Author: ctopper
Date: Wed Aug 8 12:14:23 2018
New Revision: 339281
URL: http://llvm.org/viewvc/llvm-project?rev=339281&view=rev
Log:
[CodeGen][Timers] Enable llvm::TimePassesIsEnabled when -ftime-report is
specified
r330571 added a new FrontendTimesIsEnabled variable and replaced many usages of
Author: ctopper
Date: Wed Aug 8 12:55:52 2018
New Revision: 339282
URL: http://llvm.org/viewvc/llvm-project?rev=339282&view=rev
Log:
[Builtins] Implement __builtin_clrsb to be compatible with gcc
gcc defines an intrinsic called __builtin_clrsb which counts the number of
extra sign bits on a num
Author: ctopper
Date: Wed Aug 8 13:59:40 2018
New Revision: 339287
URL: http://llvm.org/viewvc/llvm-project?rev=339287&view=rev
Log:
[Builtins] Add __builtin_clrsb support to IntExprEvaluator::VisitBuiltinCallExpr
This addresses a FIXME that has existed since before clang supported the
builtin.
Author: ctopper
Date: Wed Aug 8 14:21:21 2018
New Revision: 339289
URL: http://llvm.org/viewvc/llvm-project?rev=339289&view=rev
Log:
Revert r339287 "[Builtins] Add __builtin_clrsb support to
IntExprEvaluator::VisitBuiltinCallExpr"
This add an additional unintended change in it.
Modified:
c
Author: ctopper
Date: Wed Aug 8 15:31:14 2018
New Revision: 339296
URL: http://llvm.org/viewvc/llvm-project?rev=339296&view=rev
Log:
[VFS] Remove superfluous semicolon from unittest.
Modified:
cfe/trunk/unittests/Basic/VirtualFileSystemTest.cpp
Modified: cfe/trunk/unittests/Basic/VirtualFil
Author: ctopper
Date: Wed Aug 8 15:31:12 2018
New Revision: 339295
URL: http://llvm.org/viewvc/llvm-project?rev=339295&view=rev
Log:
[Builtins] Add __builtin_clrsb support to IntExprEvaluator::VisitBuiltinCallExpr
This addresses a FIXME that has existed since before clang supported the
builtin.
Author: ctopper
Date: Tue Aug 14 13:21:05 2018
New Revision: 339721
URL: http://llvm.org/viewvc/llvm-project?rev=339721&view=rev
Log:
[InlineAsm] Update the min-legal-vector-width function attribute based on
inputs and outputs to inline assembly
Summary:
Another piece of my ongoing to work for p
Author: ctopper
Date: Wed Aug 15 23:20:29 2018
New Revision: 339843
URL: http://llvm.org/viewvc/llvm-project?rev=339843&view=rev
Log:
[X86] Remove masking from the 512-bit padds and psubs builtins. Use select
builtin instead.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
cfe/tr
Author: ctopper
Date: Thu Aug 16 00:28:06 2018
New Revision: 339845
URL: http://llvm.org/viewvc/llvm-project?rev=339845&view=rev
Log:
[X86] Remove masking from the 512-bit paddus/psubus builtins. Use a select
builtin instead.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
cfe/tr
Author: ctopper
Date: Wed May 9 09:57:48 2018
New Revision: 331893
URL: http://llvm.org/viewvc/llvm-project?rev=331893&view=rev
Log:
[X86] Only enable the __ud2 and __int2c builtins if intrin.h has been included.
Differential Revision: https://reviews.llvm.org/D46332
Modified:
cfe/trunk/inc
Author: ctopper
Date: Wed May 9 17:05:13 2018
New Revision: 331943
URL: http://llvm.org/viewvc/llvm-project?rev=331943&view=rev
Log:
[Builtins] Improve the IR emitted for MSVC compatible rotr/rotl builtins to
match what the middle and backends understand
Previously we emitted something like
ro
Author: ctopper
Date: Wed May 9 22:43:43 2018
New Revision: 331958
URL: http://llvm.org/viewvc/llvm-project?rev=331958&view=rev
Log:
[X86] Change the implementation of scalar masked load/store intrinsics to not
use a 512-bit intermediate vector.
