Gitiles
Code Review
Sign In
review.blissroms.org
/
platform_external_llvm80
/
6023f6ca35a59b01f0114d1ab49ca75b0bc40aa3
6023f6c
[Pipeliner] Fix assert caused by pipeliner serialization
by Krzysztof Parzyszek
· 7 years ago
0c137c5
[InstCombine] reassociate loop invariant GEP chains to enable LICM
by Sebastian Pop
· 7 years ago
42f8f94
[Pipeliner] Enable more base+offset dependence changes in pipeliner
by Krzysztof Parzyszek
· 7 years ago
1e6d98e
[Pipeliner] Fix calculation when reusing phis
by Krzysztof Parzyszek
· 7 years ago
f0568a8
[X86][Btver2] Account for the "+i" integer pipe transfer costs (1cy use of JALU0 for GPR PRF write)
by Simon Pilgrim
· 7 years ago
2fc30a3
[Pipeliner] Fix check for order dependences when finalizing instructions
by Krzysztof Parzyszek
· 7 years ago
d5ceb00
[Pipeliner] Fix in the pipeliner phi reuse code
by Krzysztof Parzyszek
· 7 years ago
a13b439
[Pipeliner] Pipeliner should mark physical registers as used
by Krzysztof Parzyszek
· 7 years ago
b2c8940
[Pipeliner] Correctly update memoperands in the epilog
by Krzysztof Parzyszek
· 7 years ago
a76b756
[demangler] Fix a bug in r328464 found by oss-fuzz.
by Erik Pilkington
· 7 years ago
eb1ea2f
[Hexagon] Give priority to post-incremementing memory accesses in LSR
by Krzysztof Parzyszek
· 7 years ago
c6c033f
[X86][Btver2] Add CVTSD2SI/CVTSS2SI scheduler costs
by Simon Pilgrim
· 7 years ago
474404b
Migrate dockerfiles to use multi-stage builds.
by Ilya Biryukov
· 7 years ago
3e5c120
[InstCombine] distribute fmul over fadd/fsub
by Sanjay Patel
· 7 years ago
da4fb16
[X86][Btver2] Fix YMM BLENDPD/BLENDPS + UNPCKPD/UNPCKP instructions costs
by Simon Pilgrim
· 7 years ago
7fa3ae5
[llvm-mca] Fix how views are added to the InstructionTables.
by Andrea Di Biagio
· 7 years ago
4432622
[InstCombine] check uses before creating instructions for fmul distribution
by Sanjay Patel
· 7 years ago
74c171f
[X86][Btver2] Add (V)SQRTPD/(V)SQRTSD costs
by Simon Pilgrim
· 7 years ago
6e532c9
AMDGPU: Introduce common SOP_Pseudo and VOP_Pseudo TableGen base classes
by Nicolai Haehnle
· 7 years ago
fafdf4a
[llvm-mca] Add a flag -instruction-info to enable/disable the instruction info view.
by Andrea Di Biagio
· 7 years ago
33dfb9d
[llvm-mca] Update the commandline docs after r328305.
by Andrea Di Biagio
· 7 years ago
d1449c9
[X86][Btver2] Double the AGU and schedule pipe resources for YMM
by Simon Pilgrim
· 7 years ago
48972e6
[LSR] Allow giving priority to post-incrementing addressing modes
by Krzysztof Parzyszek
· 7 years ago
181ce9f
[llvm-mca] Add flag -instruction-tables to print the theoretical resource pressure distribution for instructions (PR36874)
by Andrea Di Biagio
· 7 years ago
df730ad
Test commit - adding a new line.
by Carlos Alberto Enciso
· 7 years ago
edec8af
[LoopUnroll] Fix dangling pointers in SCEV
by Max Kazantsev
· 7 years ago
7b0d030
Revert r328386 "[X86] Fix Windows `i1 zeroext` conventions to use i8 instead of i32"
by Hans Wennborg
· 7 years ago
120cf63
[DeadArgElim] Strip allocsize attributes when deleting an argument.
by Benjamin Kramer
· 7 years ago
7144163
[IRCE] Enable increasing loops of variable bounds
by Sam Parker
· 7 years ago
2c3af0c
[ARM] Simplify constructing the ARMArchFeature string. NFC.
by Martin Storsjo
· 7 years ago
640ccb5
[X86] Fix the SchedRW for intrinsic register form of SQRT/RCP/RSQRT.
by Craig Topper
· 7 years ago
0738cbc
[X86] Merge the SSE and AVX versions of fp divs and sqrts in the SandyBridge/Haswell/Broadwell/Skylake scheduler models.
