- e4d5c44 This mega patch converts us from using Function::a{iterator|begin|end} to by Chris Lattner · 19 years ago
- 572dd08 Silence some uninit variable warnings. by Chris Lattner · 19 years ago
- 3576c84 We can fold promoted and non-promoted loads into divs also! by Chris Lattner · 20 years ago
- 44129b5 Fold promoted loads into binary ops for FP, allowing us to generate m32 forms by Chris Lattner · 20 years ago
- 0a07883 Silence a warning. by Chris Lattner · 20 years ago
- fb0f53f Speed up folding operations into loads. by Chris Lattner · 20 years ago
- 67b1c3c The ever-important vanity pass name :) by Chris Lattner · 20 years ago
- b62e1e2 Fix a FIXME: realize that argument stores are all independent (don't alias) by Chris Lattner · 20 years ago
- 19ad062 Implement ADD_PARTS/SUB_PARTS so that 64-bit integer add/sub work. This by Chris Lattner · 20 years ago
- bf52d49 Fix a crash compiling 134.perl. by Chris Lattner · 20 years ago
- c3c021b Fix a problem where were were literally selecting for INCREASED register by Chris Lattner · 20 years ago
- 4053b1e Implement Regression/CodeGen/X86/rotate.ll: emit rotate instructions (which by Chris Lattner · 20 years ago
- 30ea1e9 Match 16-bit shld/shrd instructions as well, implementing shift-double.llx:test5 by Chris Lattner · 20 years ago
- 8571637 Codegen long >> 2 to this: by Chris Lattner · 20 years ago
- 009b55b X86 shifts mask the amount. by Chris Lattner · 20 years ago
- 477c931 Code to handle FP_EXTEND is dead now. X86 doesn't support any data types to by Chris Lattner · 20 years ago
- 3dd6815 Remove more dead code. by Chris Lattner · 20 years ago
- a7a4c0f The selection dag code handles the promotions from F32 to F64 for us, so we by Chris Lattner · 20 years ago
- b422aea Fix 124.m88ksim. by Chris Lattner · 20 years ago
- 811482a Do not emit loads multiple times, potentially in the wrong places. by Chris Lattner · 20 years ago
- b71f8fc Eliminate bad assertions. by Chris Lattner · 20 years ago
- 4a10866 * Eliminate the TokenSet and just use the ExprMap for both tokens and values. by Chris Lattner · 20 years ago
- 883c86f Rely on the code in MatchAddress to do this work. Otherwise we fail to by Chris Lattner · 20 years ago
- 98a8ba0 Fix a problem where probing for addressing modes caused expressions to be by Chris Lattner · 20 years ago
- 869e043 Don't bother using max here. by Chris Lattner · 20 years ago
- 90d1be7 Do not give token factor nodes outrageous weights by Chris Lattner · 20 years ago
- 5c65981 Two changes: by Chris Lattner · 20 years ago
- e10269b Refactor load/op/store folding into it's own method, no functionality changes. by Chris Lattner · 20 years ago
- ae3ffba Fix a major regression last night that prevented us from producing [mem] op= reg by Chris Lattner · 20 years ago
- 2a4e508 Codegen this: by Chris Lattner · 20 years ago
- 4ff348b Fix test/Regression/CodeGen/X86/2005-01-17-CycleInDAG.ll and 132.ijpeg. by Chris Lattner · 20 years ago
- cb1aa8d Do not fold a load into a comparison that is used by more than one place. by Chris Lattner · 20 years ago
- c98279d Do not codegen 'xor bool, true' as 'not reg'. not reg inverts the upper bits by Chris Lattner · 20 years ago
- 4df0de9 Set up the shift and setcc types. by Chris Lattner · 20 years ago
- da2ce11 * Adjust to changes in TargetLowering interfaces. by Chris Lattner · 20 years ago
- e9ef81d Add support for truncstore and *extload. by Chris Lattner · 20 years ago
- 7f2afac Adjust to CopyFromREg changes. by Chris Lattner · 20 years ago
- 18c2f13 Add new ImplicitDef node, rename CopyRegSDNode class to RegSDNode. by Chris Lattner · 20 years ago
- 1d50b7f Codegen factor nodes more intelligently according to perceived register pressure. by Chris Lattner · 20 years ago
- c358071 Initial trivial (but stupid) codegen for this node. by Chris Lattner · 20 years ago
- 636e79a Add some really pedantic assertions to the load folding code. Fix a bunch by Chris Lattner · 20 years ago
- a0bb692 We can only fold a load into an op if there is exactly one use of the value. by Chris Lattner · 20 years ago
- 9bbd992 Try both ways to fold an add together. This allows us to generate this code by Chris Lattner · 20 years ago
- 2b93786 Fix a major miscompilation where we were overwriting the scale reg. by Chris Lattner · 20 years ago
- 48034fd Do not use the type of the RHS constant to determine the type of the operation. by Chris Lattner · 20 years ago
- 603fea9 Fix C++ more compilatiom errors by Jeff Cohen · 20 years ago
