summaryrefslogtreecommitdiffstats
path: root/llvm/lib
Commit message (Collapse)AuthorAgeFilesLines
* Add explicit keywords.Dan Gohman2008-03-2516-18/+19
| | | | llvm-svn: 48801
* A quick nm audit turned up several fixed tables and objects that wereDan Gohman2008-03-259-14/+16
| | | | | | | marked read-write. Use const so that they can be allocated in a read-only segment. llvm-svn: 48800
* Avoid outputing spaces at the ends of lines.Dan Gohman2008-03-251-4/+4
| | | | llvm-svn: 48797
* Do not align loops if optimizing for size.Devang Patel2008-03-251-1/+1
| | | | llvm-svn: 48794
* Add optimize-for-size knob.Devang Patel2008-03-251-0/+6
| | | | llvm-svn: 48793
* Handle a special case xor undef, undef -> 0. Technically this should be ↵Evan Cheng2008-03-253-2/+15
| | | | | | transformed to undef. But this is such a common idiom (misuse) we are going to handle it. llvm-svn: 48792
* Handle a special case xor undef, undef -> 0. Technically this should be ↵Evan Cheng2008-03-251-1/+6
| | | | | | transformed to undef. But this is such a common idiom (misuse) we are going to handle it. llvm-svn: 48791
* Fix typos.Dan Gohman2008-03-251-3/+3
| | | | llvm-svn: 48779
* Add CMP32mr and friends to the load-unfolding table. AmongDan Gohman2008-03-251-0/+4
| | | | | | | | other things, this allows the scheduler to unfold a load operand in the 2008-01-08-SchedulerCrash.ll testcase, so it now successfully clones the comparison to avoid a pushf+popf. llvm-svn: 48777
* Fix PR2062: Don't build Intrinsics.gen in the source directory. Do it in theBill Wendling2008-03-251-2/+2
| | | | | | object directory. llvm-svn: 48766
* lastRegisterUse() should ignore identity copies. Those will be erased.Evan Cheng2008-03-251-7/+14
| | | | llvm-svn: 48759
* Remove an unneeded test.Evan Cheng2008-03-241-1/+1
| | | | llvm-svn: 48755
* Add \t after .set. Fix by Jay Freeman.Evan Cheng2008-03-241-1/+1
| | | | llvm-svn: 48753
* If the coalescer commuted a def MI to allow coalescing, it can changed a ↵Evan Cheng2008-03-241-1/+2
| | | | | | previously coalesced copy into an non-identity copy. llvm-svn: 48752
* Add an assertion to catch register of illegal class.Evan Cheng2008-03-241-0/+1
| | | | llvm-svn: 48751
* Use the bit size of the operand instead of the hard-coded 32 to generate theBill Wendling2008-03-241-2/+5
| | | | | | mask. llvm-svn: 48750
* - SSE4.1 extractfps extracts a f32 into a gr32 register. Very useful! Not. ↵Evan Cheng2008-03-242-8/+27
| | | | | | Fix the instruction specification and teaches lowering code to use it only when the only use is a store instruction. llvm-svn: 48746
* Revert r48676. I had plans for using it, but now it's just dead code.Owen Anderson2008-03-242-70/+0
| | | | llvm-svn: 48743
* Remove #include<iostream>, which I was using for debugging.Owen Anderson2008-03-241-2/+0
| | | | llvm-svn: 48739
* Add incoming value from header only if phi node has any use inside the loop.Devang Patel2008-03-241-2/+3
| | | | llvm-svn: 48738
* Shrink the size of AllocationInst by using its SubclassDataDan Gohman2008-03-241-4/+10
| | | | | | | field to store the alignment value instead of haing a separate field. llvm-svn: 48727
* APIntify SelectionDAG's EXTRACT_ELEMENT code.Dan Gohman2008-03-241-2/+4
| | | | llvm-svn: 48726
* Increasing the inline limit from (overly conservative) 200 to 300. Given ↵Evan Cheng2008-03-243-9/+38
| | | | | | | | each BB costs 20 and each instruction costs 5, 200 means a 4 BB function + 24 instructions (actually less because caller's size also contributes to it). Furthermore, double the limit when more than 10% of the callee instructions are vector instructions. Multimedia kernels tend to love inlining. llvm-svn: 48725
