (which is O(n) in the size of the loop), use a sorted vector, which is O(log(N)) for each query. This speeds up Duraid's horrible testcase from ~72s to ~31s in a debug build. llvm-svn: 29645
* Reimplement the loopsimplify code which deletes edges from unreachableChris Lattner2006-08-121-29/+53
| | | | | | | | | | | | | | | | | | | blocks that target loop blocks. Before, the code was run once per loop, and depended on the number of predecessors each block in the loop had. Unfortunately, scanning preds can be really slow when huge numbers of phis exist or when phis with huge numbers of inputs exist. Now, the code is run once per function and scans successors instead of preds, which is far faster. In addition, the new code is simpler and is goto free, woo. This change speeds up a nasty testcase Duraid provided me from taking hours to taking ~72s with a debug build. The functionality this implements is already tested in the testsuite as Transforms/CodeExtractor/2004-03-13-LoopExtractorCrash.ll. llvm-svn: 29644
* Track # nodes in a CSEMap.Chris Lattner2006-08-121-0/+1
| | | | llvm-svn: 29643
* Switch to using SuperFastHash instead of adding all elements together. ThisChris Lattner2006-08-121-6/+24
| | | | | | doesn't significantly improve performance but it helps a small amount. llvm-svn: 29642
* Lexer.cpp is a generated file both in lib/AsmParser and projects/StackerReid Spencer2006-08-111-0/+1
| | | | | | so we remove it from being counted. llvm-svn: 29641
* Switch NodeID to track 32-bit chunks instead of 8-bit chunks, for a 2.5%Chris Lattner2006-08-112-31/+18
| | | | | | speedup in isel time. llvm-svn: 29640
* remove IncludeFile turds in MathExtras.h, which bloats every .o file thatChris Lattner2006-08-113-16/+9
| | | | | | #includes it. llvm-svn: 29639
* Weed out some cruft and add in some missing extensions.Reid Spencer2006-08-111-18/+40
| | | | llvm-svn: 29638
* Weed out cruft that should not be counted and add in extensions weReid Spencer2006-08-111-106/+50
| | | | | | previously missed. llvm-svn: 29637
* move code out of line so that GCC doesn't inline it at -O3Chris Lattner2006-08-111-61/+68
| | | | llvm-svn: 29636
* Added information for the new nightly tester.Patrick Jenkins2006-08-111-19/+35
| | | | llvm-svn: 29635
* Split SmallVector into SmallVector and SmallVectorImpl, which allows us toChris Lattner2006-08-111-40/+45
| | | | | | eliminate code duplication due to the 'N' parameter. llvm-svn: 29634
* Added two flags (-submit-server and -submit-script) which allow the user to ↵Patrick Jenkins2006-08-111-5/+18
| | | | | | specify where the test results are submitted too. If these are not specified they default to the LLVM.org server. llvm-svn: 29633
* Make this script not care where it is run from by getting the llvm srcReid Spencer2006-08-111-1/+1
| | | | | | root from the llvm-config script. llvm-svn: 29632
* Remove 8 more std::map's.Chris Lattner2006-08-113-108/+74
| | | | llvm-svn: 29631
* Make llvmdo and llvmgrep invulnerable to where they are run from by gettingReid Spencer2006-08-112-5/+11
| | | | | | | | the llvm source root from the llvm-config command. The dependency now is that the correct llvm-config command is in the path. For most developers this is the case. llvm-svn: 29630
* Don't print double newline for --src-root and --obj-root as it tends toReid Spencer2006-08-111-2/+2
| | | | | | confuse shells that might use this output. llvm-svn: 29629
* remove old piece of the V9 backend.Chris Lattner2006-08-111-39/+0
| | | | llvm-svn: 29628
* eliminate extraneous blank lineChris Lattner2006-08-111-1/+0
| | | | llvm-svn: 29627
* Move the BBNodes, GlobalValues, TargetGlobalValues, Constants, TargetConstants,Chris Lattner2006-08-112-88/+83
| | | | | | RegNodes, and ValueNodes maps into the CSEMap. llvm-svn: 29626
* Start moving leaf nodes over to the CSEMap.Chris Lattner2006-08-112-14/+20
