aboutsummaryrefslogtreecommitdiff
path: root/test/CodeGen/X86
AgeCommit message (Collapse)Author
2012-05-29Add intrinsics, code gen, assembler and disassembler support for the SSE4a ↵Benjamin Kramer
extrq and insertq instructions. This required light surgery on the assembler and disassembler because the instructions use an uncommon encoding. They are the only two instructions in x86 that use register operands and two immediates. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157634 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-27These tests used intrinsics with the wrong prototype. They weren't caught ↵Chris Lattner
because the old verifier just checked that something "was a pointer", but not that the pointee was correct. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157544 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-26SelectionDAGBuilder: When emitting small compare chains for switches order ↵Benjamin Kramer
them by using edge weights. SimplifyCFG tends to form a lot of 2-3 case switches when merging branches. Move the most likely condition to the front so it is checked first and the others can be skipped. This is currently not as effective as it could be because SimplifyCFG destroys profiling metadata when merging branches and switches. Merging branch weight metadata is tricky though. This code touches at most 3 cases so I didn't use a proper sorting algorithm. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157521 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-25test/CodeGen/X86/bigstructret.ll: Suppress one test. It is ↵NAKAMURA Takumi
msvc-incompatible. (compatible to mingw32 and netbsd, though) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157474 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-25test/CodeGen/X86/bigstructret.ll: Relax stack offsets for hosts of ↵NAKAMURA Takumi
stack-align=8, eg. win32 and netbsd. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157471 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-25Simplify code for calling a function where CanLowerReturn fails, fixing a ↵Eli Friedman
small bug in the process. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157446 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-24Fix for CHECK-NOT misspelling.David Blaikie
Patch by Nicklas Bo Jensen. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157421 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-23Correctly deal with identity copies in RegisterCoalescer.Jakob Stoklund Olesen
Now that the coalescer keeps live intervals and machine code in sync at all times, it needs to deal with identity copies differently. When merging two virtual registers, all identity copies are removed right away. This means that other identity copies must come from somewhere else, and they are going to have a value number. Deal with such copies by merging the value numbers before erasing the copy instruction. Otherwise, we leave dangling value numbers in the live interval. This fixes PR12927. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157340 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-22revert my previous patches that introduced an additional parameter to the ↵Nuno Lopes
objectsize intrinsic. After a lot of discussion, we realized it's not the best option for run-time bounds checking git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157255 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-22Only erase virtregs with no uses left.Jakob Stoklund Olesen
Also make sure registers aren't erased twice if the dead def mentions the register twice. This fixes PR12911. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157254 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-21Allow 256-bit shuffles to still be split even if only half of the shuffle ↵Craig Topper
comes from two 128-bit pieces. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157175 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-20When legalising shifts, do not pre-build a list of operands whichPeter Collingbourne
may be RAUW'd by the recursive call to LegalizeOps; instead, retrieve the other operands when calling UpdateNodeOperands. Fixes PR12889. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157162 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-20Properly constrain register classes for sub-registers.Jakob Stoklund Olesen
Not all GR64 registers have sub_8bit sub-registers. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157150 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-20Properly constrain register classes in 2-addr.Jakob Stoklund Olesen
X86 has 2-addr instructions with different constraints on the tied def and use operands. One is GR32, one is GR32_NOSP. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157149 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-19Fix 12892.Jakob Stoklund Olesen
Dead code elimination during coalescing could cause a virtual register to be split into connected components. The following rewriting would be confused about the already joined copies present in the code, but without a corresponding value number in the live range. Erase all joined copies instantly when joining intervals such that the MI and LiveInterval representations are always in sync. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157135 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-19Erase joined copies immediately.Jakob Stoklund Olesen
The late dead code elimination is no longer necessary. The test changes are cause by a register hint that can be either %rdi or %rax. The choice depends on the use list order, which this patch changes. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157131 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-19On Haswell, perfer storing YMM registers using a single instruction.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157129 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-19Add support for additional in-reg vbroadcast patternsNadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157127 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-18Simplify handling of v16i8 shuffles and fix a missed optimization.Craig Topper
