| Commit message (Expand) | Author | Age | Files | Lines |
|---|
| * | Don't replace all dominated uses if there is only one use, since that | Duncan Sands | 2011-10-15 | 1 | -4/+9 |
| * | Fix indvars randomness by removing iteration over a map. | Andrew Trick | 2011-10-15 | 1 | -12/+12 |
| * | Avoid undefined behavior in negation in LSR. Patch by Ahmed Charles. | Eli Friedman | 2011-10-13 | 1 | -2/+2 |
| * | Enhance the memdep interface so that users can tell the difference between a ... | Eli Friedman | 2011-10-13 | 2 | -13/+13 |
| * | Fix a couple hash functions so that they do not depend on undefined shifts. ... | Eli Friedman | 2011-10-12 | 1 | -2/+2 |
| * | Add missing space. | Nick Lewycky | 2011-10-12 | 1 | -1/+1 |
| * | Fix PR11106 by correcting a typo that has been in the code for over a year. This | Cameron Zwarich | 2011-10-11 | 1 | -2/+2 |
| * | Remove a lot of the fancy scalar replacement code for dealing with llvm-gcc's | Cameron Zwarich | 2011-10-11 | 1 | -205/+16 |
| * | Add experimental -enable-lsr-phielim option. | Andrew Trick | 2011-10-11 | 1 | -0/+22 |
| * | Move replaceCongruentIVs into SCEVExapander and bias toward "expanded" | Andrew Trick | 2011-10-11 | 1 | -87/+6 |
| * | Add a natural stack alignment field to TargetData, and prevent InstCombine from | Lang Hames | 2011-10-10 | 1 | -2/+6 |
| * | LSR should only reuse phis that match its formula. | Andrew Trick | 2011-10-07 | 1 | -0/+1 |
| * | Teach GVN to also propagate switch cases. For example, in this code | Duncan Sands | 2011-10-07 | 1 | -31/+59 |
| * | Remove the old atomic instrinsics. autoupgrade functionality is included wit... | Eli Friedman | 2011-10-06 | 1 | -98/+1 |
| * | PR11061: Make simplifylibcalls fold strcmp("", x) correctly. | Eli Friedman | 2011-10-05 | 1 | -13/+18 |
| * | Re-commit 141203, but much more conservative. | Jim Grosbach | 2011-10-05 | 1 | -3/+3 |
| * | Revert 141203. InstCombine is looping on unit tests. | Jim Grosbach | 2011-10-05 | 1 | -9/+8 |
| * | Update InstCombine worklist after instruction transform is complete. | Jim Grosbach | 2011-10-05 | 1 | -8/+9 |
| * | GVN does simple propagation of conditions: when it sees a conditional | Duncan Sands | 2011-10-05 | 1 | -14/+111 |
| * | Generalize GVN's conditional propagation logic slightly: | Duncan Sands | 2011-10-05 | 1 | -4/+29 |
| * | LSR should avoid redundant edge splitting. | Andrew Trick | 2011-10-04 | 2 | -3/+6 |
| * | whitespace | Andrew Trick | 2011-10-04 | 1 | -18/+18 |
| * | Add a new icmp+select optz'n. Also shows off the load(cst) folding added in | Nick Lewycky | 2011-10-02 | 1 | -0/+6 |
| * | Enhance a couple places where we were doing constant folding of instructions, | Nick Lewycky | 2011-10-02 | 2 | -7/+8 |
| * | Inlining and unrolling heuristics should be aware of free truncs. | Andrew Trick | 2011-10-01 | 3 | -3/+11 |
| * | whitespace | Andrew Trick | 2011-10-01 | 2 | -18/+18 |
| * | Don't modify constant in-place. | Jim Grosbach | 2011-09-30 | 1 | -3/+4 |
| * | float comparison to double 'zero' constant can just be a float 'zero.' | Jim Grosbach | 2011-09-30 | 1 | -3/+6 |
| * | Tidy up. Trailing whitespace. | Jim Grosbach | 2011-09-30 | 1 | -220/+220 |
| * | Inlining often produces landingpad instructions with repeated | Duncan Sands | 2011-09-30 | 2 | -0/+338 |
| * | Fold two identical set lookups into one. No functionality change. | Nick Lewycky | 2011-09-29 | 1 | -4/+2 |
| * | When eliminating unnecessary retain+autorelease on return values, | Dan Gohman | 2011-09-29 | 1 | -1/+2 |
| * | Don't eliminate objc_retainBlock calls on stack objects if the | Dan Gohman | 2011-09-29 | 1 | -8/+29 |
| * | Clean up uses of switch instructions so they are not dependent on the operand... | Eli Friedman | 2011-09-29 | 2 | -7/+13 |
| * | typo + pasto | Andrew Trick | 2011-09-29 | 1 | -2/+2 |
| * | LSR: rewrite inner loops only. | Andrew Trick | 2011-09-29 | 1 | -5/+21 |
| * | indvars should hoist [sz]ext because licm is not rerun. | Andrew Trick | 2011-09-28 | 1 | -17/+27 |
| * | Stop emitting instructions with the name "tmp" they eat up memory and have to... | Benjamin Kramer | 2011-09-27 | 13 | -102/+79 |
| * | Split the landing pad basic block with the correct function. Also merge the | Bill Wendling | 2011-09-27 | 1 | -3/+23 |
| * | Disable LSR retry by default. | Andrew Trick | 2011-09-27 | 1 | -0/+16 |
| * | LSR, one of the new Cost::isLoser() checks did not get merged in the previous... | Andrew Trick | 2011-09-26 | 1 | -2/+6 |
| * | LSR cost metric minor fix and verification. | Andrew Trick | 2011-09-26 | 1 | -3/+26 |
| * | LSR minor bug fix in RateRegister. | Andrew Trick | 2011-09-23 | 1 | -1/+1 |
| * | PR10987: add a missed safety check to isSafePHIToSpeculate in scalarrepl. | Eli Friedman | 2011-09-22 | 1 | -7/+11 |
| * | Make sure IPSCCP never marks a tracked call as overdefined in SCCPSolver::Res... | Eli Friedman | 2011-09-20 | 1 | -9/+35 |
| * | Relax this condition. | Bill Wendling | 2011-09-20 | 1 | -2/+1 |
| * | Place the check for an exit landing pad where it will be run on both code pat... | Bill Wendling | 2011-09-20 | 1 | -6/+15 |
| * | Omit extracting a loop if one of the exits is a landing pad. | Bill Wendling | 2011-09-20 | 1 | -5/+11 |
| * | Check the terminator, not the basic block. | Bill Wendling | 2011-09-20 | 1 | -1/+2 |
| * | When extracting a basic block that ends in an 'invoke' instruction, we need to | Bill Wendling | 2011-09-20 | 1 | -7/+46 |
