| Commit message (Expand) | Author | Age | Files | Lines |
* | Turn store -> null/undef into the LLVM unreachable instruction! This simple | Chris Lattner | 2004-10-18 | 1 | -0/+27 |
* | Turn things with obviously undefined semantics into 'store -> null' | Chris Lattner | 2004-10-18 | 1 | -26/+34 |
* | My friend the invoke instruction does not dominate all basic blocks if it | Chris Lattner | 2004-10-18 | 1 | -1/+2 |
* | Getting ADCE to interact well with unreachable instructions seems like a nont... | Chris Lattner | 2004-10-17 | 1 | -2/+3 |
* | Remove printout, realize that instructions in the entry block dominate all | Chris Lattner | 2004-10-17 | 1 | -6/+6 |
* | hasConstantValue will soon return instructions that don't dominate the PHI node, | Chris Lattner | 2004-10-17 | 1 | -4/+18 |
* | Fix a type violation | Chris Lattner | 2004-10-16 | 1 | -1/+1 |
* | Kill the bogon that slipped into my buffer before I committed. | Chris Lattner | 2004-10-16 | 1 | -1/+1 |
* | Implement InstCombine/getelementptr.ll:test9, which is the source of many | Chris Lattner | 2004-10-16 | 1 | -0/+18 |
* | Optimize instructions involving undef values. For example X+undef == undef. | Chris Lattner | 2004-10-16 | 1 | -27/+125 |
* | Handle undef values as undefined on the constant lattice | Chris Lattner | 2004-10-16 | 1 | -3/+6 |
* | Add note | Chris Lattner | 2004-10-16 | 1 | -0/+1 |
* | Update to reflect changes in Makefile rules. | Reid Spencer | 2004-10-13 | 1 | -4/+2 |
* | Transform memmove -> memcpy when the source is obviously constant memory. | Chris Lattner | 2004-10-12 | 1 | -16/+33 |
* | Fix a REALLY obscure bug in my previous checkin, which was splicing the END | Chris Lattner | 2004-10-12 | 1 | -1/+1 |
* | Handle a common case more carefully. In particular, instead of transforming | Chris Lattner | 2004-10-11 | 1 | -4/+33 |
* | Reenable the transform, turning X/-10 < 1 into X > -10 | Chris Lattner | 2004-10-11 | 1 | -5/+5 |
* | Initial version of automake Makefile.am file. | Reid Spencer | 2004-10-10 | 1 | -0/+50 |
* | Use DEBUG instead of DebugFlag directly, as DebugFlag does not respect | Chris Lattner | 2004-10-09 | 1 | -5/+3 |
* | Implement sub.ll:test17, -X/C -> X/-C | Chris Lattner | 2004-10-09 | 1 | -0/+5 |
* | Temporarily disable a buggy transformation until it can be fixed. This fixes | Chris Lattner | 2004-10-08 | 1 | -0/+5 |
* | Instcombine (X & FF00) + xx00 -> (X+xx00) & FF00, implementing and.ll:test27 | Chris Lattner | 2004-10-08 | 1 | -0/+25 |
* | Little patch to turn (shl (add X, 123), 4) -> (add (shl X, 4), 123 << 4) | Chris Lattner | 2004-10-08 | 1 | -0/+3 |
* | Instcombine: -(X sdiv C) -> (X sdiv -C), tested by sub.ll:test16 | Chris Lattner | 2004-10-06 | 1 | -0/+8 |
* | Reduce code growth implied by the tail duplication pass by not duplicating | Chris Lattner | 2004-10-06 | 1 | -0/+75 |
* | Hrm, debugging printouts do not need to be in here | Chris Lattner | 2004-09-29 | 1 | -2/+0 |
* | * Pull range optimization code out into new InsertRangeTest function. | Chris Lattner | 2004-09-29 | 1 | -80/+219 |
* | Fold binary expressions and casts into PHI nodes that have all constant inputs. | Chris Lattner | 2004-09-29 | 1 | -1/+86 |
* | Hrm, really, all tests passed without this, but it is scary to think how... | Chris Lattner | 2004-09-29 | 1 | -2/+2 |
* | Remove debugging printout | Chris Lattner | 2004-09-29 | 1 | -1/+38 |
* | Fold (X setcc C1) | (X setcc C2) | Chris Lattner | 2004-09-28 | 1 | -1/+102 |
* | Fold (and (setcc X, C1), (setcc X, C2)) | Chris Lattner | 2004-09-28 | 1 | -20/+121 |
* | Implement X / C1 / C2 folding | Chris Lattner | 2004-09-28 | 1 | -8/+54 |
* | shl is always zero extending, so always use a zero extending shift right. | Chris Lattner | 2004-09-28 | 1 | -5/+12 |
* | Fix two bugs: one where a condition was mistakenly swapped, and another | Chris Lattner | 2004-09-27 | 1 | -14/+11 |
* | Fold: (setcc (shr X, ShAmt), CI), where 'cc' is eq or ne. This xform | Chris Lattner | 2004-09-27 | 1 | -0/+45 |
* | Implement shift-and combinations, implementing InstCombine/and.ll:test19-21 | Chris Lattner | 2004-09-24 | 1 | -7/+44 |
* | Move LHSI->hasOneUse() into the arms of the conditional, reindenting code. | Chris Lattner | 2004-09-23 | 1 | -71/+72 |
* | Implement Transforms/InstCombine/and.ll:test18, a case that occurs 20 times | Chris Lattner | 2004-09-23 | 1 | -0/+45 |
* | Implement select.ll:test16: fold load (select C, X, null) -> load X | Chris Lattner | 2004-09-23 | 1 | -0/+14 |
* | Do not fold (X + C1 != C2) if there are other users of the add. Doing | Chris Lattner | 2004-09-21 | 1 | -2/+3 |
* | Fix potential miscompilations: InstCombine/2004-09-20-BadLoadCombine*.llx | Chris Lattner | 2004-09-20 | 1 | -7/+19 |
* | Fix loop condition so that we don't decrement off the beginning of the | Alkis Evlogimenos | 2004-09-20 | 1 | -5/+5 |
* | Prototype these functions more accurately | Chris Lattner | 2004-09-20 | 10 | -11/+13 |
* | Make isSafeToLoadUnconditionally a bit smarter, implementing PR362 and | Chris Lattner | 2004-09-19 | 1 | -6/+28 |
* | Make instruction combining a bit more aggressive in the face of volatile | Chris Lattner | 2004-09-19 | 1 | -3/+60 |
* | Convert code to compile with vc7.1. | Reid Spencer | 2004-09-15 | 6 | -13/+18 |
* | Fix a bug in the previous checkin that broke 255.vortex | Chris Lattner | 2004-09-15 | 1 | -1/+1 |
* | Make sure to update alias analysis information as we transform the function. | Chris Lattner | 2004-09-15 | 1 | -21/+57 |
* | Remove a long-dead pass. Actually, this pass was never used at all. | Chris Lattner | 2004-09-14 | 1 | -182/+0 |