| Commit message (Expand) | Author | Age | Files | Lines |
|---|
| * | Make sure type is not extended or untyped before create a constant of the typ... | Evan Cheng | 2012-06-26 | 1 | -0/+4 |
| * | Rename -allow-excess-fp-precision flag to -fuse-fp-ops, and switch from a | Lang Hames | 2012-06-22 | 1 | -2/+2 |
| * | Fix potential crash if DAGCombine on stores sees a half type | Pete Cooper | 2012-06-21 | 1 | -1/+2 |
| * | Add users of a MERGE_VALUE node to the worklist to process again when the nod... | Pete Cooper | 2012-06-20 | 1 | -0/+3 |
| * | Fix DAGCombine to deal with ext-conversion of pre/post_inc loads. | Hal Finkel | 2012-06-20 | 1 | -1/+8 |
| * | Add DAG-combines for aggressive FMA formation. | Lang Hames | 2012-06-19 | 1 | -0/+43 |
| * | Make comment slightly more helpful. | Lang Hames | 2012-06-14 | 1 | -1/+1 |
| * | Switch the canonical FMA term operand order to match both the comment I wrote... | Owen Anderson | 2012-05-30 | 1 | -1/+1 |
| * | Teach DAGCombine to canonicalize the position of a constant in the term opera... | Owen Anderson | 2012-05-30 | 1 | -0/+4 |
| * | DAGCombiner should not change the type of an extract_vector index. | Jim Grosbach | 2012-05-08 | 1 | -3/+4 |
| * | Teach DAG combine to fold x-x to 0.0 when unsafe FP math is enabled. | Owen Anderson | 2012-05-07 | 1 | -0/+4 |
| * | Teach DAGCombine the same multiply-by-1.0 folding trick when doing FMAs, just... | Owen Anderson | 2012-05-02 | 1 | -0/+18 |
| * | Teach DAG combine that multiplication by 1.0 can always be constant folded. | Owen Anderson | 2012-05-02 | 1 | -0/+3 |
| * | ZERO_EXTEND/SIGN_EXTEND/TRUNCATE optimization for AVX2 | Elena Demikhovsky | 2012-04-22 | 1 | -0/+2 |
| * | Register DAGUpdateListeners with SelectionDAG. | Jakob Stoklund Olesen | 2012-04-20 | 1 | -48/+30 |
| * | Remove dead SD nodes after the combining pass. Fixes PR12201. | Hal Finkel | 2012-04-16 | 1 | -0/+1 |
| * | Reapply 154397. Original message: | Nadav Rotem | 2012-04-11 | 1 | -11/+18 |
| * | Add a comment noting that the fdiv -> fmul conversion won't generate | Duncan Sands | 2012-04-10 | 1 | -3/+3 |
| * | Revert r154397, which was causing make check failures on the buildbots. | Owen Anderson | 2012-04-10 | 1 | -13/+6 |
| * | Fix a dagcombine optimization which assumes that the vsetcc result type is al... | Nadav Rotem | 2012-04-10 | 1 | -6/+13 |
| * | Transform div to mul with reciprocal only when fp imm is legal. | Anton Korobeynikov | 2012-04-10 | 1 | -2/+9 |
| * | Don't try to zExt just to check if an integer constant is zero, it might | Rafael Espindola | 2012-04-10 | 1 | -2/+2 |
| * | Pattern match a setcc of boolean value with 0 as a truncate. | Rafael Espindola | 2012-04-09 | 1 | -9/+48 |
| * | Remove unnecessary type check when combining and/or/xor of swizzles. Move som... | Craig Topper | 2012-04-09 | 1 | -13/+12 |
| * | Remove unnecessary 'else' on an 'if' that always returns | Craig Topper | 2012-04-09 | 1 | -1/+2 |
| * | Optimize code slightly. No functionality change. | Craig Topper | 2012-04-09 | 1 | -6/+7 |
| * | Replace some explicit checks with asserts for conditions that should never ha... | Craig Topper | 2012-04-09 | 1 | -14/+7 |
| * | Silence sign-compare warning. | Benjamin Kramer | 2012-04-08 | 1 | -1/+1 |
| * | Only have codegen turn fdiv by a constant into fmul by the reciprocal | Duncan Sands | 2012-04-08 | 1 | -5/+3 |
| * | 1. Remove the part of r153848 which optimizes shuffle-of-shuffle into a new | Nadav Rotem | 2012-04-07 | 1 | -12/+14 |
| * | Convert floating point division by a constant into multiplication by the | Duncan Sands | 2012-04-07 | 1 | -0/+13 |
| * | Always compute all the bits in ComputeMaskedBits. | Rafael Espindola | 2012-04-04 | 1 | -15/+10 |
| * | Add predicates for checking whether targets have free FNEG and FABS operation... | Owen Anderson | 2012-04-02 | 1 | -3/+5 |
| * | Optimizing swizzles of complex shuffles may generate additional complex shuff... | Nadav Rotem | 2012-04-02 | 1 | -1/+9 |
| * | This commit contains a few changes that had to go in together. | Nadav Rotem | 2012-04-01 | 1 | -0/+92 |
| * | fix what looks like a real logic bug, found by PVS-Studio (part of PR12357) | Chris Lattner | 2012-03-27 | 1 | -2/+2 |
| * | When combining (vextract shuffle (load ), <1,u,u,u>), 0) -> (load ), add user... | Craig Topper | 2012-03-20 | 1 | -0/+1 |
| * | Fix DAG combine which creates illegal vector shuffles. Patch by Heikki Kultala. | Duncan Sands | 2012-03-19 | 1 | -0/+6 |
| * | When optimizing certain BUILD_VECTOR nodes into other BUILD_VECTOR nodes, add... | Nadav Rotem | 2012-03-15 | 1 | -0/+4 |
| * | Add a xform to the DAG combiner. | Bill Wendling | 2012-03-15 | 1 | -0/+17 |
| * | Fortify r152675 a bit. Although I'm not able to come up with a test case that... | Evan Cheng | 2012-03-13 | 1 | -3/+11 |
| * | DAG combine incorrectly optimize (i32 vextract (v4i16 load $addr), c) to | Evan Cheng | 2012-03-13 | 1 | -4/+19 |
| * | Give dagcombiner's worklist some inline capacity. | Benjamin Kramer | 2012-03-10 | 1 | -3/+2 |
| * | Extend r148086 to check for [r +/- reg] address mode. This fixes queens perfo... | Evan Cheng | 2012-03-06 | 1 | -4/+7 |
| * | Make it possible for a target to mark FSUB as Expand. This requires providin... | Owen Anderson | 2012-03-06 | 1 | -16/+29 |
| * | Teach the DAGCombiner that certain loadext nodes followed by ANDs can be conv... | James Molloy | 2012-02-20 | 1 | -0/+82 |
| * | Remove extraneous #include and spelling mistake introduced in r150669. | James Molloy | 2012-02-16 | 1 | -2/+1 |
| * | Modify the algorithm when traversing the DAGCombiner's worklist to be O(log N... | James Molloy | 2012-02-16 | 1 | -13/+36 |
| * | Fix a bug in DAGCombine for the optimization of BUILD_VECTOR. We cant generat... | Nadav Rotem | 2012-02-13 | 1 | -2/+6 |
| * | This patch addresses the problem of poor code generation for the zext | Nadav Rotem | 2012-02-12 | 1 | -14/+29 |
