maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/clang/lib/Headers/avx512pfintrin.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

98 lines
4.5 KiB
C
Raw Normal View History

[X86] Fix some inconsistent formatting in the first line of our intrinsics headers. Some were too long and some were too short. llvm-svn: 331559
2018-05-05 05:45:25 +08:00
/*===------------- avx512pfintrin.h - PF intrinsics ------------------------===
[Clang][AVX512][BuiltIn] Adding intrinsics of VGATHER{DPS|DPD} , VPGATHER{QD|QQ|DD|DQ} and VGATHERPF{0|1}{DPS|QPS|DPD|QPD} instruction set . Differential Revision: http://reviews.llvm.org/D19224 llvm-svn: 266983
2016-04-21 20:47:27 +08:00
*
*
Move the builtin headers to use the new license file header. Summary: These all had somewhat custom file headers with different text from the ones I searched for previously, and so I missed them. Thanks to Hal and Kristina and others who prompted me to fix this, and sorry it took so long. Reviewers: hfinkel Subscribers: mcrosier, javed.absar, cfe-commits Tags: #clang Differential Revision: https://reviews.llvm.org/D60406 llvm-svn: 357941
2019-04-09 04:51:30 +08:00
* Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
* See https://llvm.org/LICENSE.txt for license information.
* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
[Clang][AVX512][BuiltIn] Adding intrinsics of VGATHER{DPS|DPD} , VPGATHER{QD|QQ|DD|DQ} and VGATHERPF{0|1}{DPS|QPS|DPD|QPD} instruction set . Differential Revision: http://reviews.llvm.org/D19224 llvm-svn: 266983
2016-04-21 20:47:27 +08:00
*
*===-----------------------------------------------------------------------===
*/
#ifndef __IMMINTRIN_H
#error "Never use <avx512pfintrin.h> directly; include <immintrin.h> instead."
#endif
#ifndef __AVX512PFINTRIN_H
#define __AVX512PFINTRIN_H
/* Define the default attributes for the functions in this file. */
#define __DEFAULT_FN_ATTRS __attribute__((__always_inline__, __nodebug__, __target__("avx512pf")))
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_mask_prefetch_i32gather_pd(index, mask, addr, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_gatherpfdpd((__mmask8)(mask), (__v8si)(__m256i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void const *)(addr), (int)(scale), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(hint))
#define _mm512_prefetch_i32gather_pd(index, addr, scale, hint) \
[Clang][avx512][Intrinsics] adding prefetch gather intrinsics Differential Revision: http://reviews.llvm.org/D21322 llvm-svn: 272667
2016-06-14 21:45:17 +08:00
__builtin_ia32_gatherpfdpd((__mmask8) -1, (__v8si)(__m256i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void const *)(addr), (int)(scale), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(hint))
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_mask_prefetch_i32gather_ps(index, mask, addr, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_gatherpfdps((__mmask16)(mask), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(__v16si)(__m512i)(index), (void const *)(addr), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(scale), (int)(hint))
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_prefetch_i32gather_ps(index, addr, scale, hint) \
[Clang][avx512][Intrinsics] adding prefetch gather intrinsics Differential Revision: http://reviews.llvm.org/D21322 llvm-svn: 272667
2016-06-14 21:45:17 +08:00
__builtin_ia32_gatherpfdps((__mmask16) -1, \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(__v16si)(__m512i)(index), (void const *)(addr), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(scale), (int)(hint))
[Clang][avx512][Intrinsics] adding prefetch gather intrinsics Differential Revision: http://reviews.llvm.org/D21322 llvm-svn: 272667
2016-06-14 21:45:17 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_mask_prefetch_i64gather_pd(index, mask, addr, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_gatherpfqpd((__mmask8)(mask), (__v8di)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void const *)(addr), (int)(scale), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(hint))
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_prefetch_i64gather_pd(index, addr, scale, hint) \
[Clang][avx512][Intrinsics] adding prefetch gather intrinsics Differential Revision: http://reviews.llvm.org/D21322 llvm-svn: 272667
2016-06-14 21:45:17 +08:00
__builtin_ia32_gatherpfqpd((__mmask8) -1, (__v8di)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void const *)(addr), (int)(scale), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(hint))
#define _mm512_mask_prefetch_i64gather_ps(index, mask, addr, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_gatherpfqps((__mmask8)(mask), (__v8di)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void const *)(addr), (int)(scale), (int)(hint))
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_prefetch_i64gather_ps(index, addr, scale, hint) \
[Clang][avx512][Intrinsics] adding prefetch gather intrinsics Differential Revision: http://reviews.llvm.org/D21322 llvm-svn: 272667
2016-06-14 21:45:17 +08:00
__builtin_ia32_gatherpfqps((__mmask8) -1, (__v8di)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void const *)(addr), (int)(scale), (int)(hint))
[Clang][avx512][Intrinsics] adding prefetch gather intrinsics Differential Revision: http://reviews.llvm.org/D21322 llvm-svn: 272667
