2006-02-05 |
Evan Cheng | * Added SDNode::isOnlyUse().
|
commit | commitdiff | tree |
2006-02-05 |
Evan Cheng | Re-commit the last bit of change that was backed out.
|
commit | commitdiff | tree |
2006-02-05 |
Evan Cheng | Re-committing the last bit of change. It shouldn't...
|
commit | commitdiff | tree |
2006-02-04 |
Evan Cheng | Complex pattern's custom matcher should not call Select...
|
commit | commitdiff | tree |
2006-02-04 |
Evan Cheng | Get rid of some memory leaks identified by Valgrind
|
commit | commitdiff | tree |
2006-02-04 |
Evan Cheng | Remove an unnecessary predicate.
|
commit | commitdiff | tree |
2006-02-04 |
Evan Cheng | Separate FILD and FILD_FLAG, the later is only used...
|
commit | commitdiff | tree |
2006-02-03 |
Evan Cheng | Added a (store (op (load ...) ...) ...) folding test...
|
commit | commitdiff | tree |
2006-02-03 |
Evan Cheng | (store (op (load ...))) folding problem. In the generated...
|
commit | commitdiff | tree |
2006-02-03 |
Evan Cheng | Added case HANDLENODE to getOperationName().
|
commit | commitdiff | tree |
2006-02-02 |
Evan Cheng | Update.
|
commit | commitdiff | tree |
2006-02-02 |
Evan Cheng | Fix a erroneous comment.
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | Tell codegen MOVAPSrr and MOVAPDrr are copies.
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | Added SSE entries to foldMemoryOperand().
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | Rearrange code to my liking. :)
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | - Use xor to clear integer registers (set R, 0).
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | Remove another entry.
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | If a pattern's root node is a constant, its size should...
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | Return's chain should be matching either the chain...
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | When folding a load into a return of SSE value, check...
|
commit | commitdiff | tree |
2006-02-01 |
Evan Cheng | Remove an item. It's done.
|
commit | commitdiff | tree |
2006-01-31 |
Evan Cheng | Be smarter about whether to store the SSE return value...
|
commit | commitdiff | tree |
2006-01-31 |
Evan Cheng | - Allow XMM load (for scalar use) to be folded into...
|
commit | commitdiff | tree |
2006-01-31 |
Evan Cheng | Remove entries on fabs and fneg. These are done.
|
commit | commitdiff | tree |
2006-01-31 |
Evan Cheng | Allow the specification of explicit alignments for...
|
commit | commitdiff | tree |
2006-01-31 |
Evan Cheng | Allow custom lowering of fabs. I forgot to check in...
|
commit | commitdiff | tree |
2006-01-31 |
Evan Cheng | Added custom lowering of fabs
|
commit | commitdiff | tree |
2006-01-30 |
Evan Cheng | Don't generate complex sequence for SETOLE, SETOLT...
|
commit | commitdiff | tree |
2006-01-30 |
Evan Cheng | Don't generate (or setp, setae) for SETUGE. Simply...
|
commit | commitdiff | tree |
2006-01-30 |
Evan Cheng | i64 -> f32, f32 -> i64 and some clean up.
|
commit | commitdiff | tree |
2006-01-30 |
Evan Cheng | Always use FP stack instructions to perform i64 to...
|
commit | commitdiff | tree |
2006-01-30 |
Evan Cheng | One more getTargetNode() variant shouldn't hurt...
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | A bit of wisdom from Chris on the last entry.
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | AT&T assembly convention: registers are in lower case.
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | Use Intel assembly syntax to look for ST
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | (store (op (load ...)), ...)
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | Added notes about a x86 isel deficiency.
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | Added a temporary option -enable-x86-sse to enable...
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | Bye bye Pattern ISel, hello DAG ISel.
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | X86 dag isel is now (soon) the default.
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | These are fp stack test cases.
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | XFAIL tailcall test cases until it's implemented.
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | A better workaround
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | x86 CPU detection and proper subtarget support
|
commit | commitdiff | tree |
2006-01-27 |
Evan Cheng | Subtarget feature can now set any variable to any value
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | Another folding problem: if a node r/w chain or flag...
