- 0d5fcae Defer some shl transforms to DAGCombine. by Jakob Stoklund Olesen · 12 years ago
- 0d77b9c Extract the broken part of XFAILed test into its own file. by Jakob Stoklund Olesen · 12 years ago
- f5782e2 FileCheckize by Jakob Stoklund Olesen · 12 years ago
- 377bf1a Nobody likes shifty instructions, but that was a bit strong. by Jakob Stoklund Olesen · 12 years ago
- 35ee7d2 Added support for disassembling unpredictable swp/swpb ARM instructions. by Silviu Baranga · 12 years ago
- 6b9f97d Fix the bahavior of the disassembler when decoding unpredictable mrs instructions on ARM. Now the diasassembler emmits warnings instead of errors. by Silviu Baranga · 12 years ago
- fa1ebc6 Added support for unpredictable mcrr/mcrr2/mrrc/mrrc2 ARM instruction in the disassembler. Since the upredicability conditions are complex, C++ code was added to handle them. by Silviu Baranga · 12 years ago
- e546c4c Fixed decoding for the ARM cdp2 instruction. The restriction on the coprocessor number was removed for this instruction. by Silviu Baranga · 12 years ago
- 9e71231 Add suport for unpredicatble cases of the cmp, tst, teq and cmnz ARM instructions in the disassembler. by Silviu Baranga · 12 years ago
- 41c3e9a FileCheckify, un-XFAIL SimplifyLibCalls/floor test Fixes build on MSVC by Joe Groff · 12 years ago
- d15c581 Move win32 SimplifyLibcall test under Transforms by Joe Groff · 12 years ago
- d5bda5e fix pr12559: mark unavailable win32 math libcalls by Joe Groff · 12 years ago
- ecdc9d5 Add disassembler to MIPS. by Akira Hatanaka · 12 years ago
- 93751c8 Force cmov on test so block placement doesn't shuffle the code around. by Benjamin Kramer · 12 years ago
- 72aadc0 Fix bad EXTRACT_SUBREG in instruction selection for extending-loads on NEON. by James Molloy · 12 years ago
- 86df062 Revert "SCEV: When expanding a GEP the final addition to the base pointer has NUW but not NSW." by Benjamin Kramer · 12 years ago
- 8ca441a Test cases that assume layout should use -disable-code-place. by Andrew Trick · 12 years ago
- c5a2a33 Fix ARM disassembly of VLD2 (single 2-element structure to all lanes) by Kevin Enderby · 12 years ago
- 8975f51 by Preston Gurd · 12 years ago
- fd2e4e6 Disable the atom scheduling test after r154874 broke it. by Chandler Carruth · 12 years ago
- bf42f24 ARM two-operand forms for vhadd and vhsub instructions. by Jim Grosbach · 12 years ago
- 177bea5 Relax this test a touch to cope with different assembly variants. by Chandler Carruth · 12 years ago
- f1a60c7 Fix updateTerminator to be resiliant to degenerate terminators where by Chandler Carruth · 12 years ago
- 68f89a6 MC assembly parser handling for trailing comma in macro instantiation. by Jim Grosbach · 12 years ago
- 39ac325 FileCheckize these tests. by Jakob Stoklund Olesen · 12 years ago
- fbefc91 Disable code placement for this test. by Jakob Stoklund Olesen · 12 years ago
- 2867c85 Remove support for the special 'fast' value for fpmath accuracy for the moment. by Duncan Sands · 12 years ago
- 2c651fe Fix incorrect atomics codegen introduced in r154705, and extend test to catch it. by Richard Smith · 12 years ago
- 1fbfea7 This patch fixes 3 problems: by Akira Hatanaka · 12 years ago
- 199366a ARM assembly two-operand forms for VRSHL. by Jim Grosbach · 12 years ago
- 695eca6 Tidy up. Test formatting. by Jim Grosbach · 12 years ago
- 3ef7edc Do not add offset in applyFixup. This has already been accounted for in Value. by Akira Hatanaka · 12 years ago
- 705e257 ARM two-operand aliases for VRHADD instructions. by Jim Grosbach · 12 years ago
- dbd6ba3 Tidy up. Testcase formatting. by Jim Grosbach · 12 years ago