by Craig Topper
· 7 years ago
a514d32
[X86] Add itinerary to intrinsic version of sqrtss, rcpss, and rsqrtss instructions.
by Craig Topper
· 7 years ago
c2d0f67
[X86] Correct the itineraries for the dot production instructions.
by Craig Topper
· 7 years ago
1f70ff8
[X86] Use the same itinerary for VCVTDQ2PD as the SSE version so that the generated scheduler classes will merge.
by Craig Topper
· 7 years ago
8816046
[X86] Swap the itineraries on the memory and register forms of CVTDQ2PD.
by Craig Topper
· 7 years ago
1214440
[X86] Give VMOVSX/ZX the same itinerary as the SSE version so they'll reuse the same generated scheduler class.
by Craig Topper
· 7 years ago
629da0d
[X86] Give vpmsadbw the same itinerary as the SSE version so they'll be able to share the same generated scheduler class.
by Craig Topper
· 7 years ago
f415690
[X86] Move (v)movss to port 5 only for Skylake. Move (v)movups/d to port 015 for Skylake.
by Craig Topper
· 7 years ago
21f2826
[demangler] Use a back-patching scheme to resolve forward references.
by Erik Pilkington
· 7 years ago
4990859
[demangler] Tweak how parameter pack sizes are determined.
by Erik Pilkington
· 7 years ago
e7042cf
[demangler] Support for clang's enable_if attribute.
by Erik Pilkington
· 7 years ago
1017678
[PatternMatch] allow undef elements when matching vector FP +0.0
by Sanjay Patel
· 7 years ago
636b39a
[X86] Use WriteResPair for WriteIDiv to cleanup sched defs. NFCI.
by Simon Pilgrim
· 7 years ago
fb87505
[SchedModel] Remove instregex entries that don't match any instructions
by Simon Pilgrim
· 7 years ago
9c2d3ad
[X86][SkylakeClient] Fix missing comma
by Simon Pilgrim
· 7 years ago
cfec5dc
[ARM] Remove sched model instregex entries that don't match any instructions (D44687)
by Simon Pilgrim
· 7 years ago
0989cb2
[X86] Add missing full stop to comment. NFCI.
by Simon Pilgrim
· 7 years ago
1241873
[InstSimplify, InstCombine] add/update tests with FP +0.0 vector with undef; NFC
by Sanjay Patel
· 7 years ago
d40afef
[X86][SkylakeClient] Fix a set of regular expressions that were checking for optionally starting with 'Y' instead of 'V'
by Craig Topper
· 7 years ago
feb3905
[X86][MMX] MOVQ2DQ/MOVDQ2Q are better described as WriteVecMove than WriteMove
by Simon Pilgrim
· 7 years ago
96cdb99
[X86][SkylakeServer] Merge multiple instregex. NFCI
by Simon Pilgrim
· 7 years ago
e192eb0
[X86] Update cost model for Goldmont. Add fsqrt costs for Silvermont
by Craig Topper
· 7 years ago
9c0bd4b
[InstCombine] adjust test comments; NFC
by Sanjay Patel
· 7 years ago
eb6f6af
[InstCombine] consolidate casted icmp vector tests
by Sanjay Patel
· 7 years ago
aef1b11
[InstCombine] peek through more icmp of FP cast + bitcast
by Sanjay Patel
· 7 years ago
7a1a91f
[X86] Add the ability to override memory folding latency to schedules and add 1uop for memory folds for Intel models
by Simon Pilgrim
· 7 years ago
04dad85
[X86] Consistently prefix all defs in X86ScheduleSLM.td with 'SLM'.
by Craig Topper
· 7 years ago
6c053b7
[X86] Update a partially stale comment, since SVN r328386. NFC.
by Martin Storsjo
· 7 years ago
419fa24
[SchedModel] Remove an unneeded temporary vector.
by Craig Topper
· 7 years ago
284e1ff
[SchedModel] Use std::move in a couple places to reduce copying
by Craig Topper
· 7 years ago
af807a6
[SchedModel] Use std::move to replace a vector instead of vector::swap
by Craig Topper
· 7 years ago
a078148
Fix module.modulemap after r328395
by Eric Fiselier
· 7 years ago
9201665
[SchedModel] Remove std::vectors that were created with 1 element and then passed to an ArrayRef parameter.
by Craig Topper
· 7 years ago
ded7cc4
[SchedModel] Record::getName() returns StringRef - avoid std::string creation. NFCI.
by Simon Pilgrim
· 7 years ago
79483b2
[SchedModel] Avoid std::string creation for instregex patterns that don't contain regex metas. NFCI.
