- 11af4ea fold fmul X, +2.0 -> fadd X, X; by Nate Begeman · 20 years ago
- 854077d add a trivial fold by Chris Lattner · 20 years ago
- 750dbd5 Fix this logic. by Chris Lattner · 20 years ago
- 85d63bb Add a case we were missing that was causing us to fail CodeGen/PowerPC/rlwinm.ll:test3 by Chris Lattner · 20 years ago
- 8136cda Use getExtLoad here instead of getNode, as extloads produce two values. This by Chris Lattner · 20 years ago
- 216def8 fold sext_in_reg, sext_in_reg where both have the same VT. This was by Nate Begeman · 20 years ago
- 5054f16 Relax the checking on zextload generation a bit, since as sabre pointed out by Nate Begeman · 20 years ago
- 24edbb7 Fix the trunc(load) case, finally allowing crafty and povray to pass by Chris Lattner · 20 years ago
- f988405 Fix some bugs in (sext (load x)) by Chris Lattner · 20 years ago
- 9ad8481 When ExpandOp'ing a [SZ]EXTLOAD, make sure to remember that the chain by Chris Lattner · 20 years ago
- 4c8f8f0 Implement PromoteOp for *EXTLOAD, allowing MallocBench/gs to Legalize by Chris Lattner · 20 years ago
- bfd65a0 Fix the remaining DAGCombiner issues pointed out by sabre. This should fix by Nate Begeman · 20 years ago
- 67a44cd Fix a minor bug in the dag combiner that broke pcompress2 and some other tests. by Chris Lattner · 20 years ago
- 144ff66 Add support to Legalize for expanding i64 sextload/zextload into hi and lo by Nate Begeman · 20 years ago
- 53c523c Inhibit instructions from being pushed before function calls. This will by Jim Laskey · 20 years ago
- ded4963 Move some Legalize functionality over to the DAGCombiner where it belongs. by Nate Begeman · 20 years ago
- 765784a Fix a potential bug with two combine-to's back to back that chris pointed by Nate Begeman · 20 years ago
- 3df4d52 More cool stuff for the dag combiner. We can now finally handle things by Nate Begeman · 20 years ago
- fab66f6 Finally committing to the new scheduler. Still -sched=none by default. by Jim Laskey · 20 years ago
- a179ab3 Fix a powerpc crash on CodeGen/Generic/llvm-ct-intrinsics.ll by Chris Lattner · 20 years ago
- 05b5743 Add a canonicalization that got lost, fixing PowerPC/fold-li.ll:SUB by Chris Lattner · 20 years ago
- 04ecf6d clean up some corner cases by Chris Lattner · 20 years ago
- 87514ca Implement trivial DSE. If two stores are neighbors and store to the same by Chris Lattner · 20 years ago
- 01a2202 Add support for CombineTo, allowing the dag combiner to replace nodes with by Chris Lattner · 20 years ago
- 07ed417 Teach the DAGCombiner several new tricks, teaching it how to turn by Nate Begeman · 20 years ago
- d739075 Fix comment by Chris Lattner · 20 years ago
- bba9aa3 Add ISD::ADD to MaskedValueIsZero by Chris Lattner · 20 years ago
- 3892560 This function is now dead by Chris Lattner · 20 years ago
- af21d55 Enable Nate's excellent DAG combiner work by default. This allows the by Chris Lattner · 20 years ago
- a19cea0 add a todo for something I noticed by Chris Lattner · 20 years ago
- ee899e6 (X & Y) & C == 0 if either X&C or Y&C are zero by Chris Lattner · 20 years ago
- 089c25c When emiting a CopyFromReg and the source is already a vreg, do not bother by Chris Lattner · 20 years ago
- f845b45 Lo and behold, the last bits of SelectionDAG.cpp have been moved over. by Nate Begeman · 20 years ago
- 9a9719e remove debugging code by Chris Lattner · 20 years ago
- c4ced26 implement CodeGen/PowerPC/div-2.ll:test2-4 by propagating zero bits through by Chris Lattner · 20 years ago
- 6c4dad0 fix indentation by Chris Lattner · 20 years ago
- 094c8fc Turn sdivs into udivs when we can prove the sign bits are clear. This by Chris Lattner · 20 years ago
- cf5c79b silence a bogus GCC warning by Chris Lattner · 20 years ago
- 32fca00 Make the legalizer completely non-recursive by Chris Lattner · 20 years ago
- 0558f61 Let the combiner handle more cases by Nate Begeman · 20 years ago
- 829cb81 Remove some bad code from Legalize by Nate Begeman · 20 years ago
- e17daeb Check in some more DAGCombiner pieces by Nate Begeman · 20 years ago
- 3ea0b47 implement visitBR_CC so that PowerPC/inverted-bool-compares.ll passes by Chris Lattner · 20 years ago
- ad13715 fix some pastos by Chris Lattner · 20 years ago
- 9503859 Add a new HandleNode class, which is used to handle (haha) cases in the by Chris Lattner · 20 years ago
- 5c46f74 Implement the code for PowerPC/inverted-bool-compares.ll, even though it by Chris Lattner · 20 years ago
- d48050a make sure that -view-isel-dags is the input to the isel, not the input to by Chris Lattner · 20 years ago
- 9155902 Fix a crash compiling Olden/tsp by Chris Lattner · 20 years ago
- 9d528dc Reverting to version - until problem isolated. by Jim Laskey · 20 years ago
- 1aa1972 Fix some faulty logic in the libcall inserter. by Nate Begeman · 20 years ago
- 27d404c Add back a workaround that fixes some breakages from chris's last change. by Nate Begeman · 20 years ago