This is unnecessary for AVX512VL supporting CPUs
Author: ctopper
Date: Fri May 11 10:37:40 2018
New Revision: 332108
URL: http://llvm.org/viewvc/llvm-project?rev=332108&view=rev
Log:
[X86] Fix the file header name on fmaintrin.h
Modified:
cfe/trunk/lib/Headers/fmaintrin.h
Modified: cfe/trunk/lib/Headers/fmaintrin.h
URL:
http://llvm.org/vi
Author: ctopper
Date: Sun May 13 16:03:30 2018
New Revision: 332203
URL: http://llvm.org/viewvc/llvm-project?rev=332203&view=rev
Log:
[X86] Emit better code for _mm_cvtu32_sd, _mm_cvtu64_sd, _mm_cvtu32_ss, and
_mm_cvtu64_ss.
We can use direct C code for these that will use uitofp and inserteleme
Author: ctopper
Date: Sun May 13 21:05:06 2018
New Revision: 332210
URL: http://llvm.org/viewvc/llvm-project?rev=332210&view=rev
Log:
[X86] Use __builtin_convertvector to implement _mm512_cvtps_pd.
If we're using default rounding mode we can let __builtin_convertvector to
generate an fpextend. T
Author: ctopper
Date: Sun May 13 21:57:46 2018
New Revision: 332213
URL: http://llvm.org/viewvc/llvm-project?rev=332213&view=rev
Log:
[X86] Use select instrution and fpextend in the implementation of
_mm512_mask_cvtps_pd and _mm512_maskz_cvtps_pd.
Modified:
cfe/trunk/lib/Headers/avx512fintri
Author: ctopper
Date: Mon May 14 10:50:40 2018
New Revision: 332266
URL: http://llvm.org/viewvc/llvm-project?rev=332266&view=rev
Log:
[X86] Use __builtin_convertvector to replace some of the avx512 truncate
builtins.
As long as the destination type is a 256 or 128 bit vector with the same number
Author: ctopper
Date: Mon May 14 20:17:52 2018
New Revision: 332322
URL: http://llvm.org/viewvc/llvm-project?rev=332322&view=rev
Log:
[X86] Revert part of r332266: Use __builtin_convertvector to replace some of
the avx512 truncate builtins.
The masking doesn't work right in the backend for the o
Author: ctopper
Date: Fri May 18 10:18:46 2018
New Revision: 332738
URL: http://llvm.org/viewvc/llvm-project?rev=332738&view=rev
Log:
[X86] Fix a bad cast from mask16 to mask8 in _mm256_mask_cvtepi16_epi8
introduced in r332266.
Modified:
cfe/trunk/lib/Headers/avx512vlbwintrin.h
Modified: cf
Author: ctopper
Date: Sun May 20 16:34:10 2018
New Revision: 332825
URL: http://llvm.org/viewvc/llvm-project?rev=332825&view=rev
Log:
[X86] Remove mask arguments from permvar builtins/intrinsics. Use a select in
IR instead.
Someday maybe we'll use selects for all the builtins.
Modified:
cfe
Author: ctopper
Date: Sun May 20 20:36:57 2018
New Revision: 332829
URL: http://llvm.org/viewvc/llvm-project?rev=332829&view=rev
Log:
[X86] Remove some unused builtins.
These were upgraded to native shufflevectors months ago.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
cfe/tr
Author: ctopper
Date: Sun May 20 23:07:49 2018
New Revision: 332830
URL: http://llvm.org/viewvc/llvm-project?rev=332830&view=rev
Log:
[X86] Remove some preprocessor feature checks from intrinsic headers
Summary:
These look to be a couple things that weren't removed when we switched to
target att
Author: ctopper
Date: Mon May 21 13:19:17 2018
New Revision: 332882
URL: http://llvm.org/viewvc/llvm-project?rev=332882&view=rev
Log:
[X86] Use __builtin_convertvector to implement some of the packed integer to
packed float conversion intrinsics.
I believe this is safe assuming default default F
Author: ctopper
Date: Mon May 21 13:58:23 2018
New Revision: 332891
URL: http://llvm.org/viewvc/llvm-project?rev=332891&view=rev
Log:
[X86] Remove masking from pternlog llvm intrinsics and use a select instruction
instead.
