Commits

Show all
Author Commit Message Labels Comments Date
Dan Gohman
Initial documentation for first-class aggregates changes.
Tags
2 tags
Dan Gohman
Fix a malformed %gt;
Nate Begeman
Teach Legalize how to scalarize VSETCC Teach X86 a few more vsetcc patterns. Custom lowering for unsupported ones is next.
Evan Cheng
Xform bitconvert(build_pair(load a, load b)) to a single load if the load locations are at the right offset from each other.
Dale Johannesen
New test for tail merging
Dale Johannesen
Be more aggressive about tail-merging small blocks if those blocks consist entirely of common instructions; merging will not add an extra branch in this case.
Bill Wendling
Constify isSourceDefinedByImplicitDef function. Otherwise, just formatting changes that don't change functionality.
Bill Wendling
Constify the machine instruction passed into the "is{Trivially,Really}ReMaterializable" methods.
Nate Begeman
Initial X86 codegen support for VSETCC.
Dale Johannesen
Further rework of tail merge algorithm. Not quite semantically identical, but little difference in either results or execution speed; but it's much easier to read, at least IMO.
Nate Begeman
Simplify some checks
Dan Gohman
Fix a copy+paste bug; pseudo-instructions shouldn't have encoding information.
Nate Begeman
Pointer comparisons should use icmp, not vicmp
Owen Anderson
Go back to passing the analyses around as parameters.
Nate Begeman
Pointer comparisons should be handled by icmp, not vicmp :)
Evan Cheng
Forgot this.
Nate Begeman
Hard code CmpInst back to i1 for now while I go track down what in the bitcode reader/writer is assuming i1 This was breaking a bunch of tests
Evan Cheng
Refactor isConsecutiveLoad from X86 to TargetLowering so DAG combiner can make use of it.
Bill Wendling
Revert the previous commit. Go ahead and hoist rematerializable instructions.
Nate Begeman
Add support for vicmp/vfcmp codegen, more legalize support coming. This is necessary to unbreak the build.
Bill Wendling
One real change - don't hoist something that's trivially rematerializable. It's possible for it to produce worse code than before. The rest of this patch is code cleanup.
Nate Begeman
Fix build breakage
Nate Begeman
Add two new instructions to the llvm IR, vicmp and vfcmp. see updated LangRef for details. CodeGen support coming in a follow up patch
Dan Gohman
Make firstEightPowers const.
Dan Gohman
Update comments.
Mikhail Glushenkov
Filter option names to escape symbols not allowed as C++ identifiers. Makes it possible to use options with names like "Wa,". Also fixes the -Wall option handling as a side-effect.
Mikhail Glushenkov
Make it possible to choose between different compilation graph definitions at compile-time.
Mikhail Glushenkov
Fix some error messages; Make LLVMC pass through the exit code of a failed tool.
Dan Gohman
Fix a compile error on compilers that still want a return value in a non-void function that calls abort.
Dan Gohman
Fix a missing break in the ISD::FLT_ROUNDS_ handling. Patch by giuma!
  1. Prev
  2. Next