- 8ba732b Refactor gathering node info and emission. by Jim Laskey · 20 years ago
- ab510a7 Fix a problem where the legalizer would run out of stack space on extremely by Chris Lattner · 20 years ago
- dcd5abc silence a bogus warning by Chris Lattner · 20 years ago
- a3638c0 Add assertions to the trivial scheduler to check that the value types match by Chris Lattner · 20 years ago
- 14765be Codegen CopyFromReg using the regclass that matches the valuetype of the by Chris Lattner · 20 years ago
- 505277a Add some very paranoid checking for operand/result reg class matchup by Chris Lattner · 20 years ago
- 2aeaf4e Fix VC++ warnings. by Jeff Cohen · 20 years ago
- 37345fe add a method by Chris Lattner · 20 years ago
- 22f6212 typo by Jim Laskey · 20 years ago
- 41755e2 1. Simplify the gathering of node groups. by Jim Laskey · 20 years ago
- b6d4c2c 1. Made things node-centric (from operand). by Jim Laskey · 20 years ago
- 473a990 Fix two bugs in my patch earlier today that broke int->fp conversion on X86. by Chris Lattner · 20 years ago
- fef80f4 Silence VC++ redeclaration warnings. by Jeff Cohen · 20 years ago
- 01b3d73 Add FP versions of the binary operators, keeping the int and fp worlds seperate. by Chris Lattner · 20 years ago
- 6c38b33 If the target prefers it, use _setjmp/_longjmp should be used instead of setjmp/longjmp for llvm.setjmp/llvm.longjmp. by Chris Lattner · 20 years ago
- 8e6be8b initialize new flag by Chris Lattner · 20 years ago
- 5324fec Remove some redundancies. by Jim Laskey · 20 years ago
- e6b90fb Addition of a simple two pass scheduler. This version is currently hacked up by Jim Laskey · 20 years ago
- 5ae7911 Turn (X^C1) == C2 into X == C1^C2 iff X&~C1 = 0 (and move a function) by Chris Lattner · 20 years ago
- 44728a7 Stub out the rest of the DAG Combiner. Just need to fill in the by Nate Begeman · 20 years ago
- 452d7be More DAG combining. Still need the branch instructions, and select_cc by Nate Begeman · 20 years ago
- fa57702 If a function has liveins, and if the target requested that they be plopped by Chris Lattner · 20 years ago
- 13d58e7 Allow targets to say they don't support truncstore i1 (which includes a mask by Chris Lattner · 20 years ago
- a500fc6 Add a missing #include, patch courtesy of Baptiste Lepilleur. by Chris Lattner · 20 years ago
- 3ec5d74 Fix a problem duraid encountered on itanium where this folding: by Chris Lattner · 20 years ago
- 08addbd Fix a crash viewing dags that have target nodes in them by Chris Lattner · 20 years ago
- 39ee1ac Last round of 2-node folds from SD.cpp. Will move on to 3 node ops such by Nate Begeman · 20 years ago
- 223df22 Move yet more folds over to the dag combiner from sd.cpp by Nate Begeman · 20 years ago
- 9980119 Another round of dag combiner changes. This fixes some missing XOR folds by Nate Begeman · 20 years ago
- f89d78d Implement a common missing fold, (add (add x, c1), c2) -> (add x, c1+c2). by Nate Begeman · 20 years ago
- 1e111c7 Fix a bug nate ran into with replacealluseswith. In the recursive cse case, by Chris Lattner · 20 years ago
- 2300f55 Add an option to the DAG Combiner to enable it for beta runs, and turn on by Nate Begeman · 20 years ago
- 83e75ec Next round of DAGCombiner changes. This version now passes all the tests by Nate Begeman · 20 years ago
- 6a8a21c Fix a checking failure in gs by Chris Lattner · 20 years ago
- 646d7e2 Next round of DAG Combiner changes. Just need to support multiple return by Nate Begeman · 20 years ago
- 6fdcb25 Clean up some code from the last checkin by Chris Lattner · 20 years ago
- 99c25b8 Fix a bug in legalize where it would emit two calls to libcalls that return by Chris Lattner · 20 years ago
- 70b9b10 Make sure to auto-cse nullary ops by Chris Lattner · 20 years ago
- 6621e3b Fix some buggy logic where we would try to remove nodes with two operands by Chris Lattner · 20 years ago
- fd88f64 Don't create zero sized stack objects even for array allocas with a zero by Chris Lattner · 20 years ago
- a639a43 Fix the release build, noticed by Eric van Riet Paap by Chris Lattner · 20 years ago
- 08951a3 Make sure to legalize assert[zs]ext's operand correctly by Chris Lattner · 20 years ago
- fab0887 For values that are live across basic blocks and need promotion, use ANY_EXTEND by Chris Lattner · 20 years ago
- 13c78e2 legalize ANY_EXTEND appropriately by Chris Lattner · 20 years ago
- 4ed11b4 Add support for ANY_EXTEND and add a few minor folds for it by Chris Lattner · 20 years ago
- 39f60a2 Fix some code in the current node combining code, spotted when it was moved by Nate Begeman · 20 years ago
- 4ebd805 Implement first round of feedback from chris (there's still a couple things by Nate Begeman · 20 years ago
- da8abb0 It is NDEBUG not _NDEBUG by Chris Lattner · 20 years ago
- 4942a96 Add the rest of the currently implemented visit routines to the switch by Nate Begeman · 20 years ago