Because the intrinsics in the headers are implemented as macros, we can't
Author: ctopper
Date: Mon May 21 15:10:02 2018
New Revision: 332909
URL: http://llvm.org/viewvc/llvm-project?rev=332909&view=rev
Log:
[X86] Remove a builtin that should have been removed in r332882.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
cfe/trunk/test/CodeGen/builtins-x8
Author: ctopper
Date: Mon May 21 19:02:13 2018
New Revision: 332929
URL: http://llvm.org/viewvc/llvm-project?rev=332929&view=rev
Log:
[X86] Prevent inclusion of __wmmintrin_aes.h and __wmmintrin_pclmul.h without
including wmmintrin.h
Modified:
cfe/trunk/lib/Headers/__wmmintrin_aes.h
cfe/
Author: ctopper
Date: Tue May 22 11:54:19 2018
New Revision: 333014
URL: http://llvm.org/viewvc/llvm-project?rev=333014&view=rev
Log:
[X86] Move 128-bit f16c intrinsics to __emmintrin_f16c.h include from
emmintrin.h. Move 256-bit f16c intrinsics back to f16cintrin.h
Intel documents the 128-bit v
Author: ctopper
Date: Tue May 22 13:19:05 2018
New Revision: 333020
URL: http://llvm.org/viewvc/llvm-project?rev=333020&view=rev
Log:
[X86] Add __emmintrin_f16c.h to module map and CMakeLists.
I missed this in r333014
Modified:
cfe/trunk/lib/Headers/CMakeLists.txt
cfe/trunk/lib/Headers/m
Author: ctopper
Date: Tue May 22 13:33:04 2018
New Revision: 333023
URL: http://llvm.org/viewvc/llvm-project?rev=333023&view=rev
Log:
[X86] Add two missing #endif directives to immintrin.h that should have been in
r333014.
Modified:
cfe/trunk/lib/Headers/immintrin.h
Modified: cfe/trunk/lib/
Author: ctopper
Date: Tue May 22 13:48:20 2018
New Revision: 333026
URL: http://llvm.org/viewvc/llvm-project?rev=333026&view=rev
Log:
[X86] Another attempt at fixing the intrinsic module map for rr333014.
Modified:
cfe/trunk/lib/Headers/module.modulemap
Modified: cfe/trunk/lib/Headers/module
Author: ctopper
Date: Tue May 22 13:48:24 2018
New Revision: 333027
URL: http://llvm.org/viewvc/llvm-project?rev=333027&view=rev
Log:
[X86] Remove mask argument from some builtins that are handled completely in
CGBuiltin.cpp. Just wrap a select builtin around them in the header file
instead.
Mo
Author: ctopper
Date: Tue May 22 15:19:19 2018
New Revision: 333033
URL: http://llvm.org/viewvc/llvm-project?rev=333033&view=rev
Log:
[X86] As mentioned in post-commit feedback in D47174, move the 128 bit f16c
intrinsics into f16cintrin.h and remove __emmintrin_f16c.h
These were included in emmi
Author: ctopper
Date: Tue May 22 22:51:52 2018
New Revision: 333062
URL: http://llvm.org/viewvc/llvm-project?rev=333062&view=rev
Log:
[X86] In the floating point max reduction intrinsics, negate infinity before
feeding it to set1.
Previously we negated the whole vector after splatting infinity.
Author: ctopper
Date: Tue May 22 23:31:36 2018
New Revision: 333064
URL: http://llvm.org/viewvc/llvm-project?rev=333064&view=rev
Log:
[X86] Undef the vector reduction helper macros when we're done with them.
These are implementation helper macros we shouldn't expose them to user code if
we don't
Author: ctopper
Date: Wed May 23 11:32:58 2018
New Revision: 333110
URL: http://llvm.org/viewvc/llvm-project?rev=333110&view=rev
Log:
[X86] Move all Intel defined intrinsic includes into immintrin.h
This matches the Intel documentation which shows them available by importing
immintrin.h. x86intr
Author: ctopper
Date: Wed May 23 14:04:26 2018
New Revision: 333124
URL: http://llvm.org/viewvc/llvm-project?rev=333124&view=rev
Log:
[X86] Move the include of clzerointrin.h from immintrin.h back to x86intrin.h.
This is an AMD intrinsic not an Intel intrinsic so it shouldn't be in
immintrin.h
Author: ctopper
Date: Thu May 24 10:32:49 2018
New Revision: 333211
URL: http://llvm.org/viewvc/llvm-project?rev=333211&view=rev
Log:
[X86] Fix a bad cast in _mm512_mask_abs_epi32 and _mm512_maskz_abs_epi32.
