summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2013-05-08R600: Improve texture handlingnativeVincent Lejeune9-198/+510
2013-05-06R600: Anti dep better handled in tex clauseVincent Lejeune1-6/+4
2013-05-06factorize swizzleVincent Lejeune3-136/+140
2013-05-05R600: Some factorizationVincent Lejeune5-85/+91
2013-05-04R600:packetize data depVincent Lejeune2-0/+250
2013-05-04various fix for pvVincent Lejeune1-4/+3
2013-05-04use a more pv friendly schedulingVincent Lejeune1-1/+1
2013-05-04R600: Don't emit empty then clause and use alu_pop_afterVincent Lejeune4-13/+50
2013-05-04R600: transonly in a separate ig ?Vincent Lejeune1-1/+1
2013-05-04R600: Recompute schedule graph non order dependencies.Vincent Lejeune9-29/+572
2013-05-04Add a case to LiveIntervalAnalysis::HandleMoveVincent Lejeune1-4/+123
Some target like R600 may want to swap LiveRange order for a same register. Subregister writes are independant and can be reorder to improve scheduling. This patch allows LiveIntervals to support moving intervals accross some others.
2013-05-04R600: Use bottom up scheduling algorithmVincent Lejeune5-115/+106
2013-05-04R600: Factorize Fetch size limit inside AMDGPUSubTargetVincent Lejeune4-13/+13
2013-05-04R600: prettier dump of clampVincent Lejeune2-4/+4
2013-05-02R600: Signed literals are 64bits wideVincent Lejeune2-3/+168
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180960 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02R600: If previous bundle is dot4, PV valid chan is always XVincent Lejeune2-38/+295
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180959 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02R600: Add a test to check that use_kill is emittedVincent Lejeune1-0/+1
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180958 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02R600: Improve asmPrint of ALU clauseVincent Lejeune3-4/+20
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180957 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02R600: Prettier asmPrint of AluVincent Lejeune38-114/+172
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180956 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Hexagon: Add multiclass/encoding bits for the New-Value Jump instructions.Jyotsna Verma5-404/+305
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180953 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02[GV] Remove dead code which is really difficult to decipher.Shuxin Yang1-26/+2
Actually it took me couple of hours trying to make sense of them and only to find they are dead code. I guess the original author used "allSingleSucc" to indicate if there are any critial edge emanating from some blocks, and tried to perform code motion (actually speculation) in the presence of these critical edges; but later on he/she changed mind and decided to perform edge-splitting first. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180951 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02We don't want FP elimination when doing an Apple-style build.Bill Wendling2-2/+2
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180949 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Replaced usages of size_type with size_t to be more consistent.Aaron Ballman1-8/+8
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180947 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Hexagon - Add peephole optimizations for zero extends.Pranav Bhandarkar4-6/+64
* lib/Target/Hexagon/HexagonInstrInfo.td: Add patterns to combine a sequence of a pair of i32->i64 extensions followed by a "bitwise or" into COMBINE_rr. * lib/Target/Hexagon/HexagonPeephole.cpp: Copy propagate Rx in the instruction Rp = COMBINE_Ir_V4(0, Rx) to the uses of Rp:subreg_loreg. * test/CodeGen/Hexagon/union-1.ll: New test. * test/CodeGen/Hexagon/combine_ir.ll: Fix test. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180946 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02[mips] Fix the head Mips16RegisterInfo.cpp commentRichard Sandiford1-2/+1
...aka a test commit. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180936 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02TBAA: remove !tbaa from testing cases if not used.Manman Ren9-182/+138
This will make it easier to turn on struct-path aware TBAA since the metadata format will change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180935 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Provide an API to temporarily suppress DebugLocations from being attachedAdrian Prantl1-0/+21
to emitted instructions. Use this if you want an instruction to be counted towards the prologue or if there is no useful source location. rdar://problem/13442648 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180929 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Hexagon: Honor __builtin_expect by using branch probabilities.Jyotsna Verma4-37/+122
* lib/Target/Hexagon/HexagonInstrInfo.cpp (GetDotNewPredOp): Given a jump opcode return the right pred.new jump opcode with a taken vs not-taken hint based on branch probabilities provided by the target independent module. * lib/Target/Hexagon/HexagonVLIWPacketizer.cpp: Use the above function. * lib/Target/Hexagon/HexagonNewValueJump.cpp(getNewvalueJumpOpcode): Enhance existing function use branch probabilities like HexagonInstrInfo::GetDotNewPredOp but for New Value (GPR) Jumps. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180923 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02R600: Use new tablegen syntax for patternsTom Stellard3-447/+381
All but two patterns have been converted to the new syntax. The remaining two patterns will require COPY_TO_REGCLASS instructions, which the VLIW DAG Scheduler cannot handle. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180922 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02R600/SI: remove nonsense select patternTom Stellard1-8/+1
Fortunately this pattern never matched, otherwise we would have generated incorrect code. Signed-off-by: Christian K??nig <christian.koenig@amd.com> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180921 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Added table of contents declaration in CommandLine Library documentation.Tobias Grosser1-0/+3
Contributed-by: Dan Liew <daniel.liew@imperial.ac.uk> git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180919 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-0280-col fixup.Michael Liao1-2/+3
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180915 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Avoid duplicating logic on frame register selecting when lowering eh_returnMichael Liao1-9/+11
