| Commit message (Expand) | Author | Age | Files | Lines |
| * | Increase default setting of tail-merge-threshold to | Dale Johannesen | 2008-10-27 | 1 | -1/+1 |
| * | Do not shrink wrap live interval in a mbb if it's livein any of its successor... | Evan Cheng | 2008-10-26 | 1 | -6/+21 |
| * | Handle cases where there aren't uses in the barrier mbb. | Evan Cheng | 2008-10-25 | 1 | -1/+5 |
| * | SDNodes may have at most one Flag result. Update this comment | Dan Gohman | 2008-10-25 | 1 | -2/+2 |
| * | Move the code that adds the DeadMachineInstructionElimPass from | Dan Gohman | 2008-10-25 | 1 | -4/+0 |
| * | If val# def is ~0U, meaning it's defined by a PHI, and it's previously split,... | Evan Cheng | 2008-10-25 | 1 | -9/+11 |
| * | Fix a pasto. | Evan Cheng | 2008-10-24 | 1 | -1/+1 |
| * | Fix a end() dereference; remove an abort() that wasn't meant to be left in. | Evan Cheng | 2008-10-24 | 1 | -5/+4 |
| * | Avoid splitting an interval multiple times; avoid splitting re-materializable... | Evan Cheng | 2008-10-24 | 2 | -51/+115 |
| * | Initialize uninitialized variable. | Dale Johannesen | 2008-10-24 | 1 | -1/+1 |
| * | Committing a good chunk of the pre-register allocation live interval splittin... | Evan Cheng | 2008-10-23 | 2 | -12/+578 |
| * | Fix thinko - the operand number has nothing to do | Duncan Sands | 2008-10-23 | 2 | -3/+3 |
| * | LegalizeTypes soft-float support for fpow. | Duncan Sands | 2008-10-22 | 2 | -1/+15 |
| * | Be nice to CellSPU: for this target getSetCCResultType | Duncan Sands | 2008-10-22 | 1 | -2/+8 |
| * | Port from LegalizeDAG the logic to only generate | Duncan Sands | 2008-10-22 | 1 | -8/+44 |
| * | Add some comments explaining the meaning of a boolean | Duncan Sands | 2008-10-22 | 1 | -4/+2 |
| * | Temporarily allow the operands of a BUILD_VECTOR | Duncan Sands | 2008-10-22 | 1 | -4/+8 |
| * | Move Print*Pass to use raw_ostream. | Daniel Dunbar | 2008-10-22 | 1 | -2/+2 |
| * | Privatize PrintModulePass and PrintFunctionPass and add | Daniel Dunbar | 2008-10-21 | 1 | -3/+4 |
| * | Add an SSE2 algorithm for uint64->f64 conversion. | Dale Johannesen | 2008-10-21 | 1 | -0/+16 |
| * | Fix SelectionDAGBuild lowering of Select instructions to | Dan Gohman | 2008-10-21 | 1 | -8/+22 |
| * | Don't create TargetGlobalAddress nodes with offsets that don't fit | Dan Gohman | 2008-10-21 | 1 | -1/+1 |
| * | Optimized FCMP_OEQ and FCMP_UNE for x86. | Dan Gohman | 2008-10-21 | 1 | -15/+19 |
| * | When the coalescer is doing rematerializing, have it remove | Dan Gohman | 2008-10-21 | 1 | -2/+3 |
| * | Make the NaN test come second, heuristically assuming | Dan Gohman | 2008-10-21 | 1 | -12/+12 |
| * | CMake: updated lib/CodeGen/CMakeLists.txt | Oscar Fuentes | 2008-10-21 | 1 | -0/+1 |
| * | Fix gcc.c-torture/compile/920520-1.c by inserting bitconverts | Chris Lattner | 2008-10-21 | 1 | -7/+30 |
| * | Add skeleton for the pre-register allocation live interval splitting pass. | Evan Cheng | 2008-10-20 | 1 | -0/+81 |
| * | Fast-isel no longer an experiment. | Dan Gohman | 2008-10-20 | 1 | -1/+1 |
| * | Add a register class -> virtual registers map. | Evan Cheng | 2008-10-20 | 1 | -0/+1 |
| * | Support operations like fp_to_uint with a vector | Duncan Sands | 2008-10-20 | 2 | -1/+53 |
| * | LegalizeTypes support for atomic operation promotion. | Duncan Sands | 2008-10-20 | 2 | -3/+78 |
| * | Use DAG.getIntPtrConstant rather than DAG.getConstant | Duncan Sands | 2008-10-20 | 2 | -6/+4 |
| * | Always use either MVT::i1 or getSetCCResultType for | Duncan Sands | 2008-10-20 | 1 | -15/+51 |
| * | Formatting - no functional change. | Duncan Sands | 2008-10-20 | 2 | -7/+6 |
| * | Don't use a random type for the select condition, | Duncan Sands | 2008-10-20 | 1 | -2/+1 |
| * | Set N->OperandList to 0 after deletion. Otherwise, it's possible that it will | Bill Wendling | 2008-10-19 | 1 | -15/+26 |
| * | Fix comment. Other formatting changes. No functionality changes. | Bill Wendling | 2008-10-19 | 1 | -5/+6 |
| * | Vector shuffle mask elements may be "undef". Handle | Duncan Sands | 2008-10-19 | 1 | -13/+24 |
| * | Use a legal integer type for vector shuffle mask | Duncan Sands | 2008-10-19 | 1 | -4/+4 |
| * | Reapply r57699 with a fix to not crash on asms with multiple results. Unlike | Chris Lattner | 2008-10-18 | 1 | -16/+48 |
| * | Don't truncate GlobalAddress offsets to int in debug output. | Dan Gohman | 2008-10-18 | 2 | -2/+2 |
| * | By min, I mean max. | Evan Cheng | 2008-10-18 | 1 | -1/+1 |
| * | When creating intervals, leave min(1, numdefs) holes after each instruction. | Evan Cheng | 2008-10-18 | 1 | -5/+13 |
| * | Teach DAGCombine to fold constant offsets into GlobalAddress nodes, | Dan Gohman | 2008-10-18 | 3 | -2/+47 |
| * | Revert r57699. It's causing regressions in | Dan Gohman | 2008-10-18 | 1 | -43/+15 |
| * | Factor out the code for mapping LLVM IR condition opcodes to | Dan Gohman | 2008-10-17 | 2 | -34/+47 |
| * | Fix PR2898. Spiller delete a store for reuse before it knows for sure the reu... | Evan Cheng | 2008-10-17 | 1 | -11/+26 |
| * | add support for 128 bit aggregates. | Chris Lattner | 2008-10-17 | 1 | -0/+1 |
| * | The Dwarf writer was comparing mangled and unmangled names for C++ code when we | Bill Wendling | 2008-10-17 | 1 | -1/+4 |