| | | | llvm-svn: 29625
* Clean up the HTML output. Add a -filedetails option to report the number ofReid Spencer2006-08-111-50/+61
| | | | | | lines contributed by each user to each file. llvm-svn: 29624
* Remove blank line at end of file, removing test user "foo" from CVS history.Reid Spencer2006-08-111-1/+0
| | | | llvm-svn: 29623
* new method for adding intsChris Lattner2006-08-111-0/+6
| | | | llvm-svn: 29622
* eliminate the NullaryOps map, use CSEMap instead.Chris Lattner2006-08-112-24/+18
| | | | llvm-svn: 29621
* Drop the -details option, it was broken. If you want to know about aReid Spencer2006-08-111-54/+42
| | | | | | | | particular directory, just pass that directory to the script. Also, clean up the HTML output, put list of skipped dirs at the end, and make a few other minor improvements. llvm-svn: 29620
* Use SmallVector instead.Evan Cheng2006-08-111-1/+1
| | | | llvm-svn: 29619
* change internal impl of dag combiner so that calls to CombineTo never have toChris Lattner2006-08-111-12/+9
| | | | | | make a temporary vector. llvm-svn: 29618
* Change one ReplaceAllUsesWith method to take an array of operands to ↵Chris Lattner2006-08-111-1/+1
| | | | | | replaceinstead of a vector of operands. llvm-svn: 29617
* Change one ReplaceAllUsesWith method to take an array of operands to replaceChris Lattner2006-08-112-6/+5
| | | | | | instead of a vector of operands. llvm-svn: 29616
* Remove now-dead method.Chris Lattner2006-08-111-4/+0
| | | | llvm-svn: 29615
* Eliminate use of getNode that takes a vector.Chris Lattner2006-08-115-33/+47
| | | | llvm-svn: 29614
* remove a (now) unused getNode method.Chris Lattner2006-08-111-5/+0
| | | | llvm-svn: 29613
* elimiante use of getNode that takes vector of operands.Chris Lattner2006-08-111-3/+4
| | | | llvm-svn: 29612
* eliminate use of getNode that takes vector of operands.Chris Lattner2006-08-111-2/+4
| | | | llvm-svn: 29611
* eliminate use of getNode that takes vector<SDOperand>. Wrap a really long line.Chris Lattner2006-08-111-4/+6
| | | | llvm-svn: 29610
* Convert vectors to fixed sized arrays and smallvectors. Eliminate use of ↵Chris Lattner2006-08-111-37/+42
| | | | | | getNode that takes a vector. llvm-svn: 29609
* Fix miscompilation of float vector returns. Compile code to this:Chris Lattner2006-08-111-4/+4
| | | | | | | | | | | | | | | | | _func: vsldoi v2, v3, v2, 12 vsldoi v2, v2, v2, 4 blr instead of: _func: vsldoi v2, v3, v2, 12 vsldoi v2, v2, v2, 4 *** vor f1, v2, v2 blr llvm-svn: 29607
* New testcaseChris Lattner2006-08-111-0/+8
| | | | llvm-svn: 29606
* Minor cleanupsChris Lattner2006-08-111-4/+4
| | | | llvm-svn: 29605
* Match tablegen changes.Evan Cheng2006-08-117-129/+118
| | | | llvm-svn: 29604
* CALLSEQ_* produces chain even if that's not needed.Evan Cheng2006-08-115-9/+17
| | | | llvm-svn: 29603
* - Prevent some functions from being inlined to eliminate the code size bloatEvan Cheng2006-08-111-67/+96
| | | | | | | | | | | | | introduced by previous commit. - SelectCode now returns a SDNode*. If it is not null, the selected node produces the same number of results as the input node. The seletion loop is responsible for calling ReplaceAllUsesWith() to replace the input node with the output target node. For other cases, e.g. when load is folded, the selection code is responsible for calling ReplaceAllUsesOfValueWith() and SelectCode returns NULL. - Other clean ups. llvm-svn: 29602
* Convert more calls of getNode() that takes a vector to pass in the start of ↵Evan Cheng2006-08-111-13/+14
| | | | | | an array. llvm-svn: 29601
* Fix a typo noticed by Gordon Henriksen. Thanks!Reid Spencer2006-08-101-1/+1
| | | | llvm-svn: 29597
* llvm2cpp no longer reads LLVM assembly but LLVM bytecode instead. AdjustReid Spencer2006-08-101-6/+6
| | | | | | the documentation to reflect this. llvm-svn: 29596
OpenPOWER on IntegriCloud