Modified:
cfe/trunk/lib/Headers/avx512fintrin.h
Modified: cfe/trunk/lib/Headers/avx5
Author: ctopper
Date: Fri May 25 15:07:43 2018
New Revision: 18
URL: http://llvm.org/viewvc/llvm-project?rev=18&view=rev
Log:
[X86] Mark a few more builtins const that were missed in r331814.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
Modified: cfe/trunk/include/clang/Ba
Author: ctopper
Date: Fri May 25 15:43:20 2018
New Revision: 20
URL: http://llvm.org/viewvc/llvm-project?rev=20&view=rev
Log:
[X86] Correct the target features on two avx512bw builtins that were
incorrectly labeled as avx512f.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
Author: ctopper
Date: Fri May 25 15:52:29 2018
New Revision: 21
URL: http://llvm.org/viewvc/llvm-project?rev=21&view=rev
Log:
[X86] Add const to another builtin that was missed from r331814.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
Modified: cfe/trunk/include/clang/Bas
Author: ctopper
Date: Sat May 26 11:55:26 2018
New Revision: 48
URL: http://llvm.org/viewvc/llvm-project?rev=48&view=rev
Log:
[X86] Remove mask from avx512ifma builtins. Use a select instruction instead.
This reduces from 12 builtins to 6 since we no longer need a mask and maskz
version.
Author: ctopper
Date: Mon May 28 20:26:38 2018
New Revision: 87
URL: http://llvm.org/viewvc/llvm-project?rev=87&view=rev
Log:
[X86] Merge the 3 different flavors of masked vpermi2var/vpermt2var builtins to
a single version without masking. Use select builtins with appropriate operand
ins
Author: ctopper
Date: Tue May 29 11:23:22 2018
New Revision: 333446
URL: http://llvm.org/viewvc/llvm-project?rev=333446&view=rev
Log:
[X86] Tag some 128/256 load/store instructions as requiring avx512vl instead of
avx512f.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
Modified: cf
Author: ctopper
Date: Tue May 29 20:38:15 2018
New Revision: 333497
URL: http://llvm.org/viewvc/llvm-project?rev=333497&view=rev
Log:
[X86] Fix the names of a bunch of icelake intrinsics.
Mostly this fixes the names of all the 128-bit intrinsics to start with _mm_
instead of _mm128_ as is the co
Author: ctopper
Date: Tue May 29 22:26:04 2018
New Revision: 333509
URL: http://llvm.org/viewvc/llvm-project?rev=333509&view=rev
Log:
[X86] Remove masking from the AVX512VNNI builtins. Use a select in IR instead.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
cfe/trunk/lib/Header
Author: ctopper
Date: Wed Oct 11 09:06:08 2017
New Revision: 315470
URL: http://llvm.org/viewvc/llvm-project?rev=315470&view=rev
Log:
[X86] Correct type for argument to clflushopt intrinsic.
Summary: According to Intel docs this should take void const *. We had char*.
The lack of const is the ma
Author: ctopper
Date: Wed Oct 11 14:42:02 2017
New Revision: 315517
URL: http://llvm.org/viewvc/llvm-project?rev=315517&view=rev
Log:
[X86] Add support for 'amdfam17h' to __builtin_cpu_is to match gcc.
The compiler-rt implementation already supported it, it just wasn't exposed.
Modified:
cfe
Author: ctopper
Date: Wed Oct 11 19:06:17 2017
New Revision: 315547
URL: http://llvm.org/viewvc/llvm-project?rev=315547&view=rev
Log:
[X86] Remove a few unnecessary check lines from the predefined-arch-macros test.
These were testing OS macros and clang/llvm macros.
Modified:
cfe/trunk/test/
Author: ctopper
Date: Thu Oct 12 10:21:01 2017
New Revision: 315594
URL: http://llvm.org/viewvc/llvm-project?rev=315594&view=rev
Log:
[X86] Use -ffreestanding instead of using the mm_malloc.h include guard hack on
more of the builtin tests.
Modified:
cfe/trunk/test/CodeGen/adc-builtins.c
Author: ctopper
Date: Thu Oct 12 11:57:15 2017
New Revision: 315607
URL: http://llvm.org/viewvc/llvm-project?rev=315607&view=rev
Log:
[X86] Add CLWB intrinsic. clang part
Reviewers: RKSimon, zvi, igorb
Reviewed By: RKSimon
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.or
Author: ctopper
Date: Fri Oct 13 11:14:24 2017
New Revision: 315723
URL: http://llvm.org/viewvc/llvm-project?rev=315723&view=rev
Log:
[X86] Add skeleton support for knm cpu
This adds support Knights Mill CPU. Preprocessor defines match gcc's
implementation.
Differential Revision: https://review
Author: ctopper
Date: Fri Oct 13 11:38:10 2017
New Revision: 315729
URL: http://llvm.org/viewvc/llvm-project?rev=315729&view=rev
Log:
[X86] Remove 'knm' defines from predefined-arch-macros.c test.
