| Commit message (Expand) | Author | Age | Files | Lines |
|---|
| * | Unbreak VC++ build. | Jeff Cohen | 2006-12-02 | 1 | -7/+7 |
| * | disable transformations that are invalid for fp vectors. This fixes | Chris Lattner | 2006-12-02 | 1 | -4/+4 |
| * | Remove 4 FIXMEs to hack around cast-to-bool problems which no longer exist. | Reid Spencer | 2006-11-30 | 1 | -46/+3 |
| * | make it clear that this is always a zext | Chris Lattner | 2006-11-30 | 1 | -1/+1 |
| * | One more bugfix, 3 cases of making casts explicit. | Chris Lattner | 2006-11-30 | 1 | -5/+8 |
| * | Fix a bug in globalopt due to the recent cast patch. | Chris Lattner | 2006-11-30 | 1 | -1/+2 |
| * | implement cast.ll:test35. With this, we recognize: | Chris Lattner | 2006-11-29 | 1 | -0/+16 |
| * | Teach instcombine to turn trunc(srl x, c) -> srl (trunc(x), c) when safe. | Chris Lattner | 2006-11-29 | 1 | -1/+33 |
| * | Implement Regression/Transforms/InstCombine/bswap-fold.ll, | Chris Lattner | 2006-11-29 | 1 | -1/+24 |
| * | Join a split line. | Reid Spencer | 2006-11-29 | 1 | -2/+1 |
| * | Undo the last patch until 253.perlbmk passes with these changes. | Reid Spencer | 2006-11-28 | 1 | -3/+46 |
| * | Remove 4 FIXME's from the CAST patch now that the back end is correctly | Reid Spencer | 2006-11-28 | 1 | -46/+3 |
| * | Fix PR1014 and InstCombine/2006-11-27-XorBug.ll. | Chris Lattner | 2006-11-27 | 1 | -10/+8 |
| * | For PR950: | Reid Spencer | 2006-11-27 | 21 | -786/+908 |
| * | Remove #include <iostream> and use llvm_* streams instead. | Bill Wendling | 2006-11-26 | 3 | -40/+37 |
| * | Replace #include <iostream> with llvm_* streams. | Bill Wendling | 2006-11-26 | 6 | -69/+62 |
| * | Removed #include <iostream> and replaced with llvm_* streams. | Bill Wendling | 2006-11-26 | 11 | -115/+100 |
| * | Removed #include <iostream> and used the llvm_cerr/DOUT streams instead. | Bill Wendling | 2006-11-26 | 7 | -44/+34 |
| * | Update to new predicate simplifier VRP design. Fixes PR966 and PR967. | Nick Lewycky | 2006-11-22 | 1 | -574/+1105 |
| * | This xform is handled by FoldOpIntoPhi in visitCastInst in a more elegant way. | Chris Lattner | 2006-11-21 | 1 | -30/+1 |
| * | Do not convert massive blocks on phi nodes into select statements. Instead | Chris Lattner | 2006-11-18 | 1 | -0/+27 |
| * | If an indvar with a variable stride is used by the exit condition, go ahead | Chris Lattner | 2006-11-17 | 1 | -4/+0 |
| * | Fix a gcc 4.2 warning. | Chris Lattner | 2006-11-15 | 1 | -0/+2 |
| * | implement InstCombine/shift-simplify.ll by transforming: | Chris Lattner | 2006-11-14 | 1 | -3/+46 |
| * | implement InstCombine/and-compare.ll:test1. This compiles: | Chris Lattner | 2006-11-14 | 1 | -0/+26 |
| * | Fix InstCombine/2006-11-10-ashr-miscompile.ll a miscompilation introduced | Chris Lattner | 2006-11-10 | 1 | -3/+3 |
| * | second patch to fix PR992/993. | Chris Lattner | 2006-11-09 | 1 | -4/+17 |
| * | Minimal patch to fix PR992/PR993 | Chris Lattner | 2006-11-09 | 1 | -2/+1 |
| * | Teach ShrinkDemandedConstant how to handle X+C. This implements: | Chris Lattner | 2006-11-09 | 1 | -1/+100 |
| * | reenable factoring of GEP expressions, being more precise about the | Chris Lattner | 2006-11-08 | 1 | -5/+10 |
| * | make this code more efficient by not creating a phi node we are just going to | Chris Lattner | 2006-11-08 | 1 | -36/+33 |
| * | Remove redundant <cmath>. | Jim Laskey | 2006-11-08 | 1 | -1/+0 |
| * | disable this factoring optzn for GEPs for now, this severely pessimizes some | Chris Lattner | 2006-11-08 | 1 | -1/+1 |
| * | For PR950: | Reid Spencer | 2006-11-08 | 5 | -215/+194 |
| * | scalarrepl should not split the two elements of the vsiidx array: | Chris Lattner | 2006-11-07 | 1 | -3/+7 |
| * | Unbreak VC++ build. | Jeff Cohen | 2006-11-05 | 2 | -6/+6 |
| * | Remove commented line from earlier debugging. | Nick Lewycky | 2006-11-05 | 1 | -2/+0 |
| * | The wrong parameter was being tested to deturmine i32 vs i64 | Andrew Lenharth | 2006-11-03 | 1 | -1/+1 |
| * | remove dead code | Chris Lattner | 2006-11-03 | 1 | -13/+0 |
| * | For PR786: | Reid Spencer | 2006-11-02 | 22 | -47/+22 |
| * | For PR950: | Reid Spencer | 2006-11-02 | 3 | -119/+134 |
| * | There can be more than one PHINode at the start of the block. | Devang Patel | 2006-11-01 | 1 | -5/+4 |
| * | Handle PHINode with only one incoming value. | Devang Patel | 2006-11-01 | 1 | -5/+9 |
| * | Fix GlobalOpt/2006-11-01-ShrinkGlobalPhiCrash.ll and McGill/chomp | Chris Lattner | 2006-11-01 | 1 | -8/+14 |
| * | Factor gep instructions through phi nodes. | Chris Lattner | 2006-11-01 | 1 | -10/+39 |
| * | Turn a phi of many loads into a phi of the address and a single load of the | Chris Lattner | 2006-11-01 | 1 | -41/+30 |
| * | Fix a bug in the previous patch | Chris Lattner | 2006-11-01 | 1 | -3/+6 |
| * | Fold things like "phi [add (a,b), add(c,d)]" into two phi's and one add. | Chris Lattner | 2006-11-01 | 1 | -3/+57 |
| * | generalize the fix for PR977 to also fix | Chris Lattner | 2006-10-31 | 1 | -28/+26 |
| * | Fix PR977 and Transforms/LCSSA/2006-10-31-UnreachableBlock.ll | Chris Lattner | 2006-10-31 | 1 | -1/+8 |