* Temporarily disabling memset forming optimization. Add an option.Evan Cheng2008-03-241-0/+8
| | | | llvm-svn: 48720
* Be sure to remove intervals after we've joined them. Also, remove some ↵Owen Anderson2008-03-241-46/+18
| | | | | | | | | duplicated code. With this pass, StrongPHIElim can compile very simple testcases correctly. There's still a ways to go before it's ready for prime time, though. llvm-svn: 48719
* Transform (zext (or (icmp), (icmp))) to (or (zext (cimp), (zext icmp))) if ↵Evan Cheng2008-03-241-96/+129
| | | | | | at least one of the (zext icmp) can be transformed to eliminate an icmp. llvm-svn: 48715
* Remove duplicated entries.Evan Cheng2008-03-231-35/+0
| | | | llvm-svn: 48714
* Objective Caml bindings for basic block, function, global, and arg iterators.Gordon Henriksen2008-03-231-4/+36
| | | | llvm-svn: 48711
* Minor typo fixes. Also add another FIXME.Anton Korobeynikov2008-03-231-5/+6
| | | | llvm-svn: 48710
* Preserve calling convention during function cloningAnton Korobeynikov2008-03-231-0/+3
| | | | llvm-svn: 48708
* Add license headerAnton Korobeynikov2008-03-231-0/+13
| | | | llvm-svn: 48707
* Add Win64 compilation callback. This allows easy examples to be JITed on Win64!Anton Korobeynikov2008-03-232-5/+58
| | | | llvm-svn: 48706
* Provide a JIT selector on win64Anton Korobeynikov2008-03-231-1/+1
| | | | llvm-svn: 48704
* Hack out the PIC mode on Win64 targets. This needs to be investigated later.Anton Korobeynikov2008-03-231-1/+3
| | | | llvm-svn: 48703
* Code cleanup. Provide generic way of selecting JIT pointer bitwidth regardlessAnton Korobeynikov2008-03-231-22/+30
| | | | | | of compiler used. llvm-svn: 48702
* Remove old-standing obsolete code.Anton Korobeynikov2008-03-231-9/+11
| | | | llvm-svn: 48701
* Honour built-in defines on win64 targets for automatically subtarget recognize.Anton Korobeynikov2008-03-221-2/+3
| | | | | | Force stack alignment to 16 bytes on win targets. llvm-svn: 48695
* Recognize "windows" in target triple, not only "win32"Anton Korobeynikov2008-03-221-0/+2
| | | | llvm-svn: 48694
* Add information about callee-saved registers on Win64Anton Korobeynikov2008-03-222-6/+28
| | | | llvm-svn: 48692
* Add convenient helper for win64 check. Simplify things slightly.Anton Korobeynikov2008-03-223-3/+5
| | | | llvm-svn: 48691
* Initial support for Win64 calling conventions. Still in early state.Anton Korobeynikov2008-03-222-5/+54
| | | | llvm-svn: 48690
* Register EH frames emitted in JIT when using gcc unwinding runtimeAnton Korobeynikov2008-03-221-1/+9
| | | | llvm-svn: 48688
* Fix merge errorAnton Korobeynikov2008-03-221-1/+0
| | | | llvm-svn: 48687
* Add testcase for prev. commit. Minor fixesAnton Korobeynikov2008-03-221-2/+2
| | | | llvm-svn: 48686
* Aliasing chains cleanups: update langref, add check into verifierAnton Korobeynikov2008-03-221-1/+6
| | | | llvm-svn: 48685
* Support chained aliases for LLVM IR printing. This fixes PR2145Anton Korobeynikov2008-03-221-0/+3
| | | | llvm-svn: 48684
* Another comments fixingAnton Korobeynikov2008-03-222-2/+3
| | | | llvm-svn: 48683
* Slightly increase default set size. It's cheap and won't hurt.Anton Korobeynikov2008-03-221-1/+1
| | | | llvm-svn: 48682
* implement an initial hack at a straight-line store -> memset optimization.Chris Lattner2008-03-221-6/+59
| | | | | | | | This fires dozens of times across spec and multisource, but I don't know if it actually speeds stuff up. Hopefully the testers will show something nice :) llvm-svn: 48680
* implement the logic for memset insertion and store deletion.Chris Lattner2008-03-221-11/+49
| | | | llvm-svn: 48679
OpenPOWER on IntegriCloud