Direction seems to be that we dont' want to keep adding these, but I forgot to
remove it from the t
Author: ctopper
Date: Sun Aug 26 22:27:15 2018
New Revision: 340713
URL: http://llvm.org/viewvc/llvm-project?rev=340713&view=rev
Log:
[X86] Don't set min_vector_width to 512 on intrinsics that only operate on k
registers.
Modified:
cfe/trunk/lib/Headers/avx512fintrin.h
Modified: cfe/trunk/l
Author: ctopper
Date: Sun Aug 26 22:44:45 2018
New Revision: 340714
URL: http://llvm.org/viewvc/llvm-project?rev=340714&view=rev
Log:
[X86] Undef __DEFAULT_FN_ATTRS in avx512fintrin.h.
Fixes test failure after r340713
Modified:
cfe/trunk/lib/Headers/avx512fintrin.h
Modified: cfe/trunk/lib/H
Author: ctopper
Date: Sun Aug 26 23:20:19 2018
New Revision: 340717
URL: http://llvm.org/viewvc/llvm-project?rev=340717&view=rev
Log:
[X86] Rename __DEFAULT_FN_ATTRS to a__DEFAULT_FN_ATTRS512 in avx512dqintrin.h
and avx512bwintrin.h.
This is preparation for adding removing min_vector_width 512 f
Author: ctopper
Date: Sun Aug 26 23:20:22 2018
New Revision: 340719
URL: http://llvm.org/viewvc/llvm-project?rev=340719&view=rev
Log:
[X86] Add intrinsics for kand/kandn/knot/kor/kxnor/kxor with 8, 32, and 64-bit
mask registers.
This also adds a second intrinsic name for the 16-bit mask versions
Author: ctopper
Date: Sun Aug 26 23:20:20 2018
New Revision: 340718
URL: http://llvm.org/viewvc/llvm-project?rev=340718&view=rev
Log:
[X86] Remove min_vector_width 512 from some intrinsics that operate only on
k-registers.
Modified:
cfe/trunk/lib/Headers/avx512bwintrin.h
Modified: cfe/trunk
Author: ctopper
Date: Mon Aug 27 23:28:25 2018
New Revision: 340798
URL: http://llvm.org/viewvc/llvm-project?rev=340798&view=rev
Log:
[X86] Add kortest intrinsics for 8, 32, and 64 bit masks. Add new intrinsic
names for 16 bit masks.
This matches gcc and icc despite not being documented in the I
Author: ctopper
Date: Tue Aug 28 15:32:14 2018
New Revision: 340879
URL: http://llvm.org/viewvc/llvm-project?rev=340879&view=rev
Log:
[X86] Add kadd intrinsics to match gcc and icc.
This adds the following intrinsics:
_kadd_mask64
_kadd_mask32
_kadd_mask16
_kadd_mask8
These are missing from the
Author: ctopper
Date: Fri Aug 31 11:22:52 2018
New Revision: 341234
URL: http://llvm.org/viewvc/llvm-project?rev=341234&view=rev
Log:
[X86] Add kshift intrinsics to match gcc and icc.
This adds the following intrinsics:
_kshiftli_mask8
_kshiftli_mask16
_kshiftli_mask32
_kshiftli_mask64
_kshiftri_
Author: ctopper
Date: Fri Aug 31 13:41:06 2018
New Revision: 341251
URL: http://llvm.org/viewvc/llvm-project?rev=341251&view=rev
Log:
[X86] Add k-mask conversion and load/store instrinsics to match gcc and icc.
This adds:
_cvtmask8_u32, _cvtmask16_u32, _cvtmask32_u32, _cvtmask64_u64
_cvtu32_mask8
Author: ctopper
Date: Fri Aug 31 15:29:56 2018
New Revision: 341265
URL: http://llvm.org/viewvc/llvm-project?rev=341265&view=rev
Log:
[X86] Add ktest intrinsics to match gcc and icc.
These aren't documented in the Intel Intrinsics Guide, but are supported by gcc
and icc.
Includes these intrinsi
Author: ctopper
Date: Fri Sep 7 09:58:57 2018
New Revision: 341678
URL: http://llvm.org/viewvc/llvm-project?rev=341678&view=rev
Log:
[X86] Modify addcarry/subborrow builtins to emit an 2 result and intrinsic and
an store instruction.
