projects / oota-llvm.git / search
? search:
summary | shortlog | log | commit | commitdiff | tree
first ⋅ prev ⋅ next
AMDGPU: Implement isNoopAddrSpaceCast
2015-12-01 Matt ArsenaultAMDGPU: Implement isNoopAddrSpaceCast
 commit | commitdiff | tree
2015-12-01 Matt ArsenaultAMDGPU: Disallow flat_scr in SI assembler
 commit | commitdiff | tree
2015-12-01 Matt ArsenaultAMDGPU: Optimize VOP2 operand legalization
 commit | commitdiff | tree
2015-12-01 Matt ArsenaultAMDGPU: Report extractelement as free in cost model
 commit | commitdiff | tree
2015-12-01 Matt ArsenaultSquelch unused variable warning in SIRegisterInfo.cpp.
 commit | commitdiff | tree
2015-11-30 Matt ArsenaultAMDGPU: Fix unused function
 commit | commitdiff | tree
2015-11-30 Matt ArsenaultAMDGPU: Error if too many user SGPRs used
 commit | commitdiff | tree
2015-11-30 Matt ArsenaultAMDGPU: Rework how private buffer passed for HSA
 commit | commitdiff | tree
2015-11-30 Matt ArsenaultAMDGPU: Rename enums to be consistent with HSA code...
commit | commitdiff | tree
2015-11-30 Matt ArsenaultAMDGPU: Remove SIPrepareScratchRegs
 commit | commitdiff | tree
2015-11-30 Matt ArsenaultAMDGPU: Use assert zext for workgroup sizes
 commit | commitdiff | tree
2015-11-30 Matt ArsenaultAMDGPU: Don't reserve SCRATCH_PTR input register
 commit | commitdiff | tree
2015-11-25 Matt ArsenaultAMDGPU: Check feature attributes in SIMachineFunctionInfo
 commit | commitdiff | tree
2015-11-25 Matt ArsenaultAMDGPU: Add some tests for promotion of v2i64 scalar_to_vector
 commit | commitdiff | tree
2015-11-25 Matt ArsenaultAMDGPU: Make v2i64/v2f64 legal types.
 commit | commitdiff | tree
2015-11-24 Matt ArsenaultAMDGPU: Split LDS vector loads
 commit | commitdiff | tree
2015-11-24 Matt ArsenaultAMDGPU: Split x8 and x16 vector loads instead of scalarize
 commit | commitdiff | tree
2015-11-11 Matt ArsenaultAdd target preference for GatherAllAliases max depth
 commit | commitdiff | tree
2015-11-11 Matt ArsenaultAMDGPU: Print more fields in comments
 commit | commitdiff | tree
2015-11-11 Matt ArsenaultAMDGPU: Remove dead code
 commit | commitdiff | tree
2015-11-11 Matt ArsenaultAMDGPU: Set isAllocatable = 0 on VS_32/VS_64
 commit | commitdiff | tree
2015-11-10 Matt ArsenaultLegalizeDAG: Implement promote for scalar_to_vector
 commit | commitdiff | tree
2015-11-10 Matt ArsenaultLegalizeDAG: Implement promote for insert_vector_elt
 commit | commitdiff | tree
2015-11-10 Matt ArsenaultLegalizeDAG: Implement promote for extract_vector_elt
 commit | commitdiff | tree
2015-11-10 Matt ArsenaultRemove unnecessary call to getAllocatableRegClass
 commit | commitdiff | tree
2015-11-06 Matt ArsenaultAMDGPU: Cleanup includes
 commit | commitdiff | tree
2015-11-06 Matt ArsenaultAMDGPU: Create emergency stack slots during frame lowering
 commit | commitdiff | tree
2015-11-06 Matt ArsenaultAMDGPU: Remove unused scratch resource operands
 commit | commitdiff | tree
2015-11-06 Matt ArsenaultAMDGPU: Add pass to detect used kernel features
 commit | commitdiff | tree
2015-11-06 Matt ArsenaultAMDGPU: Fix hardcoded alignment of spill.
