summaryrefslogtreecommitdiffstats
path: root/llvm
Commit message (Collapse)AuthorAgeFilesLines
* Also requires -mattr=-sse3Evan Cheng2006-03-153-3/+3
| | | | llvm-svn: 26778
* allow field sensitivity to be a tunable parameterAndrew Lenharth2006-03-151-2/+6
| | | | llvm-svn: 26777
* Update scheduling info for vrsave instructionNate Begeman2006-03-151-2/+2
| | | | llvm-svn: 26776
* Handle one offset with growth case seen in povray. Namely, if we have an ↵Andrew Lenharth2006-03-151-10/+39
| | | | | | | | | offset, and the offset lands at a field boundary in the old type, construct a new type, copying the fields masked by the offset from the old type, and unify with that. llvm-svn: 26775
* improve mem intrinsics and add a few things povray usesAndrew Lenharth2006-03-151-3/+30
| | | | llvm-svn: 26774
* Fix VC++ build error.Jeff Cohen2006-03-151-0/+1
| | | | llvm-svn: 26773
* remove typoChris Lattner2006-03-151-1/+1
| | | | llvm-svn: 26772
* Autogenerate a table of intrinsic names, so we can map from intrinsic ID toChris Lattner2006-03-152-4/+22
| | | | | | LLVM intrinsic function name. llvm-svn: 26771
* Autogenerate code to map from GCC builtin to LLVM intrinsic.Chris Lattner2006-03-153-1/+59
| | | | llvm-svn: 26770
* Set TargetPrefix on target-specific intrinsics. That way, in theory, differentChris Lattner2006-03-151-7/+7
| | | | | | targets could have different implemenations of a __builtin_foo gcc intrinsic. llvm-svn: 26769
* Fixed a typo in the vsetfp examples.Robert Bocchino2006-03-141-6/+6
| | | | llvm-svn: 26768
* Split the 'vset' instruction into two instructions, 'vsetint' andRobert Bocchino2006-03-141-38/+73
| | | | | | | 'vsetfp', to reflect the fact that the semantics are different for integer and fp values. llvm-svn: 26767
* Update this to match the documentation in the GC doc and to match actualChris Lattner2006-03-141-6/+10
| | | | | | practice. llvm-svn: 26766
* Fix the gcread/gcwrite intrinsic specifications, fixingChris Lattner2006-03-141-5/+5
| | | | | | CodeGen/Generic/GC/alloc_loop.ll llvm-svn: 26765
* Bugfix, unbreaking CodeGen/PowerPC/cttz.llChris Lattner2006-03-141-1/+1
| | | | llvm-svn: 26764
* Fix an et-forest memory leak. Patch by Daniel Berlin.Chris Lattner2006-03-141-0/+1
| | | | llvm-svn: 26763
* add a noteChris Lattner2006-03-141-0/+8
| | | | llvm-svn: 26762
* Reflect the fact that empty strings can be expressed as null.Jim Laskey2006-03-141-5/+4
| | | | llvm-svn: 26761
* 1. Use null for serialized empty strings.Jim Laskey2006-03-141-9/+17
| | | | | | | 2. Allow for user defined debug descriptors. 3. Allow for user augmented fields on debug descriptors. llvm-svn: 26760
* Bring debugging information up to date.Jim Laskey2006-03-141-636/+1166
| | | | llvm-svn: 26759
* Fix an off by one error that caused PPC LLC failures last night.Chris Lattner2006-03-141-1/+1
| | | | llvm-svn: 26758
* Implement a FIXME, recusively reassociatingChris Lattner2006-03-141-26/+65
| | | | | | | | A*A*B + A*A*C --> A*(A*B+A*C) --> A*(A*(B+C)) This implements Reassociate/mul-factor3.ll llvm-svn: 26757
* new testcase from a FIXME in the codeChris Lattner2006-03-142-2/+16
| | | | llvm-svn: 26756
* extract some code into a method, no functionality changeChris Lattner2006-03-141-50/+56
| | | | llvm-svn: 26755
* transformation implementedChris Lattner2006-03-141-10/+0
| | | | llvm-svn: 26754
* Promote shifts by a constant to multiplies so that we can reassociateChris Lattner2006-03-141-13/+17
| | | | | | | (x<<1)+(y<<1) -> (X+Y)<<1. This implements Transforms/Reassociate/shift-factor.ll llvm-svn: 26753