This is the clang side of D51769. The llvm intrinsics now ret
Author: ctopper
Date: Fri Sep 7 12:14:24 2018
New Revision: 341699
URL: http://llvm.org/viewvc/llvm-project?rev=341699&view=rev
Log:
[X86] Custom emit __builtin_rdtscp so we can emit an explicit store for the out
parameter
This is the clang side of D51803. The llvm intrinsic now returns two res
Author: ctopper
Date: Thu Feb 8 12:16:17 2018
New Revision: 324647
URL: http://llvm.org/viewvc/llvm-project?rev=324647&view=rev
Log:
[X86] Replace kortest intrinsics with native IR.
Modified:
cfe/trunk/lib/CodeGen/CGBuiltin.cpp
cfe/trunk/test/CodeGen/avx512f-builtins.c
Modified: cfe/tru
Author: ctopper
Date: Sat Feb 10 15:34:27 2018
New Revision: 324828
URL: http://llvm.org/viewvc/llvm-project?rev=324828&view=rev
Log:
[X86] Change the signature of the AVX512 packed fp compare intrinsics to return
vXi1 mask. Make bitcasts to scalar explicit in IR
Summary: This is the clang equiv
Author: ctopper
Date: Mon Feb 12 14:38:52 2018
New Revision: 324954
URL: http://llvm.org/viewvc/llvm-project?rev=324954&view=rev
Log:
[X86] Reverse the operand order of the implementation of the kunpack builtins.
The second operand needs to be in the lower bits of the concatenation. This
matches
Author: ctopper
Date: Mon Feb 19 23:28:18 2018
New Revision: 325560
URL: http://llvm.org/viewvc/llvm-project?rev=325560&view=rev
Log:
[X86] Remove mask from 512 bit pmulhrsw/pmulhw/pmulhuw builtins.
We now use a vselect node in IR around an unmasked builtin. This makes it
consistent with the 128
Author: ctopper
Date: Tue Feb 20 16:16:50 2018
New Revision: 325655
URL: http://llvm.org/viewvc/llvm-project?rev=325655&view=rev
Log:
[X86] Disable CLWB in Cannon Lake
Cannon Lake does not support CLWB, therefore it
does not include all features listed under SKX.
Patch by Gabor Buella
Different
Author: ctopper
Date: Fri Feb 23 22:46:42 2018
New Revision: 326022
URL: http://llvm.org/viewvc/llvm-project?rev=326022&view=rev
Log:
[X86] Remove __builtin_ia32_permvarsf256_mask and
__builtin_ia32_permvarsi256_mask and use the avx2 unmasked versions and a
select instead.
Modified:
cfe/tru
Author: ctopper
Date: Sat Feb 24 10:55:13 2018
New Revision: 326039
URL: http://llvm.org/viewvc/llvm-project?rev=326039&view=rev
Log:
[X86] Remove some masked cvt builtins that can be replaced with legacy sse/avx
buiiltins and a select.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
Author: ctopper
Date: Wed May 31 14:01:22 2017
New Revision: 304326
URL: http://llvm.org/viewvc/llvm-project?rev=304326&view=rev
Log:
[TableGen] Clang changes to support Record::getValueAsString and
getValueAsListOfStrings returning StringRef instead of std::string
This is the clang version of D
Author: ctopper
Date: Wed Jun 14 20:27:58 2017
New Revision: 305439
URL: http://llvm.org/viewvc/llvm-project?rev=305439&view=rev
Log:
[Basic] Use a static_assert instead of using the old array of size -1 trick.
Modified:
cfe/trunk/include/clang/Basic/AllDiagnostics.h
Modified: cfe/trunk/incl
Author: ctopper
Date: Fri Dec 15 22:02:31 2017
New Revision: 320915
URL: http://llvm.org/viewvc/llvm-project?rev=320915&view=rev
Log:
[X86] Add builtins and tests for 128 and 256 bit vpopcntdq.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
cfe/trunk/lib/CodeGen/CGBuiltin.cpp
Author: ctopper
Date: Fri Dec 15 22:10:24 2017
New Revision: 320916
URL: http://llvm.org/viewvc/llvm-project?rev=320916&view=rev
Log:
[X86] Add the two files I forgot to commit in r320915.
Added:
cfe/trunk/lib/Headers/avx512vpopcntdqvlintrin.h
cfe/trunk/test/CodeGen/avx512vpopcntdqvlintri
Author: ctopper
Date: Sat Dec 16 00:26:22 2017
New Revision: 320919
URL: http://llvm.org/viewvc/llvm-project?rev=320919&view=rev
Log:
[X86] Implement kand/kandn/kor/kxor/kxnor/knot intrinsics using native IR.
