- 99d9923 Fix overly conservative spill interval computation. by Alkis Evlogimenos · 21 years ago
- 3a9986f Beautify debug output. by Alkis Evlogimenos · 21 years ago
- d9a7349 Added the Malloc perl benchmark license information. by John Criswell · 21 years ago
- 9502658 Bug fixed by Chris Lattner · 21 years ago
- bf6fd9d New testcase for PR245 by Chris Lattner · 21 years ago
- 57cb988 Fix PR245: Linking weak and strong global variables is dependent on link order by Chris Lattner · 21 years ago
- 28a631d When we complete the bottom-up pass, make sure to merge the globals in 'main' into by Chris Lattner · 21 years ago
- 3c5df26 Since tail duplication is currently neutered, we have to help the optimizer to remove the overhead of __main by Chris Lattner · 21 years ago
- cc0d2f5 Add support for GlobalAddress's for alkis by Chris Lattner · 21 years ago
- 9b9b7fc These store to memory too. by Alkis Evlogimenos · 21 years ago
- 56ddada Remove the -disable-kill option. The register allocator is buggy with it, by Chris Lattner · 21 years ago
- acce13e These store to memory, not read from it. by Chris Lattner · 21 years ago
- 6df708d Added espresso. by John Criswell · 21 years ago
- d886ed9 Instructiosn with 1 memory operand have 4 operands in our representation.. duh! by Alkis Evlogimenos · 21 years ago
- f41dada Align case statements. by Alkis Evlogimenos · 21 years ago
- 14ffe75 Add TEST and XCHG memory operand support. by Alkis Evlogimenos · 21 years ago
- 68bff8e Add OR and XOR memory operand support. by Alkis Evlogimenos · 21 years ago
- e287a00 Peephole optimize SUBmi{16,32} into SUBmi{16,32}b when immediate is 8 bits wide. by Alkis Evlogimenos · 21 years ago
- cacca82 ADDmi{16,32} should be in the next case statement. by Alkis Evlogimenos · 21 years ago
- a7be982 Add memory operand folding support for MUL, DIV, IDIV, NEG, NOT, by Alkis Evlogimenos · 21 years ago
- 89b0214 Add memory operand folding for CMP{rm,mr,mi}{8,16,32}, INCm{8,16,32} by Alkis Evlogimenos · 21 years ago
- 509d6d6 Add CMP{rm,mr,mi}{8,16,32}, INCm{8,16,32} and DECm{8,16,32} instructions. by Alkis Evlogimenos · 21 years ago
- 503770c Add SUB{rm,mr,mi}{8,16,32} instructions. by Alkis Evlogimenos · 21 years ago
- 11390e7 Add support to the local allocator for fusing spill code into the instructions by Chris Lattner · 21 years ago
- 18bd7bb Add support for folding memory operands for ADC, SBB and SUB instructions. by Alkis Evlogimenos · 21 years ago
- 69da6db Add support for ADC{rm.mr}32 and SBB{rm,mr}32. by Alkis Evlogimenos · 21 years ago
- ee0919b Add a (hidden) option to print instructions that fail to fuse. It's looking by Chris Lattner · 21 years ago
- 17dc674 Add support for folding memory operands in MOVri{8,16,32} instructions. by Alkis Evlogimenos · 21 years ago
- 5b67252 Expand the repertoire of the forms we can print and encode. by Chris Lattner · 21 years ago
- 651fd55 Disable this peephole for now. We can't keep track of the fact that the immediate is 8 bits, by Chris Lattner · 21 years ago
- cdee287 Fix a bug in my previous refactoring change... arg! by Chris Lattner · 21 years ago
- c4601bf The CBE is now in llc, not llvm-dis by Chris Lattner · 21 years ago
- 9915cd9 The C backend is no longer in llvm-dis, it's in llc by Chris Lattner · 21 years ago
- df2cf41 The CBE is no longer in llvm-dis by Chris Lattner · 21 years ago
- a7660be Add an option to disable spill fusing in the X86 backend by Chris Lattner · 21 years ago
- e87331d Fix the mneumonics for the mov instructions to have the source and destination by Chris Lattner · 21 years ago
- 7ddc3fb Fix the last crimes against nature that used the 'ir' ordering to use the by Chris Lattner · 21 years ago
- f120ebb GRRR. Move instructions have swapped the order of the r/m operands. by Chris Lattner · 21 years ago
- 6e173a0 Rename MOVi[mr] instructions to MOV[rm]i by Chris Lattner · 21 years ago
- 90c38c8 Whoops, got my cases swapped. by Chris Lattner · 21 years ago
- 5a051f6 Change to match the newer, simpler, interface by Chris Lattner · 21 years ago
- 8026a69 Simplify and document the new interface by Chris Lattner · 21 years ago
- 0f9c491 Add support for folding memory operands into AND and IMUL's by Chris Lattner · 21 years ago
- 7c035b7 Scrunchify code, by adding helpers. No functionality changes. by Chris Lattner · 21 years ago
- 4ff7876 Add mem forms of AND instructions by Chris Lattner · 21 years ago
- 7093d37 Add LiveIntervals::Interval::empty() member function. by Alkis Evlogimenos · 21 years ago
- b499866 Add API to check and fold memory operands into instructions. by Alkis Evlogimenos · 21 years ago
- 55b5481 Rename the IMULri* instructions to IMULrri, as they are actually three address by Chris Lattner · 21 years ago
- 1e3812c Once we have a way to fold spill code reloads into instructions, we have a way to use it. :) by Chris Lattner · 21 years ago
- 355a724 Fix spilled interval update. It was too conservative. by Alkis Evlogimenos · 21 years ago