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | Added preliminary x86 subtarget support.
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | Work around some x86 Darwin assembler bugs
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | Add a FIXME comment.
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | When trying to fold X86::SETCC into a Select, make...
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | Clean up some code; improve efficiency; and fixed a...
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | Remove the uses of STATUS flag register. Rely on node...
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | Duh.
|
commit | commitdiff | tree |
2006-01-26 |
Evan Cheng | Incoming (and optional) flag bugs. They may be embedded...
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | No need to keep track of top and bottom nodes in a...
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | Set SchedulingForLatency to be the default scheduling...
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | Default scheduling preference is SchedulingForLatency.
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | X86 prefer scheduling for reduced register pressure.
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | Bottom up register usage reducing list scheduler.
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | Keep track of bottom / top element of a set of flagged...
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | If scheduler choice is the default (-sched=default...
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | Add a enum to specify target scheduling preference...
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | Some minor scheduler changes.
|
commit | commitdiff | tree |
2006-01-25 |
Evan Cheng | Fix a selectcc lowering bug. Make a copy of X86ISD...
|
commit | commitdiff | tree |
2006-01-24 |
Evan Cheng | Fix an optional in flag bug.
|
commit | commitdiff | tree |
2006-01-24 |
Evan Cheng | Optional InFlag was not being included in node.
|
commit | commitdiff | tree |
2006-01-23 |
Evan Cheng | Skeleton of the list schedule.
|
commit | commitdiff | tree |
2006-01-23 |
Evan Cheng | Minor clean up.
|
commit | commitdiff | tree |
2006-01-23 |
Evan Cheng | Remove a couple of unnecessary #include's
|
commit | commitdiff | tree |
2006-01-23 |
Evan Cheng | Factor out more instruction scheduler code to the base...
|
commit | commitdiff | tree |
2006-01-21 |
Evan Cheng | Rename fcmovae to fcmovnb and fcmova to fcmovnbe (following...
|
commit | commitdiff | tree |
2006-01-21 |
Evan Cheng | Do some code refactoring on Jim's scheduler in preparation...
|
commit | commitdiff | tree |
2006-01-20 |
Evan Cheng | Stop doing that accidental commit.
|
commit | commitdiff | tree |
2006-01-20 |
Evan Cheng | A few more SH{L|R}D peepholes.
|
commit | commitdiff | tree |
2006-01-20 |
Evan Cheng | Added a load folding bug test case.
|
commit | commitdiff | tree |
2006-01-20 |
Evan Cheng | Prevent folding of a node with multiple uses if the...
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Didn't mean to commit the last one.
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Added i16 SH{L|R}D patterns.
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Bug fix. Flag operand number may be calculated incorrectly.
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Use pattern information to determine whether the use...
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Avoid generating a redundant setcc.
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | adc and sbb need an incoming flag to ensure it reads...
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Another typo
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Two peepholes:
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Prevent unnecessary CopyToReg when the same HW register...
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | Didn't mean to check that in.
|
commit | commitdiff | tree |
2006-01-19 |
Evan Cheng | A obvious typo
|
commit | commitdiff | tree |
2006-01-18 |
Evan Cheng | SRA shift amount must be in i8
|
commit | commitdiff | tree |
2006-01-18 |
Evan Cheng | If a call return type is i1, insert a truncate from...
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | Fix lowering of calls which return f32 values.
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | Bug fix: missing LegalizeOp() on newly created nodes.
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | Suppress "no newline at end of file" warnings.
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | Emit a type matching check for ComplexPatterns.
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | Zero extending load from i1 to i8.
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | SSE does not support i64 SINT_TO_FP (FP stack doesn...
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | Added a FIXME comment about why FST is currently flagged...
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | Yet another getTargetNode() variant. I promise one...
|
commit | commitdiff | tree |
2006-01-17 |
Evan Cheng | Bug fixes: fpGETRESULT should produces a flag result...
|
commit | commitdiff | tree |
2006-01-16 |
Evan Cheng | More typo's
|
commit | commitdiff | tree |
2006-01-16 |
Evan Cheng | Some typo's
|
commit | commitdiff | tree |
next |