| 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 |
| 2006-03-13 |
Evan Cheng | Add LSR hooks.
|
commit | commitdiff | tree |
| 2006-03-13 |
Evan Cheng | Added getTargetLowering() - returns DAG lowering info.
|
commit | commitdiff | tree |
| 2006-03-13 |
Evan Cheng | Added target lowering hooks which LSR consults to make...
|
commit | commitdiff | tree |
| 2006-03-13 |
Evan Cheng | Optimizing to calloc() to malloc() + memset() can be...
|
commit | commitdiff | tree |
| 2006-03-11 |
Evan Cheng | Doh!
|
commit | commitdiff | tree |
| 2006-03-10 |
Evan Cheng | Added a parameter to control whether Constant::getStringValu...
|
commit | commitdiff | tree |
| 2006-03-09 |
Evan Cheng | Add option -enable-x86-lsr to enable x86 loop strength...
|
commit | commitdiff | tree |
| 2006-03-09 |
Evan Cheng | Add a test case for (store (op (load ..) ..) ..) folding.
|
commit | commitdiff | tree |
| 2006-03-09 |
Evan Cheng | Temporary hack to enable more (store (op (load ......
|
commit | commitdiff | tree |
| 2006-03-07 |
Evan Cheng | X86ISD::REP_STOS and X86ISD::REP_MOVS now produces...
|
commit | commitdiff | tree |
| 2006-03-07 |
Evan Cheng | Use rep/stosl; and Count 0x3; rep/stosb for memset...
|
commit | commitdiff | tree |
| 2006-03-07 |
Evan Cheng | Don't generate silly matching code like this:
|
commit | commitdiff | tree |
| 2006-03-07 |
Evan Cheng | - Emit subsections_via_symbols for Darwin.
|
commit | commitdiff | tree |
| 2006-03-07 |
Evan Cheng | Enable Dwarf debugging info.
|
commit | commitdiff | tree |
| 2006-03-06 |
Evan Cheng | Remove some code that doesn't make sense
|
commit | commitdiff | tree |
| 2006-03-06 |
Evan Cheng | Remove SUnit::Priority1: it is re-calculated on demand...
|
commit | commitdiff | tree |
| 2006-03-05 |
Evan Cheng | Back out fold (shl (add x, c1), c2) -> (add (shl x...
|
commit | commitdiff | tree |
| 2006-03-04 |
Evan Cheng | Add an entry
|
commit | commitdiff | tree |
| 2006-03-04 |
Evan Cheng | MEMSET / MEMCPY lowering bugs: we can't issue a single...
|
commit | commitdiff | tree |
| 2006-03-04 |
Evan Cheng | Typo
|
commit | commitdiff | tree |
| 2006-03-03 |
Evan Cheng | Add another test case for instruction scheduling.
|
commit | commitdiff | tree |
| 2006-03-03 |
Evan Cheng | Add more vector NodeTypes: VSDIV, VUDIV, VAND, VOR...
|
commit | commitdiff | tree |
| 2006-03-03 |
Evan Cheng | Number of NodeTypes now exceeds 128.
|
commit | commitdiff | tree |
| 2006-03-03 |
Evan Cheng | SDOperand::isOperand should not be a forwarding. It...
|
commit | commitdiff | tree |
| 2006-03-03 |
Evan Cheng | Added isOperand(N): true if this is an operand of N
|
commit | commitdiff | tree |
| 2006-03-03 |
Evan Cheng | A bit more tweaking
|
commit | commitdiff | tree |
| 2006-03-03 |
Evan Cheng | Move #include "llvm/CodeGen/AsmPrinter.h" to top since...
|
commit | commitdiff | tree |
| 2006-03-02 |
Evan Cheng | Add a regression test for bug 478.
|
commit | commitdiff | tree |
| 2006-03-02 |
Evan Cheng | - Fixed some priority calculation bugs that were causing...
|
commit | commitdiff | tree |
| 2006-03-01 |
Evan Cheng | Don't print llvm constant in assmebly file. Assembler...
|
commit | commitdiff | tree |
| 2006-03-01 |
Evan Cheng | Back out my last check-in. Wrong place to fix it.
|
commit | commitdiff | tree |
| 2006-03-01 |
Evan Cheng | AsmWriter should not print LLVM constant in comment...
|
commit | commitdiff | tree |
| 2006-03-01 |
Evan Cheng | Vector op lowering.
|
commit | commitdiff | tree |
| 2006-03-01 |
Evan Cheng | New vector type v2f32.
|
commit | commitdiff | tree |
| 2006-03-01 |
Evan Cheng | Vector ops lowering.
|
commit | commitdiff | tree |
| next |
projects / oota-llvm.git / search