Modified:
cfe/trunk/lib/CodeGen/CGBuiltin.cpp
cfe/trunk/test/CodeGen/avx512f-bui
Thanks!
~Craig
On Mon, Dec 18, 2017 at 3:29 AM, Martin Bohme via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: mboehme
> Date: Mon Dec 18 03:29:21 2017
> New Revision: 320971
>
> URL: http://llvm.org/viewvc/llvm-project?rev=320971&view=rev
> Log:
> [X86] Use {{.*}} instead of hardco
Author: ctopper
Date: Tue Dec 19 16:46:09 2017
New Revision: 321129
URL: http://llvm.org/viewvc/llvm-project?rev=321129&view=rev
Log:
[X86] Add more CPUID bits to cpuid.h to match gcc and support icelake features.
Modified:
cfe/trunk/lib/Headers/cpuid.h
Modified: cfe/trunk/lib/Headers/cpuid.
Author: ctopper
Date: Thu Dec 21 15:50:22 2017
New Revision: 321325
URL: http://llvm.org/viewvc/llvm-project?rev=321325&view=rev
Log:
[X86] Allow _mm_prefetch (both the header implementation and the builtin) to
accept bit 2 which is supposed to indicate the prefetched addresses will be
written t
Author: ctopper
Date: Thu Dec 21 20:51:00 2017
New Revision: 321341
URL: http://llvm.org/viewvc/llvm-project?rev=321341&view=rev
Log:
[X86] Add 'prfchw' to the correct CPUs to match the backend.
Modified:
cfe/trunk/lib/Basic/Targets/X86.cpp
cfe/trunk/test/Preprocessor/predefined-arch-macr
Author: ctopper
Date: Thu Dec 21 21:09:38 2017
New Revision: 321343
URL: http://llvm.org/viewvc/llvm-project?rev=321343&view=rev
Log:
[X86] Add missing check lines for the silvermont cases in
predefined-arch-macros.c test.
Modified:
cfe/trunk/test/Preprocessor/predefined-arch-macros.c
Modif
I meant if the command line says “-mvaes -mno-aes” we should make sure to
disable vaes
On Tue, Dec 26, 2017 at 9:47 AM coby via Phabricator <
revi...@reviews.llvm.org> wrote:
> coby added inline comments.
>
>
>
> Comment at: lib/Basic/Targets/X86.cpp:573
> setMMXLevel(Featur
Author: ctopper
Date: Wed Dec 27 14:25:59 2017
New Revision: 321502
URL: http://llvm.org/viewvc/llvm-project?rev=321502&view=rev
Log:
[X86] Enable avx512vpopcntdq and clwb for icelake.
Per table 1-1 of the October 2017 edition of Intel® Architecture Instruction
Set Extensions and Future Features
Author: ctopper
Date: Wed Dec 27 14:26:00 2017
New Revision: 321503
URL: http://llvm.org/viewvc/llvm-project?rev=321503&view=rev
Log:
[X86] Test that -march=skx enables PKU.
Modified:
cfe/trunk/test/Preprocessor/predefined-arch-macros.c
Modified: cfe/trunk/test/Preprocessor/predefined-arch-m
Author: ctopper
Date: Wed Dec 27 14:26:01 2017
New Revision: 321504
URL: http://llvm.org/viewvc/llvm-project?rev=321504&view=rev
Log:
[X86] Don't accidentally enable PKU on cannon lake and icelake or CLWB on
cannonlake.
We have cannonlake and icelake inheriting from skylake server in a switch us
Author: ctopper
Date: Thu Dec 28 22:39:16 2017
New Revision: 321547
URL: http://llvm.org/viewvc/llvm-project?rev=321547&view=rev
Log:
Revert r321504 "[X86] Don't accidentally enable PKU on cannon lake and icelake
or CLWB on cannonlake."
I based that commit on what was in Intel's public documenta
Author: ctopper
Date: Wed Jan 3 10:29:12 2018
New Revision: 321749
URL: http://llvm.org/viewvc/llvm-project?rev=321749&view=rev
Log:
[Docs] Re-generate command line documentation, primarily to get the icelake
feature command line options in, but there were a couple other changes too.
Modified:
Author: ctopper
Date: Mon Jan 8 14:37:56 2018
New Revision: 322038
URL: http://llvm.org/viewvc/llvm-project?rev=322038&view=rev
Log:
[X86] Replace cvt*2mask intrinsics with native IR using 'icmp slt X,
zeroinitializer.