2016-06-14 21:45:17 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_prefetch_i32scatter_pd(addr, index, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_scatterpfdpd((__mmask8)-1, (__v8si)(__m256i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void *)(addr), (int)(scale), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(hint))
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_mask_prefetch_i32scatter_pd(addr, mask, index, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_scatterpfdpd((__mmask8)(mask), (__v8si)(__m256i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void *)(addr), (int)(scale), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(hint))
[Clang][AVX512][BUILTIN] Adding intrinsics for VSCATTERPF{1|0}{DPS|QPS|DPD|QPD} instruction set Differential Revision: http://reviews.llvm.org/D19313 llvm-svn: 267398
2016-04-25 21:01:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_prefetch_i32scatter_ps(addr, index, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_scatterpfdps((__mmask16)-1, (__v16si)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void *)(addr), (int)(scale), (int)(hint))
[Clang][AVX512][BUILTIN] Adding intrinsics for VSCATTERPF{1|0}{DPS|QPS|DPD|QPD} instruction set Differential Revision: http://reviews.llvm.org/D19313 llvm-svn: 267398
2016-04-25 21:01:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_mask_prefetch_i32scatter_ps(addr, mask, index, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_scatterpfdps((__mmask16)(mask), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(__v16si)(__m512i)(index), (void *)(addr), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(scale), (int)(hint))
[Clang][AVX512][BUILTIN] Adding intrinsics for VSCATTERPF{1|0}{DPS|QPS|DPD|QPD} instruction set Differential Revision: http://reviews.llvm.org/D19313 llvm-svn: 267398
2016-04-25 21:01:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_prefetch_i64scatter_pd(addr, index, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_scatterpfqpd((__mmask8)-1, (__v8di)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void *)(addr), (int)(scale), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(hint))
[Clang][AVX512][BUILTIN] Adding intrinsics for VSCATTERPF{1|0}{DPS|QPS|DPD|QPD} instruction set Differential Revision: http://reviews.llvm.org/D19313 llvm-svn: 267398
2016-04-25 21:01:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_mask_prefetch_i64scatter_pd(addr, mask, index, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_scatterpfqpd((__mmask8)(mask), (__v8di)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void *)(addr), (int)(scale), \
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
(int)(hint))
[Clang][AVX512][BUILTIN] Adding intrinsics for VSCATTERPF{1|0}{DPS|QPS|DPD|QPD} instruction set Differential Revision: http://reviews.llvm.org/D19313 llvm-svn: 267398
2016-04-25 21:01:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_prefetch_i64scatter_ps(addr, index, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_scatterpfqps((__mmask8)-1, (__v8di)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void *)(addr), (int)(scale), (int)(hint))
[Clang][AVX512][BUILTIN] Adding intrinsics for VSCATTERPF{1|0}{DPS|QPS|DPD|QPD} instruction set Differential Revision: http://reviews.llvm.org/D19313 llvm-svn: 267398
2016-04-25 21:01:40 +08:00
[X86] Remove __extension__ from macro intrinsics when its not needed. I think this is a holdover from when we used to declare variables inside the macros. And then its been copy and pasted forward for years every time a new macro intrinsic gets added. Interestingly this caused some tests for IRGen to be slightly more optimized. We now return a zeroinitializer directly instead of going through a store+load. It also removed a bogus error message on another test. llvm-svn: 333613
2018-05-31 08:51:20 +08:00
#define _mm512_mask_prefetch_i64scatter_ps(addr, mask, index, scale, hint) \
[AVX512] Add parentheses around macro arguments in AVX512PF intrinsics. Remove leading underscores from macro argument names. Add explicit typecasts to all macro arguments and return values. And finally reformat after all the adjustments. This is a mostly mechanical change accomplished with a script. I tried to split out any changes to the typecasts that already existed into separate commits. llvm-svn: 269742
2016-05-17 12:41:40 +08:00
__builtin_ia32_scatterpfqps((__mmask8)(mask), (__v8di)(__m512i)(index), \
[X86] Make the pointer arguments to avx512 gather/scatter intrinsics 'void*' to match gcc and Intel's documentation. The avx2 gather intrinsics are documented to use 'int', 'long long', 'float', or 'double' *. So I'm leaving those. This matches gcc. llvm-svn: 350696
2019-01-09 15:36:01 +08:00
(void *)(addr), (int)(scale), (int)(hint))
[Clang][AVX512][BuiltIn] Adding intrinsics of VGATHER{DPS|DPD} , VPGATHER{QD|QQ|DD|DQ} and VGATHERPF{0|1}{DPS|QPS|DPD|QPD} instruction set . Differential Revision: http://reviews.llvm.org/D19224 llvm-svn: 266983
2016-04-21 20:47:27 +08:00
#undef __DEFAULT_FN_ATTRS
#endif