- 329b590 Re-revert r173342, without losing the compile time improvements, flat by Chandler Carruth · 13 years ago
- ceff222 Switch this code away from Value::isUsedInBasicBlock. That code either by Chandler Carruth · 13 years ago
- 1c4e323 Reapply chandlerc's r173342 now that the miscompile it was triggering is fixed. by Benjamin Kramer · 13 years ago
- 321c6a7 Revert r173342 temporarily. It appears to cause a very late miscompile by Chandler Carruth · 13 years ago
- 5f45193 Plug TTI into the speculation logic, giving it a real cost interface by Chandler Carruth · 13 years ago
- 01bffaa Address a large chunk of this FIXME by accumulating the cost for by Chandler Carruth · 13 years ago
- 8a21005 Switch the constant expression speculation cost evaluation away from by Chandler Carruth · 13 years ago
- 7481ca8 Rephrase the speculating scan of the conditional BB to be phrased in by Chandler Carruth · 13 years ago
- 76aacbd Simplify the PHI node operand rewriting. by Chandler Carruth · 13 years ago
- e2a779f Give the basic block variables here names based on the if-then-end by Chandler Carruth · 13 years ago
- 1d20c02 Lift a cheap early exit test above loops and other complex early exit by Chandler Carruth · 13 years ago
- 8a4a166 Spiff up the comment on this method, making the example a bit more by Chandler Carruth · 13 years ago
- 5924545 Initialize the components of this class. Otherwise GCC thinks that Array may be by Duncan Sands · 13 years ago
- 0b4ef9c Make SimplifyCFG simply depend upon TargetTransformInfo and pass it by Chandler Carruth · 13 years ago
- d3e7355 Move TargetTransformInfo to live under the Analysis library. This no by Chandler Carruth · 13 years ago
- 6db43e6 Switch SimplifyCFG over to the TargetTransformInfo interface rather than by Chandler Carruth · 13 years ago
- 9fb823b Move all of the header files which are involved in modelling the LLVM IR by Chandler Carruth · 13 years ago
- ed0881b Use the new script to sort the includes of every file under lib. by Chandler Carruth · 13 years ago
- d9ef81e Fix non-determinism introduced in r168970 and pointed out by Duncan. by Chandler Carruth · 13 years ago
- 77d433d Rearrange the comments, control flow, and variable names; no by Chandler Carruth · 13 years ago
- 65df808 Fix logic to determine whether to turn a switch into a lookup table. When by Evan Cheng · 13 years ago
- 7b8af0e SimplifyCFG: Don't assume non-null ScalarTargetTransformInfo. by Hans Wennborg · 13 years ago
- 7656f6d misspell by Andrew Trick · 13 years ago
- 90f5029 whitespace by Andrew Trick · 13 years ago
- c3c8d95 Only do switch-to-lookup table transformation when TargetTransformInfo by Hans Wennborg · 13 years ago
- 7ec5085 Revert the series of commits starting with r166578 which introduced the by Chandler Carruth · 13 years ago
- b71f72a Remove fixme about unreachable cases from SwitchToLookupTable by Hans Wennborg · 13 years ago
- 4fef2fe Address Duncan's comments on r167121. by Hans Wennborg · 13 years ago
- 09acdb9 Address Duncan's comments on r167115 by Hans Wennborg · 13 years ago
- 793b342 Fix false -> NULL conversion from r167115 spotted by Benjamin Kramer. by Hans Wennborg · 13 years ago
- 9e74dd9 Do simple constant propagation in lookup table formation for switches by Hans Wennborg · 13 years ago
- f325483 Use TargetTransformInfo to control switch-to-lookup table transformation by Hans Wennborg · 13 years ago
- 5bdd9dd Remove a wrapper around getIntPtrType added to GVN by Hal in commit 166624 (the by Duncan Sands · 13 years ago