* Test that we can reassociate (x<<1)+(y<<1) -> (X+Y)<<1.Chris Lattner2006-03-141-0/+11
| | | | llvm-svn: 26752
* Allow a sequence of digits at the end of the directory name when trying toReid Spencer2006-03-142-2/+3
| | | | | | | | | | | | | | | determine the top directory. This allows the tool to find the correct top directory when you have something like: /x/llvm /x/llvm2 /x/llvm3 Previously the scripts would always find /x/llvm even if you were in the llvm2 or llvm3 directory because the pattern didn't allow the digits at the end of the path. llvm-svn: 26751
* Change llvm.cs.uiuc.edu -> llvm.orgReid Spencer2006-03-141-1/+1
| | | | llvm-svn: 26750
* Add a newline at the end to avoid gcc warnings.Reid Spencer2006-03-141-1/+1
| | | | llvm-svn: 26749
* Convert llvm.cs.uiuc.edu -> llvm.orgReid Spencer2006-03-148-9/+9
| | | | llvm-svn: 26748
* Change docs for llvm.cs.uiuc.edu -> llvm.orgReid Spencer2006-03-1423-27/+27
| | | | llvm-svn: 26747
* Changes docs for llvm.cs.uiuc.edu -> llvm.orgReid Spencer2006-03-1430-94/+94
| | | | llvm-svn: 26746
* 1. Handle removal of all arguments for a morphed intrinsic.Jim Laskey2006-03-141-1/+4
| | | | | | 2. Remove the declaration of llvm.dbg.declare. llvm-svn: 26745
* Remove the use of llvm.dbg.declare.Jim Laskey2006-03-141-6/+3
| | | | llvm-svn: 26744
* PPC LSR pass should use target lowering hooks.Evan Cheng2006-03-131-2/+2
| | | | llvm-svn: 26743
* Added getTargetLowering() to TargetMachine. Refactored targets to support this.Evan Cheng2006-03-1327-66/+97
| | | | llvm-svn: 26742
* UpdateEvan Cheng2006-03-131-0/+3
| | | | llvm-svn: 26741
* Add LSR hooks.Evan Cheng2006-03-133-5/+58
| | | | llvm-svn: 26740
* Added getTargetLowering() - returns DAG lowering info.Evan Cheng2006-03-131-0/+3
| | | | llvm-svn: 26739
* Added target lowering hooks which LSR consults to make more intelligentEvan Cheng2006-03-132-27/+41
| | | | | | transformation decisions. llvm-svn: 26738
* Handle builtins that directly correspond to GCC builtins.Chris Lattner2006-03-131-2/+25
| | | | llvm-svn: 26737
* emit a mapping from LLVM intrinsic -> GCC builtins.Chris Lattner2006-03-133-8/+30
| | | | llvm-svn: 26736
* Verify that packed type operands have the right size and base type.Chris Lattner2006-03-132-6/+28
| | | | llvm-svn: 26735
* Add a first ppc altivec intrinsic. Add packed type support.Chris Lattner2006-03-131-0/+30
| | | | llvm-svn: 26734
* For functions that use vector registers, save VRSAVE, mark usedChris Lattner2006-03-134-10/+130
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | registers, and update it on entry to each function, then restore it on exit. This compiles: void func(vfloat *a, vfloat *b, vfloat *c) { *a = *b * *c + *c; } to this: _func: mfspr r2, 256 oris r6, r2, 49152 mtspr 256, r6 lvx v0, 0, r5 lvx v1, 0, r4 vmaddfp v0, v1, v0, v0 stvx v0, 0, r3 mtspr 256, r2 blr GCC produces this (which has additional stack accesses): _func: mfspr r0,256 stw r0,-4(r1) oris r0,r0,0xc000 mtspr 256,r0 lvx v0,0,r5 lvx v1,0,r4 lwz r12,-4(r1) vmaddfp v0,v0,v1,v0 stvx v0,0,r3 mtspr 256,r12 blr llvm-svn: 26733
* Optimizing to calloc() to malloc() + memset() can be bad on some platforms.Evan Cheng2006-03-131-33/+0
| | | | llvm-svn: 26732
* make sure dead token factor nodes are removed by the dag combiner.Chris Lattner2006-03-131-0/+1
| | | | llvm-svn: 26731
* Bring makefile back into compliance with standard by usingReid Spencer2006-03-131-1/+1
| | | | | | $(Echo) instead of @echo llvm-svn: 26730
* Handle the removal of the debug chain.Jim Laskey2006-03-138-65/+115
| | | | llvm-svn: 26729
OpenPOWER on IntegriCloud