- fc643c5 remove attribution from lib Makefiles. by Chris Lattner · 17 years ago
- 63079f0 Fold comparisons against a constant nan, and optimize ORD/UNORD by Chris Lattner · 17 years ago
- b253a8b make sure not to zap volatile stores, thanks a lot to Dale for noticing this! by Chris Lattner · 17 years ago
- 5938bef don't fold fp_round(fp_extend(load)) -> fp_round(extload) by Chris Lattner · 17 years ago
- 3c87285 Delete a store whose input is a load from the same pointer: by Chris Lattner · 17 years ago
- f1519e8 Flesh out the Briggs implementation a little bit more, fix a few FIXMEs. by Owen Anderson · 17 years ago
- efbcebc Sketch out an implementation of Briggs' copy placement algorithm. by Owen Anderson · 17 years ago
- 2ceb2cf initial code for forming an FGETSIGN node. This is disabled until by Chris Lattner · 17 years ago
- d268a49 improve support for fgetsign by Chris Lattner · 17 years ago
- b16f55f Tell TargetLoweringOpt whether it is running before or after legalize. by Chris Lattner · 17 years ago
- 1a3048b Add a new FGETSIGN operation, which defaults to expand on all targets. by Chris Lattner · 17 years ago
- a6c9925 Use getIntrinsicID instead of looking up intrinsic prototypes. Also by Gordon Henriksen · 17 years ago
- cae8d8d Note what still needs doing. by Owen Anderson · 17 years ago
- 22c43b2 Remove critical edge breaking. It won't be necessary as long as we are very careful when inserting copies. by Owen Anderson · 17 years ago
- 180c210 More accurate checks for two-address constraints. by Evan Cheng · 17 years ago
- 6894942 The physical register + virtual register joining requirement was much too strict. by Evan Cheng · 17 years ago
- d6c0758 Bring back a burr scheduling heuristic that's still needed. by Evan Cheng · 17 years ago
- 60ff1a3 Updated comments to reflect what "side effects" means in this situation. by Bill Wendling · 17 years ago
- 6f74b48 Simplify LowerCallTo by using a callsite. by Duncan Sands · 17 years ago
- 481dc72 The C++ exception handling personality function wants by Duncan Sands · 17 years ago
- 02132d6 Don't leave newly created nodes around if it turns out they are not needed. by Evan Cheng · 17 years ago
- 280f456 Add debugging info. Use the newly created "hasUnmodelledSideEffects" method. by Bill Wendling · 17 years ago
- feb8893 Support more insane CEP's in AsmPrinter (Yes, PyPy folks do really use them). by Anton Korobeynikov · 17 years ago
- 636e5a2 Fix PR1872: SrcValue and SrcValueOffset should not be used to compute load / store node id. by Evan Cheng · 17 years ago
- 8131013 Also print alignment and volatileness. by Evan Cheng · 17 years ago
- beec823 FIX for PR1799: When a load is unfolded from an instruction, check if it is a new node. If not, do not create a new SUnit. by Evan Cheng · 17 years ago
- 88ce93e SelectionDAG::dump() should print SrcValue of LoadSDNode and StoreSDNode. by Evan Cheng · 17 years ago
- fd7b326 Make invokes of inline asm legal. Teach codegen by Duncan Sands · 17 years ago
- 43ad6b3 Change the PointerType api for creating pointer types. The old functionality of PointerType::get() has become PointerType::getUnqual(), which returns a pointer in the generic address space. The new prototype of PointerType::get() requires both a type and an address space. by Christopher Lamb · 17 years ago
- 87a702b Break local interferences in StrongPHIElimination. One step closer... by Owen Anderson · 17 years ago
- 965b467 A few more comments. by Owen Anderson · 17 years ago
- ded2b0d Add explicit keywords, and fix a minor typo that they uncovered. by Dan Gohman · 17 years ago
- 7c07aeb Bug fix. Must also match ResNo when matching an operand with a user. by Evan Cheng · 17 years ago
- 62d67dd Add register pairs to the list to check for local interferences. by Owen Anderson · 17 years ago
- 4ba08ec Remove ugly and horrible code. It's not necessary for correctness, and can be added back later if it causes code quality issues. by Owen Anderson · 17 years ago
- 6e141fd Implicit def instructions, e.g. X86::IMPLICIT_DEF_GR32, are always re-materializable and they should not be spilled. by Evan Cheng · 17 years ago
- 8908132 Allow vector integer constants to be created with by Dan Gohman · 17 years ago
- ed2ffa2 Forgot to remove a register from the PHI-union after I'd determined that it by Owen Anderson · 17 years ago
- 39c883c If deleting a reload instruction due to reuse (value is available in register R and reload is targeting R), make sure to invalidate the kill information of the last kill. by Evan Cheng · 17 years ago
- efe2be7 Need to grow the indexed map. Added debug statements. by Bill Wendling · 17 years ago
- a17ad59 Simplify slightly. by Bill Wendling · 17 years ago
- d525f66 More progress on StrongPHIElimination. Now we actually USE the DomForest! by Owen Anderson · 17 years ago
- 12ebf14 Blark! How in the world did this work without this?! by Bill Wendling · 17 years ago
- 28bd5f0 - Update the virtual reg to machine instruction map when hoisting. by Bill Wendling · 17 years ago
- 650b052 Checking for "zero operands" during the "CanHoistInst()" method isn't necessary by Bill Wendling · 17 years ago
- 22f07ff Switch over to MachineLoopInfo. by Evan Cheng · 17 years ago
- ce25443 Pretty print shuffle mask operand. by Evan Cheng · 17 years ago