- 977f41a Also optimize large switch statements. by Jakob Stoklund Olesen · 13 years ago
- 12d9127 Add in support for getIntPtrType to get the pointer type based on the address space. by Micah Villmow · 13 years ago
- c5b0678 Simplify code. No functionality change. by Benjamin Kramer · 13 years ago
- 97c1876 PGO: create metadata for switch only if it has more than one targets. by Manman Ren · 13 years ago
- cdfe20b Move TargetData to DataLayout. by Micah Villmow · 13 years ago
- d12e82e SimplifyCFG: Enhance the "remove CFG edge that leads to null pointer dereference" optimization to also handle instructions with multiple uses. by Benjamin Kramer · 13 years ago
- 9fc3dc7 SimplifyCFG: Don't crash when forming a switch bitmap with an undef default value. by Benjamin Kramer · 13 years ago
- f064b65 SimplifyCFG: Enumerating all predecessors of a BB can be expensive (switches), avoid it if possible. by Benjamin Kramer · 13 years ago
- c2081d1 Fix a integer overflow in SimplifyCFG's look up table formation logic. by Benjamin Kramer · 13 years ago
- cd3a11f Address Duncan's comments on r164684: by Hans Wennborg · 13 years ago
- f2e2c10 Address Duncan's comments on r164682: by Hans Wennborg · 13 years ago
- 39583b8 SimplifyCFG: Make the switch-to-lookup table transformation store the by Hans Wennborg · 13 years ago
- 776d712 SimplifyCFG: Refactor the switch-to-lookup table transformation by by Hans Wennborg · 13 years ago
- 93ab649 SimplifyCFG: sink common codes from IF, ELSE blocks down to END block. by Manman Ren · 13 years ago
- f744fa9 SimplifyCFG: Don't generate invalid code for switch used to initialize by Hans Wennborg · 13 years ago
- 5657555 PGO: preserve branch-weight metadata when simplifying Switch to a sub, an icmp by Manman Ren · 13 years ago
- ce48ea7 PGO: preserve branch-weight metadata when simplifying Switch by Manman Ren · 13 years ago
- 774246a PGO: preserve branch-weight metadata when simplifying SwitchOnSelect. by Manman Ren · 13 years ago
- 2d4c10f PGO: preserve branch-weight metadata when simplifying two branches with a common by Manman Ren · 13 years ago
- 4a12706 Fix a few vars that can end up being used without initialization. by Axel Naumann · 13 years ago
- bfb9d43 PGO: preserve branch-weight metadata when simplifying two branches with a common by Manman Ren · 13 years ago
- 8691e52 PGO: preserve branch-weight metadata when simplifying a switch with a single by Manman Ren · 13 years ago
- 5e5049d Try to fix the bots by detecting inconsistant branch-weight metadata. by Manman Ren · 13 years ago
- d81b8e8 PGO: preserve branch-weight metadata when merging two switches where by Manman Ren · 13 years ago
- 571d9e4 SimplifyCFG: preserve branch-weight metadata when creating a new switch from by Manman Ren · 13 years ago
- 7fd5c844 Fix style issues from r163302 pointed out by Evan. by Hans Wennborg · 13 years ago
- d3b4d2c Remove an incorrect assert during branch weight propagation. by Andrew Trick · 13 years ago
- 08238ad SimplifyCFG: ValidLookupTableConstant should be static by Hans Wennborg · 13 years ago
- feb4d07 Fix switch_to_lookup_table.ll test from r163302. by Hans Wennborg · 13 years ago
- 8a62fc5 Build lookup tables for switches (PR884) by Hans Wennborg · 13 years ago
- ad06cee Stop casting away const qualifier needlessly. by Roman Divacky · 13 years ago
- 30c3e14 test by Michael Ilseman · 13 years ago
- 3051aa1 Preserve branch profile metadata during switch formation. by Andrew Trick · 13 years ago
- f3cf193 whitespace by Andrew Trick · 13 years ago
- 35521e2 Fix a typo (the the => the) by Sylvestre Ledru · 13 years ago
