| 2006-03-28 |
Evan Cheng | * Prefer using operation of matching types. e.g unpcklpd...
|
commit | commitdiff | tree |
| 2006-03-28 |
Evan Cheng | Use movhpd is even better than movlhps.
|
commit | commitdiff | tree |
| 2006-03-28 |
Evan Cheng | Added a couple of entries about movhps and movlhps.
|
commit | commitdiff | tree |
| 2006-03-28 |
Evan Cheng | All unpack cases are now being handled.
|
commit | commitdiff | tree |
| 2006-03-28 |
Evan Cheng | - Clean up / consoladate various shuffle masks.
|
commit | commitdiff | tree |
| 2006-03-28 |
Evan Cheng | getVectorTyppe(MVT::i64, 2) ==> MVT::v2i64.
|
commit | commitdiff | tree |
| 2006-03-28 |
Evan Cheng | Model unpack lower and interleave as vector_shuffle...
|
commit | commitdiff | tree |
| 2006-03-27 |
Evan Cheng | Intrinsics naming convention change.
|
commit | commitdiff | tree |
| 2006-03-27 |
Evan Cheng | Try again
|
commit | commitdiff | tree |
| 2006-03-27 |
Evan Cheng | Incorrect check for FP all one's
|
commit | commitdiff | tree |
| 2006-03-27 |
Evan Cheng | Use pcmpeq to generate vector of all ones.
|
commit | commitdiff | tree |
| 2006-03-27 |
Evan Cheng | Changed isBuildVectorAllOnesInteger to isBuildVectorAllOnes.
|
commit | commitdiff | tree |
| 2006-03-27 |
Evan Cheng | Change isBuildVectorAllOnesInteger to isBuildVectorAllOnes...
|
commit | commitdiff | tree |
| 2006-03-26 |
Evan Cheng | Remove X86:isZeroVector, use ISD::isBuildVectorAllZeros...
|
commit | commitdiff | tree |
| 2006-03-26 |
Evan Cheng | Remove PPC:isZeroVector, use ISD::isBuildVectorAllZeros...
|
commit | commitdiff | tree |
| 2006-03-26 |
Evan Cheng | Add immAllZerosV helper
|
commit | commitdiff | tree |
| 2006-03-26 |
Evan Cheng | Add ISD::isBuildVectorAllZeros predicate
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Add a BUILD_VECTOR with unpack and interleave testcase.
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Added missing (any_extend (load ...)) patterns.
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Build arbitrary vector with more than 2 distinct scalar...
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | X86 SSE1 cacheability support ops intrinsics
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Added SSE cachebility ops
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Instruction encoding bug
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Added a scalar to vector with zero extension testcase
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | X86 SSE1 SIMD store intrinsics.
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | X86 SSE1 SIMD load intrinsics (movhps, movlps, and...
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | X86 SSE1 conversion operations intrinsics.
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Added 128-bit packed integer subtraction.
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Added CVTTPS2PI.
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | Added CVTSS2SI.
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | X86 SSE1 comparison intrinsics.
|
commit | commitdiff | tree |
| 2006-03-25 |
Evan Cheng | X86 SSE1 arithmetic and logical operation intrinsics.
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Support for scalar to vector with zero extension.
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Added LDMXCSR
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | ldmxcsr is a SSE instruction.
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Added ldmxcsr intrinsic.
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Only to vector shuffle for {x,x,y,y} cases when SCALAR_TO_VE...
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Zero vector testcase
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Handle BUILD_VECTOR with all zero elements.
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | All v2f64 shuffle cases can be handled.
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | More efficient v2f64 shuffle using movlhps, movhlps...
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | A new entry
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Add a v2f64 shuffle case.
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Add a vector shuffle test case
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Handle more shuffle cases with SHUFP* instructions.
|
commit | commitdiff | tree |
| 2006-03-24 |
Evan Cheng | Lower BUILD_VECTOR to VECTOR_SHUFFLE if there are two...
|
commit | commitdiff | tree |
| 2006-03-23 |
Evan Cheng | Typo
|
commit | commitdiff | tree |
| 2006-03-23 |
Evan Cheng | Typo
|
commit | commitdiff | tree |
| 2006-03-23 |
Evan Cheng | Following icc's lead: use movdqa to load / store 128...
|
commit | commitdiff | tree |
| 2006-03-23 |
Evan Cheng | Add v4i32 <-> v4f32 bitconvert patterns.
|
commit | commitdiff | tree |
| 2006-03-23 |
Evan Cheng | Allow result node to be a simple leaf node. This enable...
|
commit | commitdiff | tree |
| 2006-03-23 |
Evan Cheng | Add 128-bit integer vector load and add (for testing).
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | Added a ValueType operand to isShuffleMaskLegal()....
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | Replace vector splat test case.
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | Add a v2f64 splat (using movlhps) test case.
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | SHUFP* are two address code.
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | Some clean up.
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | - Supposely movlhps is faster / better than unpcklpd.
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | - Implement X86ISelLowering::isShuffleMaskLegal()....
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | - VECTOR_SHUFFLE of v4i32 / v4f32 with undef second...
