Page MenuHomePhabricator

miyuki (Mikhail Maltsev)
User

Projects

User does not belong to any projects.

User Details

User Since
Nov 6 2017, 10:05 AM (83 w, 6 d)

Recent Activity

Fri, Jun 14

miyuki committed rGd1cc2e1543e7: [ARM] Add MVE horizontal accumulation instructions (authored by miyuki).
[ARM] Add MVE horizontal accumulation instructions
Fri, Jun 14, 7:29 AM
miyuki committed rL363403: [ARM] Add MVE horizontal accumulation instructions.
[ARM] Add MVE horizontal accumulation instructions
Fri, Jun 14, 7:29 AM
miyuki closed D62670: [ARM] Add MVE horizontal accumulation instructions..
Fri, Jun 14, 7:29 AM · Restricted Project
miyuki created D63328: [libc++] Avoid using timespec when it might not be available.
Fri, Jun 14, 4:58 AM · Restricted Project
miyuki committed rG411c630bae0e: [libc++] Keep __regex_word in sync with ctype_base (authored by miyuki).
[libc++] Keep __regex_word in sync with ctype_base
Fri, Jun 14, 2:02 AM
miyuki committed rL363363: [libc++] Keep __regex_word in sync with ctype_base.
[libc++] Keep __regex_word in sync with ctype_base
Fri, Jun 14, 2:01 AM
miyuki closed D63284: [libc++] Keep __regex_word in sync with ctype_base.
Fri, Jun 14, 2:01 AM · Restricted Project, Restricted Project

Thu, Jun 13

miyuki created D63289: [libc++] Add missing #include in tests.
Thu, Jun 13, 10:45 AM · Restricted Project, Restricted Project
miyuki created D63284: [libc++] Keep __regex_word in sync with ctype_base.
Thu, Jun 13, 10:16 AM · Restricted Project, Restricted Project
miyuki updated the diff for D62670: [ARM] Add MVE horizontal accumulation instructions..
  • Rebase
  • Avoid TableGen logic in instruction operands and constraints
Thu, Jun 13, 8:15 AM · Restricted Project
miyuki added inline comments to D62670: [ARM] Add MVE horizontal accumulation instructions..
Thu, Jun 13, 5:52 AM · Restricted Project

Wed, Jun 12

miyuki added inline comments to D62671: [ARM] Add MVE integer vector min/max instructions..
Wed, Jun 12, 8:57 AM · Restricted Project
miyuki updated the diff for D62670: [ARM] Add MVE horizontal accumulation instructions..
  • Fixed naming inconsistencies
  • Removed comments about undefined instructions
  • Factored out common parts of:
    • VMINV and VMAXV
    • VMINNMV and VMAXNMV
    • VMLADAV and VMLSDAV
    • VMLALDAV, VMLSLDAV (includes VRMLSLDAVH) and VRMLALDAVH
  • Fixed inputs and constraints of non-accumulating instructions
  • Split the assembler test into integer and floating point parts
Wed, Jun 12, 6:17 AM · Restricted Project
miyuki commandeered D62670: [ARM] Add MVE horizontal accumulation instructions..
Wed, Jun 12, 6:04 AM · Restricted Project

Tue, Jun 11

miyuki added inline comments to D62670: [ARM] Add MVE horizontal accumulation instructions..
Tue, Jun 11, 10:32 AM · Restricted Project
miyuki added inline comments to D62669: [ARM] Set up infrastructure for MVE vector instructions..
Tue, Jun 11, 6:06 AM · Restricted Project
miyuki committed rG7bd5c55cadc9: [ARM] First MVE instructions: scalar shifts. (authored by miyuki).
[ARM] First MVE instructions: scalar shifts.
Tue, Jun 11, 5:04 AM
miyuki committed rL363051: [ARM] First MVE instructions: scalar shifts..
[ARM] First MVE instructions: scalar shifts.
Tue, Jun 11, 5:01 AM
miyuki closed D62668: [ARM] First MVE instructions: scalar shifts..
Tue, Jun 11, 5:01 AM · Restricted Project
miyuki set the repository for D62668: [ARM] First MVE instructions: scalar shifts. to rG LLVM Github Monorepo.
Tue, Jun 11, 4:53 AM · Restricted Project
miyuki updated the diff for D62668: [ARM] First MVE instructions: scalar shifts..
Tue, Jun 11, 3:39 AM · Restricted Project
miyuki updated the diff for D62668: [ARM] First MVE instructions: scalar shifts..