- 57ca13e Move to X86 directory because this fails on non-X86 platforms. by Bill Wendling · 12 years ago
- 8883c43 Make it possible to indicate relaxed floating point requirements at the IR level by Duncan Sands · 12 years ago
- 9e67db4 Flip the new block-placement pass to be on by default. by Chandler Carruth · 12 years ago
- 0de089a Remove an overly brittle test. This test will no longer be interesting by Chandler Carruth · 12 years ago
- e773e8c Add a somewhat hacky heuristic to do something different from whole-loop by Chandler Carruth · 12 years ago
- d0c478d Add -disassemble support for -show-inst and -show-encode capability llvm-mc. Also refactor so all MC paraphernalia are created once for all uses as much as possible. by Richard Barton · 12 years ago
- 16295fc Tweak the loop rotation logic to check whether the loop is naturally by Chandler Carruth · 12 years ago
- 31490ba Remove dead SD nodes after the combining pass. Fixes PR12201. by Hal Finkel · 12 years ago
- 70daea9 Rewrite how machine block placement handles loop rotation. by Chandler Carruth · 12 years ago
- 2cb1e9d Remove AVX2 vpermq and vpermpd intrinsics. These can now be handled with normal shuffle vectors. by Craig Topper · 12 years ago
- f16af0a Fix PR12529. The Vxx family of instructions are only supported by AVX. by Nadav Rotem · 12 years ago
- 3ab32ea When emulating vselect using OR/AND/XOR make sure to bitcast the result back to the original type. by Nadav Rotem · 12 years ago
- 73c504a Added VPERM optimization for AVX2 shuffles by Elena Demikhovsky · 12 years ago
- 5e5c5f8 Rename "fpaccuracy" metadata to the more generic "fpmath". That's because I'm by Duncan Sands · 12 years ago
- bba23ed Fix an error in BBVectorize important for vectorizing pointer types. by Hal Finkel · 12 years ago
- f3f5a1e Enhance BBVectorize to more-properly handle pointer values and vectorize GEPs. by Hal Finkel · 12 years ago
- 42fc29e Fix X86 codegen for 'atomicrmw nand' to generate *x = ~(*x & y), not *x = ~*x & y. by Richard Smith · 12 years ago
- fc3665c Add support to BBVectorize for vectorizing selects. by Hal Finkel · 12 years ago
- 7ece953 On Darwin targets, only use vfma etc. if the source use fma() intrinsic explicitly. by Evan Cheng · 12 years ago
- 4423477 Consider ObjC runtime calls objc_storeWeak and others which make a copy of by Dan Gohman · 12 years ago
- e92077f Catch the Python exception when subprocess.Popen is failing. by Sylvestre Ledru · 12 years ago
- 6c189ec Use the new Use-aware dominates method to apply the objc runtime by Dan Gohman · 12 years ago
- 511568d Don't move objc_autorelease calls past autorelease pool boundaries when by Dan Gohman · 12 years ago
- 2f69e4c Disable Hexagon test temporarily. by Sirish Pande · 12 years ago
- c68dda8 This patch improves the MCJIT runtime dynamic loader by adding new handling by Preston Gurd · 12 years ago
- bf596c9 Fix 128-bit ptest intrinsics to take v2i64 instead of v4f32 since these are integer instructions. by Craig Topper · 12 years ago
- ed08489 Revert changes that were accidentally committed. by Akira Hatanaka · 12 years ago
- 55e0e43 Fix string that is being checked. by Akira Hatanaka · 12 years ago
- 1cc6333 Emit neg.s or neg.d only if -enable-no-nans-fp-math is supplied by user, by Akira Hatanaka · 12 years ago
- c12a6e6 Emit abs.s or abs.d only if -enable-no-nans-fp-math is supplied by user. by Akira Hatanaka · 12 years ago
- b318cc1 Fixed a case of ARM disassembly getting an assert on a bad encoding by Kevin Enderby · 12 years ago
- 056c51e Fix bugs in lowering of FCOPYSIGN nodes. by Akira Hatanaka · 12 years ago