Modified:
cfe/trunk/lib/CodeGen/CGBuiltin.cpp
cfe/trunk/test/CodeGen
Author: ctopper
Date: Wed Jan 10 17:37:57 2018
New Revision: 322244
URL: http://llvm.org/viewvc/llvm-project?rev=322244&view=rev
Log:
[X86][Sema] Remove constant range checks on on builtins that take a char.
The constant is already reduced to 8-bits by the time we get here and the
checks were ju
Author: ctopper
Date: Wed Jan 10 17:37:59 2018
New Revision: 322245
URL: http://llvm.org/viewvc/llvm-project?rev=322245&view=rev
Log:
[X86] Make -mavx512f imply -mfma and -mf16c in the frontend like it does in the
backend.
Similarly, make -mno-fma and -mno-f16c imply -mno-avx512f.
Withou this
Author: ctopper
Date: Wed Jan 10 17:38:02 2018
New Revision: 322247
URL: http://llvm.org/viewvc/llvm-project?rev=322247&view=rev
Log:
[X86][Sema] Range check the constant argument for the vpshld/vpshrd builtins to
ensure it fits in 8-bits.
Modified:
cfe/trunk/lib/Sema/SemaChecking.cpp
cf
Author: ctopper
Date: Wed Jan 10 17:38:00 2018
New Revision: 322246
URL: http://llvm.org/viewvc/llvm-project?rev=322246&view=rev
Log:
[X86] Fix vpshrd builtins to require an ICE for their constant argument to
match vpshld.
Modified:
cfe/trunk/include/clang/Basic/BuiltinsX86.def
Modified: cf
Author: ctopper
Date: Tue Mar 6 08:06:44 2018
New Revision: 326807
URL: http://llvm.org/viewvc/llvm-project?rev=326807&view=rev
Log:
[X86] Fix typo in cpuid.h, bit_AVX51SER->bit_AVX512ER.
Modified:
cfe/trunk/lib/Headers/cpuid.h
Modified: cfe/trunk/lib/Headers/cpuid.h
URL:
http://llvm.org/v
Author: ctopper
Date: Wed May 30 10:23:45 2018
New Revision: 333563
URL: http://llvm.org/viewvc/llvm-project?rev=333563&view=rev
Log:
[X86] Remove 'return' from a bunch of intrinsics that return void and use a
builtin that returns void.
Found by running the intrinsic headers through -pedantic -a
Author: ctopper
Date: Wed May 30 11:02:11 2018
New Revision: 333568
URL: http://llvm.org/viewvc/llvm-project?rev=333568&view=rev
Log:
[X86] Reduce the number of setzero intrinsics to just the set defined by the
Intel Intrinsics Guide.
We had quite a few for different element sizes of integers so
Author: ctopper
Date: Wed May 30 11:27:07 2018
New Revision: 333572
URL: http://llvm.org/viewvc/llvm-project?rev=333572&view=rev
Log:
[X86] Simplify the implementation of _mm_sqrt_ss, _mm_rcp_ss, and _mm_rsqrt_ss.
We don't need the insertion back into the original vector at the end. The
builtin
Author: ctopper
Date: Wed May 30 14:08:27 2018
New Revision: 333593
URL: http://llvm.org/viewvc/llvm-project?rev=333593&view=rev
Log:
[X86] Add __extension__ to a bunch of places in our intrinsic headers that fail
if you run it through -pedantic -ansi.
All of these are lines that create a 'compo
Author: ctopper
Date: Wed May 30 15:33:21 2018
New Revision: 333603
URL: http://llvm.org/viewvc/llvm-project?rev=333603&view=rev
Log:
[X86] Use C style comments in intrinsic headers for overall consistency.
Most of the origial comments used C style /* */ comments, but some C++ //
comments had sn
Author: ctopper
Date: Wed May 30 18:24:40 2018
New Revision: 333615
URL: http://llvm.org/viewvc/llvm-project?rev=333615&view=rev
Log:
[X86] Fix some places where macro arguments to intrinsics weren't cast to
_m512(i|d)/_m256(i|d/_m128(i|d) first.
The majority of the cases were correct. This fixe
Author: ctopper
Date: Wed May 30 22:02:08 2018
New Revision: 333626
URL: http://llvm.org/viewvc/llvm-project?rev=333626&view=rev
Log:
[X86] Make 512-bit unmasked load/store builtins more like their 128/256-bit
equivalents.
Previously we were just passing -1 mask to the masked builtin. This chang
1 - 100 of 2073 matches
Mail list logo
