| 2010-09-21 |
Chris Lattner | convert targets to the new MF.getMachineMemOperand... |
blob | commitdiff | raw |
| 2010-09-13 |
John Thompson | Added skeleton for inline asm multiple alternative... |
blob | commitdiff | raw | diff to current |
| 2010-09-08 |
Bruno Cardoso Lopes | Minor change. Fix comments and remove unused and redund... |
blob | commitdiff | raw | diff to current |
| 2010-09-08 |
Bruno Cardoso Lopes | x86 vector shuffle lowering now relies only on target... |
blob | commitdiff | raw | diff to current |
| 2010-09-07 |
Bruno Cardoso Lopes | Factor out some x86 vector shuffle rewriting and add... |
blob | commitdiff | raw | diff to current |
| 2010-09-07 |
Bruno Cardoso Lopes | Move code around to prepare for moving some of the... |
blob | commitdiff | raw | diff to current |
| 2010-09-07 |
Bill Wendling | Add an MVT::x86mmx type. It will take the place of... |
blob | commitdiff | raw | diff to current |
| 2010-09-07 |
Bruno Cardoso Lopes | decouple MMX check from regular splat checks. Some... |
blob | commitdiff | raw | diff to current |
| 2010-09-07 |
Bruno Cardoso Lopes | Remove now useless check, because the code can be match... |
blob | commitdiff | raw | diff to current |
| 2010-09-07 |
Bruno Cardoso Lopes | Minor change. Since the checks are equivalent, use... |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | Remove the last bit of isShuffleMaskLegal checks and... |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | make explicit that we not handle several mmx shuffles |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | Emit target specific nodes to handle palignr. Do not... |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | Emit target specific nodes to handle splats starting... |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | Emit target specific nodes for isPSHUFHWMask and isPSHU... |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | Emit target specific nodes for isSHUFPMask |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | Previous isMOVLMask matching already emits targets... |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | One more check from the original isShuffleMaskLegal... |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | Remove a duplicated but useless check that i've inserte... |
blob | commitdiff | raw | diff to current |
| 2010-09-04 |
Bruno Cardoso Lopes | Refactor some code and remove the extra checks for... |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Remove check for unpckh mask |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Remove check for unpckl mask |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Inline isShuffleMaskLegal into LowerVECTOR_SHUFFLE... |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Reapply considered harmfull part of rr112934 and r112942. |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Reintroduce a simple function refactoring done in r1129... |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Reapply piecies of r112942 and r112934 which don't do |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Reapply Fix comment |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Daniel Dunbar | Revert r112934, "- Use specific nodes to match unpckl... |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Daniel Dunbar | Revert r112938 "Fix comment", which depends on r112934... |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Daniel Dunbar | Revert r112942, "Use punpckh and unpckh family of nodes... |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Use punpckh and unpckh family of nodes instead of using... |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | Fix comment |
blob | commitdiff | raw | diff to current |
| 2010-09-03 |
Bruno Cardoso Lopes | - Use specific nodes to match unpckl masks. |
blob | commitdiff | raw | diff to current |
| 2010-09-02 |
Anton Korobeynikov | Revert win64 changes. They seem to be incomplete |
blob | commitdiff | raw | diff to current |
| 2010-09-02 |
Anton Korobeynikov | Properly allocate win64 shadow reg area. |
blob | commitdiff | raw | diff to current |
| 2010-09-02 |
Bruno Cardoso Lopes | Replace unpckl_undef and unpckh_undef matching with... |
blob | commitdiff | raw | diff to current |
| 2010-09-02 |