 commit | commitdiff | tree
2015-11-06 Matt ArsenaultAMDGPU: Hack for VS_32 register pressure
 commit | commitdiff | tree
2015-11-05 Matt ArsenaultAMDGPU: Also track whether SGPRs were spilled
 commit | commitdiff | tree
2015-11-05 Matt ArsenaultAMDGPU: Print number user SGPRs
 commit | commitdiff | tree
2015-11-05 Matt ArsenaultAMDGPU: Disallow s[102:103] on VI in assembler
 commit | commitdiff | tree
2015-11-05 Matt ArsenaultAMDGPU: Fix assert when legalizing atomic operands
 commit | commitdiff | tree
2015-11-05 Matt ArsenaultAMDGPU: Make addr64 atomic operand order consistent
 commit | commitdiff | tree
2015-11-05 Matt ArsenaultAMDGPU: Add missing v2f64 fadd tests
 commit | commitdiff | tree
2015-11-05 Matt ArsenaultAMDGPU: Fix typo
 commit | commitdiff | tree
2015-11-03 Matt ArsenaultAMDGPU: Make flat_scratch name consistent
 commit | commitdiff | tree
2015-11-03 Matt ArsenaultAMDGPU: Fix asserts on invalid register ranges
 commit | commitdiff | tree
2015-11-03 Matt ArsenaultAMDGPU: Fix off by one error in register parsing
 commit | commitdiff | tree
2015-11-03 Matt ArsenaultAMDGPU: s[102:103] is unavailable on VI
 commit | commitdiff | tree
2015-11-03 Matt ArsenaultAMDGPU: Define correct number of SGPRs
 commit | commitdiff | tree
2015-11-03 Matt ArsenaultAMDGPU: Make findUsedSGPR more readable
 commit | commitdiff | tree
2015-11-03 Matt ArsenaultAMDGPU: Initialize SIFixSGPRCopies so -print-after...
commit | commitdiff | tree
2015-11-03 Matt ArsenaultAMDGPU: Alphabetize includes
 commit | commitdiff | tree
2015-11-02 Matt ArsenaultAMDGPU: Stop assuming vreg for build_vector
 commit | commitdiff | tree
2015-11-02 Matt ArsenaultAMDGPU: Error on graphics shaders with HSA
 commit | commitdiff | tree
2015-11-02 Matt ArsenaultAMDGPU: Un XFAIL a test
 commit | commitdiff | tree
2015-11-02 Matt ArsenaultAMDGPU: Distribute SGPR->VGPR copies of REG_SEQUENCE
 commit | commitdiff | tree
2015-10-24 Matt ArsenaultAMDGPU: Print modifiers when dumping AMDGPUOperand
 commit | commitdiff | tree
2015-10-23 Matt ArsenaultAMDGPU: Fix parsing of 32-bit literals with sign bit set
 commit | commitdiff | tree
2015-10-21 Matt ArsenaultAMDGPU: Fix adding redundant m0 uses
 commit | commitdiff | tree
2015-10-21 Matt ArsenaultAMDGPU: Fix verifier error in SIFoldOperands
 commit | commitdiff | tree
2015-10-21 Matt ArsenaultAMDGPU: Split DiagnosticInfoUnsupported into its own...
commit | commitdiff | tree
2015-10-21 Matt ArsenaultAMDGPU: Simplify VOP3 operand legalization.
 commit | commitdiff | tree
2015-10-21 Matt ArsenaultAMDGPU: Fix not checking implicit operands in verifyInstruction
 commit | commitdiff | tree
2015-10-21 Matt ArsenaultUse numeric_limits instead of LLONG_MAX
 commit | commitdiff | tree
2015-10-21 Matt ArsenaultLegalizeDAG: Implement promote for build_vector
 commit | commitdiff | tree
2015-10-20 Matt ArsenaultAMDGPU: Add MachineInstr overloads for instruction...
commit | commitdiff | tree
2015-10-20 Matt ArsenaultAMDGPU: Stop reserving v[254:255]
 commit | commitdiff | tree
2015-10-13 Matt ArsenaultDAGCombiner: Don't stop finding better chain on 2 aliases
 commit | commitdiff | tree
2015-10-13 Matt ArsenaultAMDGPU: Refactor isVGPRToSGPRCopy
 commit | commitdiff | tree
2015-10-12 Matt ArsenaultDAGCombiner: Combine extract_vector_elt from build_vector
 commit | commitdiff | tree
2015-10-12 Matt ArsenaultAMDGPU: Register some more passes so -print-before...