by Simon Pilgrim
· 7 years ago
9a7f9fe
[X86][SkylakeClient] Merge xmm/ymm instructions instregex entries to reduce regex matches to reduce compile time
by Simon Pilgrim
· 7 years ago
c6d3a5e
[X86][Broadwell] Merge xmm/ymm instructions instregex entries to reduce regex matches to reduce compile time
by Simon Pilgrim
· 7 years ago
df93152
[RISCV] Use init_array instead of ctors for RISCV target, by default
by Mandeep Singh Grang
· 7 years ago
329aaa6
[X86][Haswell] Merge xmm/ymm instructions instregex entries to reduce regex matches to reduce compile time
by Simon Pilgrim
· 7 years ago
43a7507
[X86][SandyBridge] Merge xmm/ymm instructions instregex entries to reduce regex matches to reduce compile time
by Simon Pilgrim
· 7 years ago
7780649
[Hexagon] Change std::sort to llvm::sort in response to r327219
by Mandeep Singh Grang
· 7 years ago
a71df88
[AMDGPU] Change std::sort to llvm::sort in response to r327219
by Mandeep Singh Grang
· 7 years ago
add7b3e
[llvm-mca] run clang-format on all files.
by Andrea Di Biagio
· 7 years ago
7525290
[llvm-mca] Remove unused field in InstrBuilder. NFC
by Andrea Di Biagio
· 7 years ago
51bf112
[InstCombine] peek through FP casts for sign-bit compares (PR36682)
by Sanjay Patel
· 7 years ago
97758c9
[InstCombine] fix formatting; NFC
by Sanjay Patel
· 7 years ago
b747538
[X86][AES] Ensure we're testing both non-VEX/VEX variants of AES instructions on AVX targets
by Simon Pilgrim
· 7 years ago
b5d6a0f
[X86][SSE] Ensure we're testing both non-VEX/VEX variants of SSE instructions on AVX targets
by Simon Pilgrim
· 7 years ago
9055038
[InstCombine] add multi-use/vector tests for intrinsic shrinking; NFC
by Sanjay Patel
· 7 years ago
02ccc54
[X86][AVX1] Ensure we don't use later instruction sets in AVX1 schedule tests
by Simon Pilgrim
· 7 years ago
f5891a0
[X86][AVX2] Ensure we don't use later instruction sets in AVX2 schedule tests
by Simon Pilgrim
· 7 years ago
7cff411
[X86] Add a new disassembler opcode map for 3DNow. Stop treating 3DNow as an attribute.
by Craig Topper
· 7 years ago
2280dbe
[X86] Use unique_ptr to simplify memory management. NFC
by Craig Topper
· 7 years ago
ce08cb3
[X86] Use X86_INSTR_MRM_MAPPING macro instead of listing all MRM_C0-MRM_FF format encodings. NFC
by Craig Topper
· 7 years ago
4f709f6
[X86] Remove an unnecessary switch around two other switches. NFC
by Craig Topper
· 7 years ago
f0c0da8
[X86] Merge the Has3DNow0F0FOpcode TSFlag into the OpMap encoding. NFC
by Craig Topper
· 7 years ago
0a0437f
Add REQUIRES lines for the targets being checked in this test.
by Eric Christopher
· 7 years ago
b97c8e6
[X86] Add a DAG combine to simplify PMULDQ/PMULUDQ nodes
by Craig Topper
· 7 years ago
3997401
Allow FDE references outside the +/-2GB range supported by PC relative
by Eric Christopher
· 7 years ago
679790c
Remove unused header from EntryExitInstrumenter
by David Blaikie
· 7 years ago
fbb1ffb
[X86] Correct the value AdSizeX in X86II enum. NFC
by Craig Topper
· 7 years ago
b91d9a7
Fix layering by moving ValueTypes.h from CodeGen to IR
by David Blaikie
· 7 years ago
b933411
Fix layering of CodeGen/TargetOpcodes.def by moving it to Support
by David Blaikie
· 7 years ago
9d9a46a
Fix layering of MachineValueType.h by moving it from CodeGen to Support
by David Blaikie
· 7 years ago
dae3bbe
Fix layering by moving Support/CodeGenCWrappers.h to Target
by David Blaikie
· 7 years ago
cb6b3af
Fix layering by moving X86DisassemblerDecoderCommon to Support
by David Blaikie
· 7 years ago
fe42bd5
Move TargetLoweringObjectFile from CodeGen to Target to fix layering
by David Blaikie
· 7 years ago
4dcdb90
[GuardWidening] Group code by class [NFC]
by Philip Reames
· 7 years ago
e8b9e31
[X86] Fix Windows `i1 zeroext` conventions to use i8 instead of i32
by Reid Kleckner
· 7 years ago
Next »