Bruno Cardoso Lopes | Move condition out to prepare for more matching |
blob | commitdiff | raw | diff to current |
| 2010-09-02 |
Bruno Cardoso Lopes | Remove checking for isUNPCKL_v_undef_Mask, the specific... |
blob | commitdiff | raw | diff to current |
| 2010-09-02 |
Bruno Cardoso Lopes | become more strict about when it's safe to use X86ISD... |
blob | commitdiff | raw | diff to current |
| 2010-09-01 |
Bruno Cardoso Lopes | Revert r112689, avoid those kind of checks cause they... |
blob | commitdiff | raw | diff to current |
| 2010-09-01 |
Bruno Cardoso Lopes | Use movlps, movlpd, movss and movsd specific nodes... |
blob | commitdiff | raw | diff to current |
| 2010-09-01 |
Bruno Cardoso Lopes | minor change, simplify some logic |
blob | commitdiff | raw | diff to current |
| 2010-09-01 |
Bruno Cardoso Lopes | Move some functions around so they can be used for... |
blob | commitdiff | raw | diff to current |
| 2010-08-31 |
Bruno Cardoso Lopes | Use x86 specific MOVSLDUP node, add more patterns to... |
blob | commitdiff | raw | diff to current |
| 2010-08-31 |
Bruno Cardoso Lopes | Use x86 specific MOVSHDUP node and add more patterns... |
blob | commitdiff | raw | diff to current |
| 2010-08-31 |
Bruno Cardoso Lopes | Use MOVHLPS node instead of matching using movhlps... |
blob | commitdiff | raw | diff to current |
| 2010-08-31 |
Bruno Cardoso Lopes | Use MOVLHPS and MOVHLPS x86 nodes whenever possible... |
blob | commitdiff | raw | diff to current |
| 2010-08-31 |
Bruno Cardoso Lopes | Use X86ISD::MOVSS and MOVSD to represent the movl mask... |
blob | commitdiff | raw | diff to current |
| 2010-08-28 |
Chris Lattner | fix the buildvector->insertp[sd] logic to not always... |
blob | commitdiff | raw | diff to current |
| 2010-08-28 |
Chris Lattner | fix the BuildVector -> unpcklps logic to not do pointle... |
blob | commitdiff | raw | diff to current |
| 2010-08-28 |
Chris Lattner | improve comments in the unpcklps generating logic,... |
blob | commitdiff | raw | diff to current |
| 2010-08-28 |
Bruno Cardoso Lopes | Clean up the logic of vector shuffles -> vector shifts. |
blob | commitdiff | raw | diff to current |
| 2010-08-27 |
Anton Korobeynikov | Properly handle passing of FP stuff to varargs function... |
blob | commitdiff | raw | diff to current |
| 2010-08-26 |
Bruno Cardoso Lopes | zap the now unused MVT::getIntVectorWithNumElements |
blob | commitdiff | raw | diff to current |
| 2010-08-26 |
Chris Lattner | implement SplitVecOp_CONCAT_VECTORS, fixing the include... |
blob | commitdiff | raw | diff to current |
| 2010-08-26 |
Chris Lattner | fix sse1 only codegen in x86-64 mode, which is something we |
blob | commitdiff | raw | diff to current |
| 2010-08-25 |
Bruno Cardoso Lopes | Revert this for now, PUNPCKLDQ dont operate on v4f32 |
blob | commitdiff | raw | diff to current |
| 2010-08-25 |
Anton Korobeynikov | Fix nasty mingw32 bug, which e.g. prevented llvm-gcc... |
blob | commitdiff | raw | diff to current |
| 2010-08-25 |
Bruno Cardoso Lopes | PUNPCKLDQ should also be used for v4f32 |
blob | commitdiff | raw | diff to current |
| 2010-08-25 |
Bruno Cardoso Lopes | teach lowering to get target specific nodes for pshufd... |
blob | commitdiff | raw | diff to current |
| 2010-08-24 |
Dan Gohman | Fix X86's isLegalAddressingMode to recognize that stati... |
blob | commitdiff | raw | diff to current |
| 2010-08-24 |
Bruno Cardoso Lopes | Use pshufhw and pshuflw in more cases and fix getTarget... |
blob | commitdiff | raw | diff to current |
| 2010-08-23 |
Bruno Cardoso Lopes | Start using target speficic nodes for shuffles: pshufhw... |
blob | commitdiff | raw | diff to current |
| 2010-08-23 |
Anton Korobeynikov | Revert invalid r111792. Jump tables are not broken... |
blob | commitdiff | raw | diff to current |
| 2010-08-23 |
Michael J. Spencer | Workaround broken jump tables on x86-64 COFF. |
blob | commitdiff | raw | diff to current |
| 2010-08-21 |
Bruno Cardoso Lopes | Prepare LowerVECTOR_SHUFFLEv8i16 to use x86 target... |
blob | commitdiff | raw | diff to current |
| 2010-08-20 |
Bruno Cardoso Lopes | This is the first step towards refactoring the x86... |
blob | commitdiff | raw | diff to current |
| 2010-08-17 |
Anton Korobeynikov | More fixes for win64: |