commit | commitdiff | tree
2015-10-12 Matt ArsenaultEnable verifier after PeepholeOptimizer
 commit | commitdiff | tree
2015-10-07 Matt ArsenaultAMDGPU: Fix missing implicit m0 uses on movrel instructions
 commit | commitdiff | tree
2015-10-07 Matt ArsenaultAMDGPU: Add comment for VOP2b operand class
 commit | commitdiff | tree
2015-10-07 Matt ArsenaultAMDGPU: Properly register passes
 commit | commitdiff | tree
2015-10-07 Matt ArsenaultAMDGPU: Use explicit register size indirect pseudos
 commit | commitdiff | tree
2015-10-07 Matt ArsenaultAMDGPU: Remove inferRegClassFromUses / inferRegClassFromDefs
 commit | commitdiff | tree
2015-10-02 Matt ArsenaultAMDGPU/SI: Add verifier check for exec reads
 commit | commitdiff | tree
2015-10-02 Matt ArsenaultAdd way to test for generic TargetOpcodes
 commit | commitdiff | tree
2015-10-01 Matt ArsenaultAMDGPU: Fix unused variable warning in release build
 commit | commitdiff | tree
2015-10-01 Matt ArsenaultAMDGPU: Move SIFixSGPRLiveRanges to be a regalloc pass
 commit | commitdiff | tree
2015-10-01 Matt ArsenaultAMDGPU: Merge if and switch
 commit | commitdiff | tree
2015-10-01 Matt ArsenaultAMDGPU: Remove dead code
 commit | commitdiff | tree
2015-10-01 Matt ArsenaultAMDGPU: Make SIInsertWaits about a factor of 4 faster
 commit | commitdiff | tree
2015-09-28 Matt ArsenaultAMDGPU: Factor switch into separate function
 commit | commitdiff | tree
2015-09-28 Matt ArsenaultAMDGPU: Fix splitting x16 SMRD loads
 commit | commitdiff | tree
2015-09-28 Matt ArsenaultAMDGPU: Fix moving SMRD loads with literal offsets...
commit | commitdiff | tree
2015-09-28 Matt ArsenaultAMDGPU: Fix splitting SMRD with large offset
 commit | commitdiff | tree
2015-09-28 Matt ArsenaultAMDGPU: Add testcases
 commit | commitdiff | tree
2015-09-28 Matt ArsenaultAMDGPU: Cleanup test
 commit | commitdiff | tree
2015-09-26 Matt ArsenaultAMDGPU: Remove hasPostISelHook from most instructions
 commit | commitdiff | tree
2015-09-26 Matt ArsenaultAMDGPU: Switch over reg class size instead of checking...
commit | commitdiff | tree
2015-09-26 Matt ArsenaultAMDGPU: Don't handle invalid reg classes in helper...
commit | commitdiff | tree
2015-09-26 Matt ArsenaultAMDGPU: Set CopyCost of register classes
 commit | commitdiff | tree
2015-09-26 Matt ArsenaultAMDGPU: VOP3b definition cleanups
 commit | commitdiff | tree
2015-09-26 Matt ArsenaultAMDGPU: Fix sched model for VOP2b instructions
 commit | commitdiff | tree
2015-09-25 Matt ArsenaultAMDGPU: Construct new buffer instruction when moving...
commit | commitdiff | tree
2015-09-25 Matt ArsenaultDAGCombiner: Check if store is volatile first
 commit | commitdiff | tree
2015-09-25 Matt ArsenaultPeepholeOptimizer: Remove redundant copies
 commit | commitdiff | tree
2015-09-25 Matt ArsenaultAMDGPU: Add some more tests for literal operands
 commit | commitdiff | tree
2015-09-25 Matt ArsenaultAMDGPU: Make getNamedOperandIdx declaration readonly
 commit | commitdiff | tree
2015-09-25 Matt ArsenaultAMDGPU: Disable some passes that are not meaningful
 commit | commitdiff | tree
2015-09-25 Matt ArsenaultAMDGPU: Handle i64->v2i32 loads/stores in PreprocessISelDAG
 commit | commitdiff | tree
2015-09-25 Matt ArsenaultAMDGPU: Fix recomputing dominator tree unnecessarily
 commit | commitdiff | tree
2015-09-25 Matt ArsenaultAMDGPU: Re-justify workaround and fix worked around...
commit | commitdiff | tree
2015-09-25 Matt ArsenaultAMDGPU: Don't create REG_SEQUENCE with SGPR dest and...
commit | commitdiff | tree
next
C/C++ LLVM-based compilers that forbids OOTA behaviors