- 42e0a8f Refactor code a bit. No functionality changes, though the comment hints at things to come. by Chris Lattner · 21 years ago
- ac89164 Adjust to recent changes by Chris Lattner · 21 years ago
- fcdb2c2 Add two missing pieces from last checkin by Chris Lattner · 21 years ago
- 3e3bcbd Rearrange code to eliminate warnings by Chris Lattner · 21 years ago
- 31bbb51 Add peephole optimizations for ADD [MEM], IMM8 instructions. by Alkis Evlogimenos · 21 years ago
- 7271e0e Add two more variants of add. Update comments. by Alkis Evlogimenos · 21 years ago
- 7727087 Preemptive additions for the MallocBench tests. by John Criswell · 21 years ago
- 304e143 Only spit out warning for functions that take pointers, not for sin and the like by Chris Lattner · 21 years ago
- 2e9bb1a Move the folding of gep null, 0, 0, 0 to a place where it can be shared and by Chris Lattner · 21 years ago
- 2ca2652 PreSelection isn't optional and so there's no way to turn it off. by Misha Brukman · 21 years ago
- eee33b2 memset and bcopy and now unified by the llvm.memset intrinsic by Chris Lattner · 21 years ago
- e0f636a Add some ADD instructions that take memory operands for Alkis by Chris Lattner · 21 years ago
- 542af5f Adjusted the default pathname to the SPEC95 benchmarks. The new directory by John Criswell · 21 years ago
- aad5c05 Add LeakDetection to MachineInstr. by Alkis Evlogimenos · 21 years ago
- 19831ec Implement test/Regression/Transforms/SimplifyCFG/UncondBranchToReturn.ll, by Chris Lattner · 21 years ago
- 5ea27d8 New testcase, details in the comments by Chris Lattner · 21 years ago
- 7059f2e Fold PHI nodes of constants which are only used by a single cast. This implements by Chris Lattner · 21 years ago
- d8bed76 A new testcase for a situation that occurs in 181.mcf by Chris Lattner · 21 years ago
- 26ca7e1 Teach LLVM to unravel the "swap idiom". This implements: by Chris Lattner · 21 years ago
- 72695f7 Add a test for the "swap idiom", which LLVM should be able to unravel. by Chris Lattner · 21 years ago
- c317d39 Implement Transforms/InstCombine/xor.ll:test19 by Chris Lattner · 21 years ago
- a783c02 Test for xor chains by Chris Lattner · 21 years ago
- 8709927 Fix a bug in the recent rewrite of the leakdetector that caused all of the by Chris Lattner · 21 years ago
- 2e6fcaf Document bug 240 by Chris Lattner · 21 years ago
- 3a0628a Moving CBE tests to test/Regression/CodeGen/CBackend by Chris Lattner · 21 years ago
- 74661c8 When the user runs 'llc foo.bc -march=c', write the output to "foo.cbe.c", not by Chris Lattner · 21 years ago
- e42cde2 Now that the lowerinvoke pass inserts calls to llvm.setjmp/llvm.longjmp, some by Chris Lattner · 21 years ago
- 9b700f7 By default, llvm.setjmp/llvm.longjmp intrinsics get lowered to their libc counterparts by Chris Lattner · 21 years ago
- 860a161 Instead of producing calls to setjmp/longjmp, produce uses of the by Chris Lattner · 21 years ago
- 588e72d Refactor code. Now the intrinsic lowering pass tries to recycle preexisting by Chris Lattner · 21 years ago
- b83530f New testcase by Chris Lattner · 21 years ago
- f11800b This pass should not require phi elimination or live variable by Alkis Evlogimenos · 21 years ago
- 5aaf1d2 Finegrainify namespacification by Chris Lattner · 21 years ago
- 859a18b Make dense maps keyed on physical registers smallerusing by Alkis Evlogimenos · 21 years ago
- 843b160 Eliminate the use of spill (reserved) registers. by Alkis Evlogimenos · 21 years ago
- 9f3c7f7 QOI bug fixed by Chris Lattner · 21 years ago
- de512b5 Adjustments to support the new ConstantAggregateZero class by Chris Lattner · 21 years ago
- cfb0fd2 Add support for the new ConstantAggregateZero class by Chris Lattner · 21 years ago
- dd2c82a Make the JIT zero out globals with memset instead of an element at a time. This by Chris Lattner · 21 years ago
- 896481e No need to scan zero initializers. This should make DSA a bit faster. by Chris Lattner · 21 years ago
- 40bbeb5 Add a new ConstantAggregateZero class, to fix PR239. This makes zero by Chris Lattner · 21 years ago
- 46ccf6d finegrainify namespacification by Chris Lattner · 21 years ago
- d6a185f Bug fixed by Chris Lattner · 21 years ago
- ca705fa ConstantArray::get and ConstantStruct::get now just return pointers to by Chris Lattner · 21 years ago
- 04d1fb6 Remove dependence on return type of ConstantStruct::get by Chris Lattner · 21 years ago
- 3710644 Remove dependence on the return type of ConstantArray::get by Chris Lattner · 21 years ago
- aab54da Keep a cache of non-abstract null arrays and structs. This speeds up llvm-dis by Chris Lattner · 21 years ago
- 07b976b Testcase for PR205: [llvmgcc] C front-end does not emit 'zeroinitializer' when possible by Chris Lattner · 21 years ago
- 8710da3 Test for actual support, not just for lack of crashage by Chris Lattner · 21 years ago
- 0e5b79c Implement llvm.(frame|return)address(0) correctly. They are used by the LLVM JIT, among other by Chris Lattner · 21 years ago