blob | commitdiff | raw | diff to current |
| 2010-08-14 |
Eric Christopher | Rework how the non-sse2 memory barrier is lowered so... |
blob | commitdiff | raw | diff to current |
| 2010-08-14 |
Chris Lattner | improve indentation |
blob | commitdiff | raw | diff to current |
| 2010-08-13 |
Bruno Cardoso Lopes | Fix comment to reflect code, and remove an unused argument |
blob | commitdiff | raw | diff to current |
| 2010-08-12 |
Bruno Cardoso Lopes | Begin to support some vector operations for AVX 256... |
blob | commitdiff | raw | diff to current |
| 2010-08-11 |
Dan Gohman | Use ISD::ADD instead of ISD::SUB with a negated constan... |
blob | commitdiff | raw | diff to current |
| 2010-08-10 |
Bruno Cardoso Lopes | Add AVX matching patterns to Packed Bit Test intrinsics. |
blob | commitdiff | raw | diff to current |
| 2010-08-10 |
Bruno Cardoso Lopes | Support AVX 256-bit load and store intrinsics |
blob | commitdiff | raw | diff to current |
| 2010-08-05 |
Bruno Cardoso Lopes | Support very basic (doesn't include ABI support in... |
blob | commitdiff | raw | diff to current |
| 2010-08-04 |
Eric Christopher | Make x86-64 membarriers work without sse and clean... |
blob | commitdiff | raw | diff to current |
| 2010-07-30 |
Bruno Cardoso Lopes | Support all 128-bit AVX vector intrinsics. Most part... |
blob | commitdiff | raw | diff to current |
| 2010-07-29 |
Jakob Stoklund Olesen | Revert r109652, and remove the offending assert in... |
blob | commitdiff | raw | diff to current |
| 2010-07-28 |
Jakob Stoklund Olesen | Create a fixed stack object for varargs that is as... |
blob | commitdiff | raw | diff to current |
| 2010-07-28 |
Nate Begeman | Implement a vectorized algorithm for <16 x i8> << ... |
blob | commitdiff | raw | diff to current |
| 2010-07-27 |
Nate Begeman | ~40% faster vector shl <4 x i32> on SSE 4.1 Larger... |
blob | commitdiff | raw | diff to current |
| 2010-07-26 |
Evan Cheng | On x86, f32 / f64 nodes share the same registers as... |
blob | commitdiff | raw | diff to current |
| 2010-07-24 |
Evan Cheng | Add an ILP scheduler. This is a register pressure aware... |
blob | commitdiff | raw | diff to current |
| 2010-07-23 |
Dale Johannesen | The only supported calling convention for X86-64 uses |
blob | commitdiff | raw | diff to current |
| 2010-07-22 |
Eric Christopher | Custom lower the memory barrier instructions and add... |
blob | commitdiff | raw | diff to current |
| 2010-07-22 |
Eric Christopher | 80-columns. |
blob | commitdiff | raw | diff to current |
| 2010-07-21 |
Nate Begeman | Fix a couple issues with Win64 ABI |
blob | commitdiff | raw | diff to current |
| 2010-07-21 |
Eric Christopher | Pulling out previous patch, must've run the tests in |
blob | commitdiff | raw | diff to current |
| 2010-07-21 |
Eric Christopher | Lower MEMBARRIER on x86 and support processors without... |
blob | commitdiff | raw | diff to current |
| 2010-07-15 |
Evan Cheng | Split -enable-finite-only-fp-math to two options: |
blob | commitdiff | raw | diff to current |
| 2010-07-14 |
Jakob Stoklund Olesen | Use TargetOpcode::COPY instead of X86-native register... |
blob | commitdiff | raw | diff to current |
| 2010-07-14 |
Evan Cheng | Fix for PR7193 was overly conservative. The only case... |
blob | commitdiff | raw | diff to current |
| 2010-07-10 |
Dan Gohman | Reapply bottom-up fast-isel, with several fixes for... |
blob | commitdiff | raw | diff to current |
| 2010-07-10 |
Jakob Stoklund Olesen | An x86 function returns a floating point value in st... |
blob | commitdiff | raw | diff to current |
| 2010-07-09 |
Bob Wilson | --- Reverse-merging r107947 into '.': |
blob | commitdiff | raw | diff to current |
| 2010-07-09 |
Dan Gohman | Fix the memoperand offsets in code generated for va_start. |
blob | commitdiff | raw | diff to current |
| 2010-07-09 |
Dan Gohman | Re-apply bottom-up fast-isel, with fixes. Be very caref... |
blob | commitdiff | raw | diff to current |
| 2010-07-08 |
Chris Lattner | Change LEA to have 5 operands for its memory operand... |
blob | commitdiff | raw | diff to current |
| 2010-07-08 |
Chris Lattner | add some long-overdue enums to refer to the parts of... |
blob | commitdiff | raw | diff to current |
| next |
projects / oota-llvm.git / history