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | Fix PSHUF* and SHUF* jit code emission problems
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | Some splat and shuffle support.
|
commit | commitdiff | tree |
| 2006-03-22 |
Evan Cheng | Add a couple more pseudo instructions.
|
commit | commitdiff | tree |
| 2006-03-21 |
Evan Cheng | Didn't mean to check this in. No MMX support yet.
|
commit | commitdiff | tree |
| 2006-03-21 |
Evan Cheng | - Use movaps to store 128-bit vector integers.
|
commit | commitdiff | tree |
| 2006-03-21 |
Evan Cheng | Don't forget to promote xform function to an explicit...
|
commit | commitdiff | tree |
| 2006-03-21 |
Evan Cheng | Combine 2 entries
|
commit | commitdiff | tree |
| 2006-03-21 |
Evan Cheng | Add a note about x86 register coallescing
|
commit | commitdiff | tree |
| 2006-03-21 |
Evan Cheng | - Remove scalar to vector pseudo ops. They are just...
|
commit | commitdiff | tree |
| 2006-03-21 |
Evan Cheng | x86 ISD::SCALAR_TO_VECTOR support.
|
commit | commitdiff | tree |
| 2006-03-21 |
Evan Cheng | Junk unused vector register classes.
|
commit | commitdiff | tree |
| 2006-03-20 |
Evan Cheng | The node wrapped in PatLeaf<> should be treated as...
|
commit | commitdiff | tree |
| 2006-03-20 |
Evan Cheng | Option -enable-x86-lsr has been removed
|
commit | commitdiff | tree |
| 2006-03-20 |
Evan Cheng | Use tblgen'd VECTOR_SHUFFLE selection code.
|
commit | commitdiff | tree |
| 2006-03-20 |
Evan Cheng | It should be ok for a xform output type to be different...
|
commit | commitdiff | tree |
| 2006-03-20 |
Evan Cheng | Move a few things around.
|
commit | commitdiff | tree |
| 2006-03-20 |
Evan Cheng | Copy matching pattern's output type info to instruction...
|
commit | commitdiff | tree |
| 2006-03-19 |
Evan Cheng | Vector undef's
|
commit | commitdiff | tree |
| 2006-03-19 |
Evan Cheng | getEnumName() missed v8i8, v4i16, and v2i32 types
|
commit | commitdiff | tree |
| 2006-03-19 |
Evan Cheng | Remind us of exit value substitution
|
commit | commitdiff | tree |
| 2006-03-19 |
Evan Cheng | Turning on LSR by default
|
commit | commitdiff | tree |
| 2006-03-19 |
Evan Cheng | Remember which tests are hurt by LSR.
|
commit | commitdiff | tree |
| 2006-03-18 |
Evan Cheng | - Fixed a bogus if condition.
|
commit | commitdiff | tree |
| 2006-03-18 |
Evan Cheng | Use the generic vector register classes VR64 / VR128...
|
commit | commitdiff | tree |
| 2006-03-18 |
Evan Cheng | Sort StrideOrder so we can process the smallest strides...
|
commit | commitdiff | tree |
| 2006-03-17 |
Evan Cheng | Darwin should use _setjmp/_longjmp instead of setjmp...
|
commit | commitdiff | tree |
| 2006-03-17 |
Evan Cheng | Move some pattern fragments to the right files.
|
commit | commitdiff | tree |
| 2006-03-17 |
Evan Cheng | Allow users of iv / stride to be rewritten with expression...
|
commit | commitdiff | tree |
| 2006-03-17 |
Evan Cheng | Add a lsr common loop invariant hoisting test case
|
commit | commitdiff | tree |
| 2006-03-17 |
Evan Cheng | Also fold MOV8r0, MOV16r0, MOV32r0 + store to MOV8mi...
|
commit | commitdiff | tree |
| 2006-03-17 |
Evan Cheng | Add some missing entries to X86RegisterInfo::foldMemoryOpera...
|
commit | commitdiff | tree |
| 2006-03-17 |
Evan Cheng | - Nuke 16-bit SBB instructions. We'll never use them.
|
commit | commitdiff | tree |
| 2006-03-16 |
Evan Cheng | A new entry.
|
commit | commitdiff | tree |
| 2006-03-16 |
Evan Cheng | Bug fix: condition inverted.
|
commit | commitdiff | tree |
| 2006-03-16 |
Evan Cheng | For each loop, keep track of all the IV expressions...
|
commit | commitdiff | tree |
| 2006-03-16 |
Evan Cheng | Added a way for TargetLowering to specify what values...
|
commit | commitdiff | tree |
| 2006-03-15 |
Evan Cheng | Also requires -mattr=-sse3
|
commit | commitdiff | tree |
| 2006-03-13 |
Evan Cheng | PPC LSR pass should use target lowering hooks.
|
commit | commitdiff | tree |
| 2006-03-13 |
Evan Cheng | Added getTargetLowering() to TargetMachine. Refactored...
|
commit | commitdiff | tree |
| 2006-03-13 |
Evan Cheng | Update
|
commit | commitdiff | tree |
| next |
projects / oota-llvm.git / search