- ad93c4f CollectorMetadata and Collector are rejiggered to get along with by Gordon Henriksen · 17 years ago
- afc6de0 A little more progress on StrongPHIElimination, now that I have a better sense of by Owen Anderson · 17 years ago
- 11a4f64 Improve branch folding by recgonizing that explict successor relationships impact the value of fall-through choices. by Christopher Lamb · 17 years ago
- 46a6e79 Duncan points out that the subtraction is unneeded since hte code by Chris Lattner · 17 years ago
- 0097555 Add support for splitting the operand of a return instruction. by Chris Lattner · 17 years ago
- 041b3f8 Reverting 44702. It wasn't correct to rename them. by Bill Wendling · 17 years ago
- eeaad40 add many new cases to SplitResult. SplitResult now handles all the cases that LegalizeDAG does. by Chris Lattner · 17 years ago
- 5018786 Implement splitting support for store, allowing us to compile: by Chris Lattner · 17 years ago
- 697b53e implement vector splitting of load, undef, and binops. by Chris Lattner · 17 years ago
- e4af7b5 implement some methods. by Chris Lattner · 17 years ago
- 13c6a17 add scaffolding for splitting of vectors. by Chris Lattner · 17 years ago
- 7514646 reorganize header to separate into functional blocks. by Chris Lattner · 17 years ago
- ee88813 split scalarization out to its own file. by Chris Lattner · 17 years ago
- cc663a8 Split expansion out into its own file. by Chris Lattner · 17 years ago
- 08d1275 Split promotion support out to its own file. by Chris Lattner · 17 years ago
- 524825b Rename LegalizeDAGTypes.cpp -> LegalizeTypes.cpp by Chris Lattner · 17 years ago
- dff67f5 Split the class definition of DAGTypeLegalizer out into a header. by Chris Lattner · 17 years ago
- 320c630 Renaming: by Bill Wendling · 17 years ago
- b48519c Incorporated comments from Evan and Chris: by Bill Wendling · 17 years ago
- 0f940c9 Initial commit of the machine code LICM pass. It successfully hoists this: by Bill Wendling · 17 years ago
- dfb1561 Add an option to control this heuristic tweak so I can test it. by Evan Cheng · 17 years ago
- e0e6fac Redo previous patch so optimization only done for i1. Simpler and safer. by Dale Johannesen · 17 years ago
- 33faddc Turning simple splitting on. Start testing new coalescer heuristics as new llcbeta. by Evan Cheng · 17 years ago
- d3fd6d2 third time around: instead of disabling this completely, by Chris Lattner · 17 years ago
- e00cbad Actually, disable this code for now. More analysis and improvements to by Chris Lattner · 17 years ago
- e70da20 implement a readme entry, compiling the code into: by Chris Lattner · 17 years ago
- 72b2df1 implement the rest of the functionality from SelectionDAGLegalize::ScalarizeVectorOp by Chris Lattner · 17 years ago
- 59f1e97 Fix PR1842. by Dale Johannesen · 17 years ago
- 5ef3a04 Fix for PR1831: if all defs of an interval are re-materializable, then it's a preferred spill candiate. by Evan Cheng · 17 years ago
- d120ffd MachineInstr can change. Store indexes instead. by Evan Cheng · 17 years ago
- adf8590 If a split live interval is spilled again, remove the kill marker on its last use. by Evan Cheng · 17 years ago
- f38d14f Clobber more bugs. by Evan Cheng · 17 years ago
- b50bb8c Fix kill info for split intervals. by Evan Cheng · 17 years ago
- 60fb5ec more scalarization by Chris Lattner · 17 years ago
- d36db02 scalarize vector binops by Chris Lattner · 17 years ago
- 018f9b0 - Mark last use of a split interval as kill instead of letting spiller track it. by Evan Cheng · 17 years ago
- d64b5c8 Add a argument to storeRegToStackSlot and storeRegToAddr to specify whether by Evan Cheng · 17 years ago
- dd86ce3 Remove a unsafe optimization. This fixes 401.bzip2. by Evan Cheng · 17 years ago
- 35a3e4a Spiller unfold optimization bug: do not clobber a reusable stack slot value unless it can be modified. by Evan Cheng · 17 years ago
- eb1819d Implement framework for scalarizing node results. This is sufficient by Chris Lattner · 17 years ago
- 13c207b start providing framework for scalarizing vectors. by Chris Lattner · 17 years ago
- 597d10d Discard split intervals made empty due to folding. by Evan Cheng · 17 years ago
- e4b3900 Bug fixes. by Evan Cheng · 17 years ago
- a3355ff Rather than having special rules like "intrinsics cannot by Duncan Sands · 17 years ago
- 31d4318 Typo by Evan Cheng · 17 years ago
- 70306f8 Update kill info for uses of split intervals. by Evan Cheng · 17 years ago
- aee4af6 Remove redundant foldMemoryOperand variants and other code clean up. by Evan Cheng · 17 years ago
- e3110d0 Fix a bug where splitting cause some unnecessary spilling. by Evan Cheng · 17 years ago
- e62f97c Allow some reloads to be folded in multi-use cases. Specifically testl r, r -> cmpl [mem], 0. by Evan Cheng · 17 years ago
- cddbb83 Do not fold reload into an instruction with multiple uses. It issues one extra load. by Evan Cheng · 17 years ago
- 962e075 Provide a way to update DescGlobals cache directly. by Devang Patel · 17 years ago
- cb3c330 Do not lose rematerialization info when spilling already split live intervals. by Evan Cheng · 17 years ago
- 1953d0c Fix a major performance issue with splitting. If there is a def (not def/use) by Evan Cheng · 17 years ago