- a56cea4 Fix a compile error with VC++, which things that static const arrays need by Chris Lattner · 20 years ago
- 7dbcb75 Fix a bug that caused us to crash on povray. We weren't emitting an FP_REG_KILL into a block that had a successor with a FP PHI node. by Chris Lattner · 20 years ago
- 085c995 Fix a crash compiling povray on UINT_TO_FP from i16. by Chris Lattner · 20 years ago
- 4292830 There are no [mem] op= reg instructions for FP, so remove their entries. by Chris Lattner · 20 years ago
- db8c368 Fix a bug where we didn't insert FP_REG_KILL instructions into MBB's that by Chris Lattner · 20 years ago
- afce430 Fold TRUNCATE (LOAD P) into a smaller load from P. by Chris Lattner · 20 years ago
- ef6806c Be more careful about order of arg evalution for CopyToReg nodes. This shrinks by Chris Lattner · 20 years ago
- 7ea64f5 Fold some more [mem] op= val operators. This allows us to things like this by Chris Lattner · 20 years ago
- dbba22f Fold loads into sign/zero extends. instead of: by Chris Lattner · 20 years ago
- 837caa7 Comment out debug code :) by Chris Lattner · 20 years ago
- c4b6a78 Treat int constants as not requiring a register, since they are almost always by Chris Lattner · 20 years ago
- a5ade06 * Factor a bunch of binary operator cases into shared code. by Chris Lattner · 20 years ago
- 947d544 Fold multiplies by 3,5,9 into addressing modes when possible. by Chris Lattner · 20 years ago
- 51a2634 Instead of generating stuff like this: by Chris Lattner · 20 years ago
- 31805bf Implement MEMCPY natively in terms of rep movs* by Chris Lattner · 20 years ago
- 989de03 Implement memset -> rep stos* by Chris Lattner · 20 years ago
- 795069d Announce that we don't support mem ops yet. by Chris Lattner · 20 years ago
- a95589b Teach the address selector to make 'reg+reg' addressing modes. by Chris Lattner · 20 years ago
- d4dab92 Emit NOT instructions. by Chris Lattner · 20 years ago
- 6c07aee Fix a bug emitting branches that broke a lot of programs. by Chris Lattner · 20 years ago
- ef7ba07 Be more careful where we set ContainsFPCode. We were missing a set in the by Chris Lattner · 20 years ago
- a3aa2e2 Fix a major bug in setcc/cmov folding, where we accidentally by Chris Lattner · 20 years ago
- 1133309 Take register pressure into account when we have to decide whether to by Chris Lattner · 20 years ago
- 24aad1b Fold setcc instructions into selects. by Chris Lattner · 20 years ago
- a13d323 Implement 8-bit multiply for X86. by Chris Lattner · 20 years ago
- bd9f0ee Codegen (Reg|imm)+&GV as an LEA, because we cannot put it into the immediate field by Chris Lattner · 20 years ago
- 2502085 Fix copy and pasto's for FP -> Int. This fixes fldry by Chris Lattner · 20 years ago
- 590d800 Initial implementation of FP->INT and INT->FP casts by Chris Lattner · 20 years ago
- 1482458 Implement varargs and returnaddress/frameaddress intrinsics. With this by Chris Lattner · 20 years ago
- b080265 Okay 15th time is the charm. Looking at the vector size is useless as it by Chris Lattner · 20 years ago
- 6cc70ef Okay, my off by one was actually off by two. This fixes Generic/2003-07-07-BadLongConst.ll by Chris Lattner · 20 years ago
- ce45d65 Fix off by one error by Chris Lattner · 20 years ago
- 5188ad7 Adjust to changes in LowerCallTo interface Minor bugfixes by Chris Lattner · 20 years ago
- 8acb1ba Reimplementation of the X86 pattern isel. This is still missing many large by Chris Lattner · 20 years ago
- d844d01 This file is now dead. by Chris Lattner · 20 years ago
- f60b91c Disable the pattern isel by Chris Lattner · 20 years ago
- 91b5ca8 Fix file header as it has been renamed. by Misha Brukman · 20 years ago
- 954da37 Add #include <iostream> since Value.h does not #include it any more. by Reid Spencer · 20 years ago
- f70e0c2 Clean up a lot of the code I added yesterday by exposing the IntrinsicLowering by Chris Lattner · 21 years ago
- 4482715 implement support for the intrinsic lowering functionality by Chris Lattner · 21 years ago
- d0fde30 Put all LLVM code into the llvm namespace, as per bug 109. by Brian Gaeke · 21 years ago
- b576c94 Added LLVM project notice to the top of every C++ source file. by John Criswell · 21 years ago
- f3c274d Adjust to new API, add expandCall stub by Chris Lattner · 21 years ago
- 19df387 Factory methods for FunctionPasses now return type FunctionPass *. by Brian Gaeke · 21 years ago
- ac0c868 Add support for a pattern matching instruction selector. This is still in by Chris Lattner · 21 years ago