No functionality change git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180914 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Avoid duplicating logic on frame register selecting when lowering frameaddrMichael Liao1-1/+3
No functionality change git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180912 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Rewrite X86 codegen regression test with FileCheckMichael Liao1-4/+7
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180910 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02Add a test for the foldSelectICmpAndOr fix committed in r180779.David Majnemer1-0/+13
This tests a case where C1 and C2 were the same but X and Y were different widths. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180907 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-02TiedTo flag can now be placed on implicit operands. isTwoAddrUse() should lookEvan Cheng1-5/+3
at all of the operands. Previously it was skipping over implicit operands which cause infinite looping when the two-address pass try to reschedule a two-address instruction below the kill of tied operand. I'm unable to come up with a reasonably sized test case. rdar://13747577 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180906 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01[mips] Rename class and functions. Simplify code.Akira Hatanaka1-26/+29
No functionality changes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180897 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01This exposes more MCJIT options via the C API:Filip Pizlo6-44/+85
CodeModel: It's now possible to create an MCJIT instance with any CodeModel you like. Previously it was only possible to create an MCJIT that used CodeModel::JITDefault. EnableFastISel: It's now possible to turn on the fast instruction selector. The CodeModel option required some trickery. The problem is that previously, we were ensuring future binary compatibility in the MCJITCompilerOptions by mandating that the user bzero's the options struct and passes the sizeof() that he saw; the bindings then bzero the remaining bits. This works great but assumes that the bitwise zero equivalent of any field is a sensible default value. But this is not the case for LLVMCodeModel, or its internal equivalent, llvm::CodeModel::Model. In both of those, the default for a JIT is CodeModel::JITDefault (or LLVMCodeModelJITDefault), which is not bitwise zero. Hence this change introduces LLVMInitializeMCJITCompilerOptions(), which will initialize the user's options struct with defaults. The user will use this in the same way that they would have previously used memset() or bzero(). MCJITCAPITest.cpp illustrates the change, as does the comment in ExecutionEngine.h. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180893 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Avoid generating tempfile(s) never usedMichael Liao1-2/+2
As DejaGNU is deprecated, it seems pipe-jam issue doesn't exist any more. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180892 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Rename 'struct LLVMTargetMachine' to 'struct LLVMOpaqueTargetMachine'. Filip Pizlo1-1/+1
This avoids namespace collisions with llvm::LLVMTargetMachine. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180891 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Revert r180737. The companion patch was reverted, and this is not relevant ↵Bill Wendling3-105/+0
right now. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180889 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Hexagon: Use multiclass for Jump instructions.Jyotsna Verma14-263/+336
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180885 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Hexagon: Clear isKill flag on the predicate register inJyotsna Verma1-1/+5
PredicateInstruction function. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180884 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Add support for other typeinfo encodings in the ExceptionDemo.Rafael Espindola1-9/+41
The old jit always uses DW_EH_PE_absptr, but MCJIT can use other encodings. This is in preparation for adding EH support to MCJIT, but not directly related, so I am committing it first. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180883 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Fix file header comment.Filip Pizlo1-1/+1
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180882 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01This patch breaks up Wrap.h so that it does not have to include all of Filip Pizlo32-145/+156
the things, and renames it to CBindingWrapping.h. I also moved CBindingWrapping.h into Support/. This new file just contains the macros for defining different wrap/unwrap methods. The calls to those macros, as well as any custom wrap/unwrap definitions (like for array of Values for example), are put into corresponding C++ headers. Doing this required some #include surgery, since some .cpp files relied on the fact that including Wrap.h implicitly caused the inclusion of a bunch of other things. This also now means that the C++ headers will include their corresponding C API headers; for example Value.h must include llvm-c/Core.h. I think this is harmless, since the C API headers contain just external function declarations and some C types, so I don't believe there should be any nasty dependency issues here. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180881 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01SROA: Generate selects instead of shuffles when blending values because this ↵Nadav Rotem2-22/+20
is the cannonical form. Shuffles are more difficult to lower and we usually don't touch them, while we do optimize selects more often. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180875 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01[inline asm] Return an undef SDValue of the expected value type, rather thanChad Rosier1-1/+1
report a fatal error. This allows us to continue processing the translation unit. Test case to come on the clang side because we need an inline asm diagnostics handler in place. rdar://13446483 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180873 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-01Optimize away nop CONCAT_VECTOR nodes.Nadav Rotem2-0/+52
Optimize CONCAT_VECTOR nodes that merge EXTRACT_SUBVECTOR values that extract from the same vector. rdar://13402653 PR15866 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180871 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-09 21:20:00 +0000