* [PATCH v0] PPC64: Remove vec_finite_alias.c as aliases it creates are no longer required.
@ 2020-02-14 16:12 GT
2020-02-14 16:14 ` GT
0 siblings, 1 reply; 3+ messages in thread
From: GT @ 2020-02-14 16:12 UTC (permalink / raw)
To: libc-alpha\@sourceware.org, tnggil@protonmail.com
[-- Attachment #1: Type: text/plain, Size: 346 bytes --]
This patch is in response to
<https://sourceware.org/ml/libc-alpha/2020-02/msg00625.html>
With this patch, branch tuliom/libmvec has all changes needed to
enable GCC to autovectorize libmvec functions. Other than sincos.
Getting GCC to vectorize sincos is work-in-progress; presently GCC
does NOT vectorize for any architecture.
Bert.
[-- Attachment #2: 0001-PPC64-Remove-vec_finite_alias.c-as-aliases-it-create.patch --]
[-- Type: application/octet-stream, Size: 3486 bytes --]
From a6b9405b03d3e553757ae8fd0588dbe037c12240 Mon Sep 17 00:00:00 2001
From: Bert Tenjy <bert.tenjy@gmail.com>
Date: Fri, 14 Feb 2020 09:49:27 -0600
Subject: [PATCH v0] PPC64: Remove vec_finite_alias.c as aliases it creates are
no longer required.
File bits/math-finite.h has been removed; consequently aliases for SIMD log/f,
exp/f, and pow/f can also be eliminated.
---
sysdeps/powerpc/powerpc64/fpu/Makefile | 8 ---
.../powerpc/powerpc64/fpu/vec_finite_alias.c | 65 -------------------
2 files changed, 73 deletions(-)
delete mode 100644 sysdeps/powerpc/powerpc64/fpu/vec_finite_alias.c
diff --git a/sysdeps/powerpc/powerpc64/fpu/Makefile b/sysdeps/powerpc/powerpc64/fpu/Makefile
index a52a9c129a..05075c2a75 100644
--- a/sysdeps/powerpc/powerpc64/fpu/Makefile
+++ b/sysdeps/powerpc/powerpc64/fpu/Makefile
@@ -5,11 +5,3 @@ CFLAGS-s_llrint.c += -fno-builtin-lrintf -fno-builtin-llrintf
# Same as before but for lroundf and llroundf
CFLAGS-s_llround.c += -fno-builtin-lroundf -fno-builtin-llroundf
endif
-
-ifeq ($(subdir),mathvec)
-libmvec-support += vec_finite_alias
-
-CFLAGS-vec_finite_alias.c += -mabi=altivec -maltivec -mvsx
-
-libmvec-static-only-routines = vec_finite_alias
-endif
diff --git a/sysdeps/powerpc/powerpc64/fpu/vec_finite_alias.c b/sysdeps/powerpc/powerpc64/fpu/vec_finite_alias.c
deleted file mode 100644
index bbf5657b3b..0000000000
--- a/sysdeps/powerpc/powerpc64/fpu/vec_finite_alias.c
+++ /dev/null
@@ -1,65 +0,0 @@
-/* These aliases added as workaround to exclude unnecessary symbol
- aliases in libmvec.so while compiler creates the vector names
- based on scalar asm name. Corresponding discussion is at
- <https://gcc.gnu.org/ml/gcc/2015-06/msg00173.html>.
- Copyright (C) 2019 Free Software Foundation, Inc.
- This file is part of the GNU C Library.
-
- The GNU C Library is free software; you can redistribute it and/or
- modify it under the terms of the GNU Lesser General Public
- License as published by the Free Software Foundation; either
- version 2.1 of the License, or (at your option) any later version.
-
- The GNU C Library is distributed in the hope that it will be useful,
- but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- Lesser General Public License for more details.