- ec7ad65 Move llvm/Support/MDBuilder.h to llvm/MDBuilder.h, to live with by Chandler Carruth · 13 years ago
- abbfe69 Make helper functions static. by Benjamin Kramer · 13 years ago
- b65acc6 Revert "IntRange:" as it appears to be breaking self hosting. by Eric Christopher · 13 years ago
- 8b9ecca IntRange: by Stepan Dyatkovskiy · 13 years ago
- aafe091 Move llvm/Support/IRBuilder.h -> llvm/IRBuilder.h by Chandler Carruth · 13 years ago
- 0a045bb Remove dyn_cast + dereference pattern by replacing it with a cast and changing by Nick Lewycky · 13 years ago
- d33f4ef SimplifyCFG: fold unconditional branch to its predecessor if profitable. by Manman Ren · 13 years ago
- 58abf4f SimplifyCFG: Turn the ad-hoc std::pair that represents switch cases into an explicit struct. by Benjamin Kramer · 13 years ago
- 65e7566 Add support for branch weight metadata to MDBuilder and use it in various places. by Benjamin Kramer · 13 years ago
- ba0a6ca Always compute all the bits in ComputeMaskedBits. by Rafael Espindola · 14 years ago
- 97b02fc llvm::SwitchInst by Stepan Dyatkovskiy · 14 years ago
- 5b648af Taken into account Duncan's comments for r149481 dated by 2nd Feb 2012: by Stepan Dyatkovskiy · 14 years ago
- d5d95b0 [unwind removal] We no longer have 'unwind' instructions being generated, so by Bill Wendling · 14 years ago
- 513aaa5 SwitchInst refactoring. by Stepan Dyatkovskiy · 14 years ago
- 3c3feaf Gracefully degrade precision in branch probability numbers. by Nick Lewycky · 14 years ago
- 219e6bc Actually, this code handles wrapped sets just fine. Noticed by inspection. by Nick Lewycky · 14 years ago
- 5ab9c0a Fix SpeculativelyExecuteBB to either speculate all or none of the phis by Dan Gohman · 14 years ago
- 5267211 Revert r56315. When the instruction to speculate is a load, this by Dan Gohman · 14 years ago
- 8640fdf Demystify this comment. by Nick Lewycky · 14 years ago
- 398255e Use false not zero, as a bool. by Nick Lewycky · 14 years ago
- c554a9b Teach simplifycfg to recompute branch weights when merging some branches, and by Nick Lewycky · 14 years ago
- 8d302df Update the branch weight metadata when reversing the order of a branch. by Nick Lewycky · 14 years ago
- e87d54c Sort includes, canonicalize whitespace, fix typos. No functionality change. by Nick Lewycky · 14 years ago
- b4039f6 Make some intrinsics safe to speculatively execute. by Nick Lewycky · 14 years ago
- 8b3deab Revert r146822 at Pete Cooper's request as it broke clang self hosting. by Kevin Enderby · 14 years ago
- eadf124 SimplifyCFG now predicts some conditional branches to true or false depending on previous branch on same comparison operands. by Pete Cooper · 14 years ago
- 75d7d5e Move Instruction::isSafeToSpeculativelyExecute out of VMCore and by Dan Gohman · 14 years ago
- 29192d0 Delete trivial landing pads that just continue unwinding the caught exception. by Duncan Sands · 14 years ago
- 0655b78 Address review comments. by Benjamin Kramer · 14 years ago
- fb212a6 SimplifyCFG: If we have a PHI node that can evaluate to NULL and do a load or store to the address returned by the PHI node then we can consider this incoming value as dead and remove the edge pointing there, unless there are instructions that can affect control flow executed in between. by Benjamin Kramer · 14 years ago
- 55d875f I think there was some confusion about what I meant. :-) Replacing the comment. by Bill Wendling · 14 years ago
- bd39703 After talking with Bill, it seems like the LandingPad handling here is likely by Eli Friedman · 14 years ago