Patchwork x86_64: Utilize vector math functions w/o -fopenmp

login
register
mail settings
Submitter Andrew Senkevich
Date Dec. 7, 2015, 3:33 p.m.
Message ID <CAMXFM3uQXKHP5Q6hpuaHZ+RnNLUCgZYgbskqbKzh=ykLYNSVfQ@mail.gmail.com>
Download mbox | patch
Permalink /patch/9914/
State New
Headers show

Comments

Andrew Senkevich - Dec. 7, 2015, 3:33 p.m.
> This shorter patch is OK with spaces added between __attribute__ and '(',
> and between __simd__ and '('.

May be such mention in NEWS also needed for this update?



--
WBR,
Andrew
Joseph Myers - Dec. 7, 2015, 6:11 p.m.
On Mon, 7 Dec 2015, Andrew Senkevich wrote:

> > This shorter patch is OK with spaces added between __attribute__ and '(',
> > and between __simd__ and '('.
> 
> May be such mention in NEWS also needed for this update?

I'll leave NEWS entry review to the release manager.

Patch

diff --git a/NEWS b/NEWS
index 37189f7..602ea7a 100644
--- a/NEWS
+++ b/NEWS
@@ -60,6 +60,10 @@  Version 2.23
   C Library is GCC 4.7.  Older GCC versions, and non-GNU compilers, can
   still be used to compile programs using the GNU C Library.

+* Vector math library can be used with GCC 6.* w/o OpenMP SIMD constructs.
+  -ftree-loop-vectorize needed instead of -fopenmp in this case (visit
+  <https://sourceware.org/glibc/wiki/libmvec> for detailed information).
+
 Security related changes:

 * The nan, nanf and nanl functions no longer have unbounded stack usage