- 1835547 ARM 'vuzp.32 Dd, Dm' is a pseudo-instruction. by Jim Grosbach · 12 years ago
- 6073b30 ARM 'vzip.32 Dd, Dm' is a pseudo-instruction. by Jim Grosbach · 12 years ago
- 14b4c03 Add more fused mul+add/sub patterns. rdar://10139676 by Evan Cheng · 12 years ago
- e611378 Reapply 154396 after fixing a test. by Nadav Rotem · 12 years ago
- bee78fe Clean up ARM fused multiply + add/sub support some more: rename some isel by Evan Cheng · 12 years ago
- 92c9045 Match (fneg (fma) to vfnma. rdar://10139676 by Evan Cheng · 12 years ago
- 0d82fe7 Add retw and lretw instructions. Also, fix Intel syntax parsing for all by Charles Davis · 12 years ago
- a0908d0 Merge fma.ll into fusedMAC.ll by Evan Cheng · 12 years ago
- a69da35 Fix ARM disassembly of VLD instructions with writebacks. And add test a case by Kevin Enderby · 12 years ago
- a5378eb ARM add missing Thumb1 two-operand aliases for shift-by-immediate. by Jim Grosbach · 12 years ago
- 82509e5 Fix a number of problems with ARM fused multiply add/subtract instructions. by Evan Cheng · 12 years ago
- 89cdaf46 Fix test to be register assignment invariant. by Jakob Stoklund Olesen · 12 years ago
- 06886aa Move the constant-folding support for FP_ROUND in SelectionDAG from the one-operand version of getNode() to the two-operand version, since it became a two-operand node at sound point. by Owen Anderson · 12 years ago
- cff60c1 [tsan] two more compile-time optimizations: by Kostya Serebryany · 12 years ago
- 3aef2ff Handle llvm.fma.* intrinsics. rdar://10914096 by Evan Cheng · 12 years ago
- 507bb7a Add a comment noting that the fdiv -> fmul conversion won't generate by Duncan Sands · 12 years ago
- a139051 Temporarily revert this patch to see if it brings the buildbots back. by Eric Christopher · 12 years ago
- 2076af0 [tsan] compile-time instrumentation: do not instrument a read if by Kostya Serebryany · 12 years ago
- 18112d8 To ensure that we have more accurate line information for a block by Eric Christopher · 12 years ago
- a23ecc2 ARM fix cc_out operand handling for t2SUBrr instructions. by Jim Grosbach · 12 years ago
- 50e64cf Modify the code that lowers shuffles to blends from using blendvXX to vblendXX. by Nadav Rotem · 12 years ago
- 999821c Transform div to mul with reciprocal only when fp imm is legal. by Anton Korobeynikov · 12 years ago
- 1fd63df Express the number of ULPs in fpaccuracy metadata as a real rather than a by Duncan Sands · 12 years ago
- d9fc1ce Fix 12513: Loop unrolling breaks with indirect branches. by Andrew Trick · 12 years ago
- fa12d0d Add proper checks. by Evan Cheng · 12 years ago
- bf010eb Fix a long standing tail call optimization bug. When a libcall is emitted by Evan Cheng · 12 years ago
- fdb230a Don't try to zExt just to check if an integer constant is zero, it might by Rafael Espindola · 12 years ago
- 23f369d Test case for PR12495. by Lang Hames · 12 years ago
- 787c3fd Have TargetLowering::getPICJumpTableRelocBase return a node that points to the by Akira Hatanaka · 12 years ago
- 7f35455 When performing a truncating store, it's possible to rearrange the data by Chad Rosier · 12 years ago
- decbc43 Pattern match a setcc of boolean value with 0 as a truncate. by Rafael Espindola · 12 years ago
- e80aa7c Lower some x86 shuffle sequences to the vblend family of instructions. by Nadav Rotem · 12 years ago
- 154819d Fix a bug in the lowering of broadcasts: ConstantPools need to use the target pointer type. by Nadav Rotem · 12 years ago
- ab5a55e Cleanup and relax a restriction on the matching of global offsets into by Chandler Carruth · 12 years ago