Fixed the tablegen file and tests according to reviewer's comments. Added missing cases to ARMRegisterBankInfo.cpp

Tue, Jun 11, 2:51 AM · Restricted Project
miyuki commandeered D62668: [ARM] First MVE instructions: scalar shifts..
Tue, Jun 11, 2:48 AM · Restricted Project

Fri, Jun 7

miyuki added inline comments to D62667: [ARM] Add the non-MVE instructions in Arm v8.1-M..
Fri, Jun 7, 10:58 AM · Restricted Project
miyuki created D63006: [libunwind][AArch64] Fix libunwind::Registers_arm64::jumpto.
Fri, Jun 7, 7:18 AM · Restricted Project

Tue, Jun 4

miyuki added a comment to D62651: [ARM] Add FP16 vector insert/extract patterns.

LGTM

We could possibly use a custom inserter to generate the vins sequence, but it would probably involve some benchmarking to make sure there aren't any unexpected performance penalties due to the weird register usage. So I'm happy to put that off for now.

(On a side-note, I think you can insert a float into element zero of a vector with two vext instructions, which is the same number of instructions, but maybe lower latency.)

Tue, Jun 4, 2:56 AM · Restricted Project
miyuki committed rG08da01b49648: [ARM] Add FP16 vector insert/extract patterns (authored by miyuki).
[ARM] Add FP16 vector insert/extract patterns
Tue, Jun 4, 2:38 AM
miyuki committed rL362482: [ARM] Add FP16 vector insert/extract patterns.
[ARM] Add FP16 vector insert/extract patterns
Tue, Jun 4, 2:37 AM
miyuki closed D62651: [ARM] Add FP16 vector insert/extract patterns.
Tue, Jun 4, 2:37 AM · Restricted Project

Fri, May 31

miyuki updated the diff for D62651: [ARM] Add FP16 vector insert/extract patterns.

Changed extraction patterns to avoid using GPRs as intermediate registers.

Fri, May 31, 3:55 AM · Restricted Project

Thu, May 30

miyuki created D62651: [ARM] Add FP16 vector insert/extract patterns.
Thu, May 30, 4:42 AM · Restricted Project

May 10 2019

miyuki added a comment to D51329: [Attribute/Diagnostics] Print macro if definition is an attribute declaration.

This caused https://bugs.llvm.org/show_bug.cgi?id=41835

May 10 2019, 9:47 AM · Restricted Project

Feb 26 2019

miyuki added inline comments to D57335: [IR] Don't assume all functions are 4 byte aligned.
Feb 26 2019, 2:04 AM · Restricted Project, Restricted Project

Feb 19 2019

miyuki added inline comments to D57896: Variable names rule.
Feb 19 2019, 3:43 AM · Restricted Project, Restricted Project

Oct 24 2018

miyuki committed rL345148: Adjust unsupported C++ versions in some tests.
Adjust unsupported C++ versions in some tests
Oct 24 2018, 8:12 AM
miyuki committed rCXX345148: Adjust unsupported C++ versions in some tests.
Adjust unsupported C++ versions in some tests
Oct 24 2018, 8:12 AM
miyuki closed D53640: Adjust unsupported C++ versions in some tests.
Oct 24 2018, 8:11 AM
miyuki closed D53640: Adjust unsupported C++ versions in some tests.
Oct 24 2018, 8:11 AM
miyuki created D53640: Adjust unsupported C++ versions in some tests.
Oct 24 2018, 4:39 AM

Sep 24 2018

Herald updated subscribers of D46443: Add missing cstdalign header.
Sep 24 2018, 8:57 AM

Sep 17 2018

miyuki committed rL342365: [Analyzer] Define and use diff_plist in tests, NFC.
[Analyzer] Define and use diff_plist in tests, NFC
Sep 17 2018, 3:22 AM
miyuki committed rC342365: [Analyzer] Define and use diff_plist in tests, NFC.
[Analyzer] Define and use diff_plist in tests, NFC
Sep 17 2018, 3:22 AM
miyuki closed D52036: [Analyzer] Use diff_plist in tests, NFC.
Sep 17 2018, 3:22 AM
miyuki closed D52036: [Analyzer] Use diff_plist in tests, NFC.
Sep 17 2018, 3:22 AM
miyuki abandoned D52086: [lit] Add a substitution for Clang Analyzer tests.
Sep 17 2018, 3:16 AM
miyuki removed a parent revision for D52036: [Analyzer] Use diff_plist in tests, NFC: D52086: [lit] Add a substitution for Clang Analyzer tests.
Sep 17 2018, 3:15 AM
miyuki removed a child revision for D52086: [lit] Add a substitution for Clang Analyzer tests: D52036: [Analyzer] Use diff_plist in tests, NFC.
Sep 17 2018, 3:15 AM