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157043 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-18Teach two-address pass to update the "source" map so it doesn't perform aEvan Cheng
non-profitable commute using outdated info. The test case would still fail because of poor pre-RA schedule. That will be fixed by MI scheduler. rdar://11472010 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157038 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-18Remove a test that was only testing for physreg joining.Jakob Stoklund Olesen
This is the same as the other tests: Clever tricks are required to make the arguments and return value line up in a single-instruction function. It rarely happens in real life. We have plenty other examples of this behavior. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157030 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-17Remove -join-physregs from the test suite.Jakob Stoklund Olesen
This option has been disabled for a while, and it is going away so I can clean up the coalescer code. The tests that required physreg joining to be enabled were almost all of the form "tiny function with interference between arguments and return value". Such functions are usually inlined in the real world. The problem exposed by phys_subreg_coalesce-3.ll is real, but fairly rare. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@157027 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-16Avoid creating a cycle when folding load / op with flag / store. PR11451474. ↵Evan Cheng
rdar://11451474 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156896 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-14Fix PR12821.Jakob Stoklund Olesen
RAFast must add an <imp-def> operand when it is rewriting a sub-register def that isn't a read-modify-write. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156777 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-14Rename @llvm.debugger to @llvm.debugtrap.Dan Gohman
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156774 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-11Fix test/CodeGen/X86/tls-pie.ll.Hans Wennborg
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156612 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-11Implement initial-exec TLS model for 32-bit PIC x86Hans Wennborg
This fixes a TODO from 2007 :) Previously, LLVM would emit the wrong code here (see the update to test/CodeGen/X86/tls-pie.ll). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156611 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-11Define a new intrinsic, @llvm.debugger. It will be similar to __builtin_trap(),Dan Gohman
but it generates int3 on x86 instead of ud2. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156593 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-10AVX2: Add an additional broadcast idiom.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156540 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-10Generate AVX/AVX2 shuffles even when there is a memory op somewhere else in ↵Nadav Rotem
the program. Starting r155461 we are able to select patterns for vbroadcast even when the load op is used by other users. Fix PR11900. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156539 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-09change the objectsize intrinsic signature: add a 3rd parameter to denote the ↵Nuno Lopes
maximum runtime performance penalty that the user is willing to accept. This commit only adds the parameter. Code taking advantage of it will follow. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156473 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-08Remove 256-bit AVX non-temporal store intrinsics. Similar was previously ↵Craig Topper
done for 128-bit. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156375 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-07Fix a regression from r147481. This combine should only happen if there is aChad Rosier
single use. rdar://11360370 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156316 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-07X86: optimization for -(x != 0)Manman Ren
This patch will optimize -(x != 0) on X86 FROM cmpl $0x01,%edi sbbl %eax,%eax notl %eax TO negl %edi sbbl %eax %eax In order to generate negl, I added patterns in Target/X86/X86InstrCompiler.td: def : Pat<(X86sub_flag 0, GR32:$src), (NEG32r GR32:$src)>; rdar: 10961709 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156312 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-07Add SSE4A MOVNTSS/MOVNTSD instructions.Craig Topper
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156281 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-06Switch the select to branch transformation on by default.Benjamin Kramer
The primitive conservative heuristic seems to give a slight overall improvement while not regressing stuff. Make it available to wider testing. If you notice any speed regressions (or significant code size regressions) let me know! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156258 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-05CodeGenPrepare: Add a transform to turn selects into branches in some cases.Benjamin Kramer