-
- You should have received a copy of the GNU Lesser General Public
- License along with the GNU C Library; if not, see
- <http://www.gnu.org/licenses/>. */
-
-#include <altivec.h>
-
-extern vector double _ZGVbN2v_log (vector double);
-extern vector float _ZGVbN4v_logf (vector float);
-extern vector double _ZGVbN2v_exp (vector double);
-extern vector float _ZGVbN4v_expf (vector float);
-extern vector double _ZGVbN2vv_pow (vector double, vector double);
-extern vector float _ZGVbN4vv_powf (vector float, vector float);
-
-vector double
-_ZGVbN2v___log_finite (vector double x)
-{
- return _ZGVbN2v_log (x);
-}
-
-vector float
-_ZGVbN4v___logf_finite (vector float x)
-{
- return _ZGVbN4v_logf (x);
-}
-
-vector double
-_ZGVbN2v___exp_finite (vector double x)
-{
- return _ZGVbN2v_exp (x);
-}
-
-vector float
-_ZGVbN4v___expf_finite (vector float x)
-{
- return _ZGVbN4v_expf (x);
-}
-
-vector double
-_ZGVbN2vv___pow_finite (vector double x, vector double y)
-{
- return _ZGVbN2vv_pow (x, y);
-}
-
-vector float
-_ZGVbN4vv___powf_finite (vector float x, vector float y)
-{
- return _ZGVbN4vv_powf (x, y);
-}
--
2.20.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH v0] PPC64: Remove vec_finite_alias.c as aliases it creates are no longer required.
2020-02-14 16:12 [PATCH v0] PPC64: Remove vec_finite_alias.c as aliases it creates are no longer required GT
@ 2020-02-14 16:14 ` GT
2020-02-20 19:37 ` Tulio Magno Quites Machado Filho
0 siblings, 1 reply; 3+ messages in thread
From: GT @ 2020-02-14 16:14 UTC (permalink / raw)
To: libc-alpha\@sourceware.org, tnggil@protonmail.com
‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Friday, February 14, 2020 11:12 AM, GT <tnggil@protonmail.com> wrote:
> This patch is in response to
> https://sourceware.org/ml/libc-alpha/2020-02/msg00625.html
>
> With this patch, branch tuliom/libmvec has all changes needed to
> enable GCC to autovectorize libmvec functions. Other than sincos.
> Getting GCC to vectorize sincos is work-in-progress; presently GCC
> does NOT vectorize for any architecture.
>
Last line above should be:
[P]resently GCC does NOT vectorize SINCOS for any architecture.
Bert.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v0] PPC64: Remove vec_finite_alias.c as aliases it creates are no longer required.
2020-02-14 16:14 ` GT
@ 2020-02-20 19:37 ` Tulio Magno Quites Machado Filho
0 siblings, 0 replies; 3+ messages in thread
From: Tulio Magno Quites Machado Filho @ 2020-02-20 19:37 UTC (permalink / raw)
To: GT, libc-alpha
GT <tnggil@protonmail.com> writes:
> On Friday, February 14, 2020 11:12 AM, GT <tnggil@protonmail.com> wrote:
>
>> This patch is in response to
>> https://sourceware.org/ml/libc-alpha/2020-02/msg00625.html
>>
>> With this patch, branch tuliom/libmvec has all changes needed to
>> enable GCC to autovectorize libmvec functions. Other than sincos.
>> Getting GCC to vectorize sincos is work-in-progress; presently GCC
>> does NOT vectorize for any architecture.
>>
> Last line above should be:
> [P]resently GCC does NOT vectorize SINCOS for any architecture.
There was a conflict between this patch and my previous merge.
This patch has been merged into tuliom/libmvec.
Thanks!
--
Tulio Magno
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-02-20 19:38 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-02-14 16:12 [PATCH v0] PPC64: Remove vec_finite_alias.c as aliases it creates are no longer required GT
2020-02-14 16:14 ` GT
2020-02-20 19:37 ` Tulio Magno Quites Machado Filho
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).