Sep 14 2018

miyuki added a child revision for D52086: [lit] Add a substitution for Clang Analyzer tests: D52036: [Analyzer] Use diff_plist in tests, NFC.
Sep 14 2018, 4:01 AM
miyuki added a parent revision for D52036: [Analyzer] Use diff_plist in tests, NFC: D52086: [lit] Add a substitution for Clang Analyzer tests.
Sep 14 2018, 4:01 AM
miyuki added a child revision for D52036: [Analyzer] Use diff_plist in tests, NFC: D52086: [lit] Add a substitution for Clang Analyzer tests.
Sep 14 2018, 4:01 AM
miyuki removed a parent revision for D52086: [lit] Add a substitution for Clang Analyzer tests: D52036: [Analyzer] Use diff_plist in tests, NFC.
Sep 14 2018, 4:01 AM
miyuki removed a child revision for D52036: [Analyzer] Use diff_plist in tests, NFC: D52086: [lit] Add a substitution for Clang Analyzer tests.
Sep 14 2018, 4:01 AM
miyuki added a parent revision for D52086: [lit] Add a substitution for Clang Analyzer tests: D52036: [Analyzer] Use diff_plist in tests, NFC.
Sep 14 2018, 4:01 AM
miyuki updated the diff for D52036: [Analyzer] Use diff_plist in tests, NFC.

Use a different approach (suggested by Artem).

Sep 14 2018, 4:01 AM
miyuki created D52086: [lit] Add a substitution for Clang Analyzer tests.
Sep 14 2018, 3:56 AM

Sep 13 2018

miyuki created D52036: [Analyzer] Use diff_plist in tests, NFC.
Sep 13 2018, 7:53 AM

Sep 12 2018

miyuki committed rL342034: [CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed.
[CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed
Sep 12 2018, 2:20 AM
miyuki committed rC342034: [CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed.
[CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed
Sep 12 2018, 2:20 AM
miyuki closed D50507: [CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed.
Sep 12 2018, 2:20 AM

Sep 11 2018

miyuki abandoned D49987: [ARM] Make FP16 vectors legal.

Abandoning in favor of https://reviews.llvm.org/D50507

Sep 11 2018, 2:35 AM

Sep 3 2018

miyuki added a comment to D50507: [CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed.

ping

Sep 3 2018, 2:27 AM

Aug 13 2018

miyuki updated the diff for D50507: [CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed.

Handle return of homogeneous aggregates

Aug 13 2018, 3:18 AM

Aug 10 2018

miyuki updated the diff for D50507: [CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed.

Fixed handling of homogeneous aggregates of FP16 vectors

Aug 10 2018, 9:04 AM

Aug 9 2018

miyuki added a comment to D49987: [ARM] Make FP16 vectors legal.

Alternative (Clang) patch: https://reviews.llvm.org/D50507

Aug 9 2018, 5:06 AM
miyuki created D50507: [CodeGen][ARM] Coerce FP16 vectors to integer vectors when needed.
Aug 9 2018, 5:03 AM

Aug 7 2018

miyuki updated the diff for D49987: [ARM] Make FP16 vectors legal.

I have added a hardfp test. Clang has some logic to handle scalar FP16 values, it performs softening when needed (e.g. tools/clang/test/CodeGen/arm-fp16-arguments.c), whereas NEON vectors are lowered into LLVM vector types. So we don't need additional changes neither in Clang nor in the LLVM scalar FP16 handling.

Aug 7 2018, 6:23 AM

Aug 1 2018

miyuki updated the diff for D49987: [ARM] Make FP16 vectors legal.

This approach worked, but required handling two more operations in PromoteFloatOperand: PromoteFloatOp_BUILD_VECTOR and PromoteFloatOp_INSERT_VECTOR_ELT, which in my implementation don't actually promote anything but rather do some FP softening. Not sure if there is a better solution.

Aug 1 2018, 8:51 AM

Jul 31 2018

miyuki added a comment to D49987: [ARM] Make FP16 vectors legal.

I tried compiling a fadd <4 x half> and it actually does not work: the type gets legalized into v4i16 and we get an unknown libcall. So, the problem is that we need to soften v4f16 to v4i16 when passing it as a function parameter, but at the same time expand to f32 when performing arithmetics on it. Do you have any suggestion how to implement this correctly? Do any other targets face a similar problem?