This came up when a change in block placement formed a cmov and slowed down a hot loop by 50%: ucomisd (%rdi), %xmm0 cmovbel %edx, %esi cmov is a really bad choice in this context because it doesn't get branch prediction. If we emit it as a branch, an out-of-order CPU can do a better job (if the branch is predicted right) and avoid waiting for the slow load+compare instruction to finish. Of course it won't help if the branch is unpredictable, but those are really rare in practice. This patch uses a dumb conservative heuristic, it turns all cmovs that have one use and a direct memory operand into branches. cmovs usually save some code size, so we disable the transform in -Os mode. In-Order architectures are unlikely to benefit as well, those are included in the "predictableSelectIsExpensive" flag. It would be better to reuse branch probability info here, but BPI doesn't support select instructions currently. It would make sense to use the same heuristics as the if-converter pass, which does the opposite direction of this transform. Test suite shows a small improvement here and there on corei7-level machines, but the actual results depend a lot on the used microarchitecture. The transformation is currently disabled by default and available by passing the -enable-cgp-select2branch flag to the code generator. Thanks to Chandler for the initial test case to him and Evan Cheng for providing me with comments and test-suite numbers that were more stable than mine :) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156234 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-04Allow v16i16 and v32i8 shuffles to be rewritten as narrower shuffles.Craig Topper
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156156 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-03Fix 256-bit vpshuflw and vpshufhw immediate encoding to handle undefs in the ↵Craig Topper
lower half correctly. Missed in r155982. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156059 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-03Fix two-address pass's aggressive instruction commuting heuristics. It's meantEvan Cheng
to catch cases like: %reg1024<def> = MOV r1 %reg1025<def> = MOV r0 %reg1026<def> = ADD %reg1024, %reg1025 r0 = MOV %reg1026 By commuting ADD, it let coalescer eliminate all of the copies. However, there was a bug in the heuristics where it ended up commuting the ADD in: %reg1024<def> = MOV r0 %reg1025<def> = MOV 0 %reg1026<def> = ADD %reg1024, %reg1025 r0 = MOV %reg1026 That did no benefit but rather ensure the last MOV would not be coalesced. rdar://11355268 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156048 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-02Revert r155853Manman Ren
The commit is intended to fix rdar://10961709. But it is the root cause of PR12720. Revert it for now. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155992 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-02Add support for selecting AVX2 vpshuflw and vpshufhw. Add decoding support ↵Craig Topper
for AsmPrinter. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155982 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-01Strip the pointer casts off of allocas so that the selection DAG can find them.Bill Wendling
PR10799 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155954 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-01X86: optimization for max-like structManman Ren
This patch will optimize the following cases on X86 (a > b) ? (a-b) : 0 (a >= b) ? (a-b) : 0 (b < a) ? (a-b) : 0 (b <= a) ? (a-b) : 0 FROM movl %edi, %ecx subl %esi, %ecx cmpl %edi, %esi movl $0, %eax cmovll %ecx, %eax TO xorl %eax, %eax subl %esi, %edi cmovll %eax, %edi movl %edi, %eax rdar: 10734411 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155919 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-30X86: optimization for -(x != 0)Manman Ren
This patch will optimize -(x != 0) on X86 FROM cmpl $0x01,%edi sbbl %eax,%eax notl %eax TO negl %edi sbbl %eax %eax git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155853 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-30test/CodeGen/X86/select.ll: remove spacesManman Ren
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155840 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-30Fix fastcc structure return with fast-isel on x86-32Derek Schuff
On x86-32, structure return via sret lets the callee pop the hidden pointer argument off the stack, which the caller then re-pushes. However if the calling convention is fastcc, then a register is used instead, and the caller should not adjust the stack. This is implemented with a check of IsTailCallConvention X86TargetLowering::LowerCall but is now checked properly in X86FastISel::DoSelectCall. (this time, actually commit what was reviewed!) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155825 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-28Reapply 155668: Fix the SD scheduler to avoid gluing the same node twice.Andrew Trick
This time, also fix the caller of AddGlue to properly handle incomplete chains. AddGlue had failure modes, but shamefully hid them from its caller. It's luck ran out. Fixes rdar://11314175: BuildSchedUnits assert. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155749 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-27Revert r155745Derek Schuff
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155746 91177308-0d34-0410-b5e6-96231b3b80d8
2012-04-27Fix fastcc structure return with fast-isel on x86-32Derek Schuff
On x86-32, structure return via sret lets the callee pop the hidden pointer argument off the stack, which the caller then re-pushes. However if the calling convention is fastcc, then a register is used instead, and the caller should not adjust the stack. This is implemented with a check of IsTailCallConvention X86TargetLowering::LowerCall but is now checked properly in X86FastISel::DoSelectCall. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@155745 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-12-30 02:25:14 +0000