Jul 31 2018, 7:22 AM

Jul 30 2018

miyuki created D49987: [ARM] Make FP16 vectors legal.
Jul 30 2018, 7:50 AM

Jul 4 2018

Herald added a reviewer for D44656: [libcxx] Add <experimental/simd> conversions between underlying non-portable types and simd/simd_mask types.: javed.absar.
Jul 4 2018, 7:34 AM
Herald updated subscribers of D41844: [libcxx] implement <experimental/simd> mask reductions.
Jul 4 2018, 7:34 AM
Herald updated subscribers of D41845: [libcxx] clean up and complete <experimental/simd>.
Jul 4 2018, 7:34 AM
Herald updated subscribers of D44658: [libcxx] Implement <experimental/simd> aligned load and store when compiled with Clang..
Jul 4 2018, 7:34 AM
Herald updated subscribers of D44657: [libcxx] fix a sanitizer-reported bug in <experimental/simd>.
Jul 4 2018, 7:34 AM
Herald updated subscribers of D44659: [libcxx] Optimize <experimental/simd> -O0 performance for operators.
Jul 4 2018, 7:34 AM
Herald updated subscribers of D44661: [libcxx] optimize reduce(), hmin(), hmax() by reordering the operations..
Jul 4 2018, 7:34 AM
Herald updated subscribers of D44660: [libcxx] unroll the loops in <experimental/simd> for Clang, until LLVM bugs are fixed.
Jul 4 2018, 7:34 AM
Herald updated subscribers of D44662: [libcxx] In <experimental/simd>, optimize masked div and rem..
Jul 4 2018, 7:34 AM
Herald updated subscribers of D44663: [libcxx] Update <experimental/simd> with R9 changes.
Jul 4 2018, 7:34 AM
Herald updated subscribers of D44664: [libcxx] Add missing <experimental/simd> __simd_reference pieces based on R9..
Jul 4 2018, 7:33 AM
Herald updated subscribers of D44665: [libcxx] Update synopsis to P0214R9.
Jul 4 2018, 7:33 AM
Herald updated subscribers of D41376: [libcxx] Implement <simd> ABI for Clang/GCC vector extension, constructors, copy_from and copy_to..
Jul 4 2018, 7:32 AM

Jun 11 2018

miyuki committed rL334414: [Driver] Add aliases for -Qn/-Qy.
[Driver] Add aliases for -Qn/-Qy
Jun 11 2018, 9:14 AM
miyuki committed rC334414: [Driver] Add aliases for -Qn/-Qy.
[Driver] Add aliases for -Qn/-Qy
Jun 11 2018, 9:14 AM
miyuki closed D48021: [Driver] Add aliases for -Qn/-Qy.
Jun 11 2018, 9:14 AM
miyuki created D48021: [Driver] Add aliases for -Qn/-Qy.
Jun 11 2018, 6:11 AM
miyuki committed rL334394: [Unittests] Change linker flags of dynamic library tests.
[Unittests] Change linker flags of dynamic library tests
Jun 11 2018, 2:20 AM
miyuki closed D47469: [Unittests] Change linker flags of dynamic library tests.
Jun 11 2018, 2:20 AM

Jun 5 2018

miyuki added a comment to D47469: [Unittests] Change linker flags of dynamic library tests.

x86_64 linux

Jun 5 2018, 12:04 AM

May 29 2018

miyuki retitled D47469: [Unittests] Change linker flags of dynamic library tests from Unittests: Change linker flags of dynamic library tests to [Unittests] Change linker flags of dynamic library tests.
May 29 2018, 3:58 AM
miyuki created D47469: [Unittests] Change linker flags of dynamic library tests.
May 29 2018, 3:51 AM

Apr 30 2018

miyuki committed rL331164: [Targets] Implement getConstraintRegister for ARM and AArch64.
[Targets] Implement getConstraintRegister for ARM and AArch64
Apr 30 2018, 2:14 AM
miyuki committed rC331164: [Targets] Implement getConstraintRegister for ARM and AArch64.
[Targets] Implement getConstraintRegister for ARM and AArch64
Apr 30 2018, 2:14 AM
miyuki closed D45965: [Targets] Implement getConstraintRegister for ARM and AArch64.
Apr 30 2018, 2:14 AM

Apr 27 2018

miyuki abandoned D44781: [IR] Add function pointer alignment to DataLayout.

Abandoning (https://reviews.llvm.org/D46110 has been committed instead)

Apr 27 2018, 2:21 AM
miyuki abandoned D44783: [Targets] Adjust ARM data layout.
Apr 27 2018, 2:18 AM
miyuki committed rL331025: [IR] Do not assume that function pointers are aligned.
[IR] Do not assume that function pointers are aligned
Apr 27 2018, 2:15 AM
miyuki closed D46110: [IR] Do not assume that function pointers are aligned.
Apr 27 2018, 2:15 AM

Apr 26 2018

miyuki updated the diff for D46110: [IR] Do not assume that function pointers are aligned.

Added a comment

Apr 26 2018, 10:03 AM
miyuki created D46110: [IR] Do not assume that function pointers are aligned.
Apr 26 2018, 2:43 AM