This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 5/N] [x86_64] Vectorized math functions


On Fri, 21 Nov 2014, Andrew Senkevich wrote:

> This is addition of new vector math library libmvec.so to Glibc build,
> install and ABI testing.
> Vector math library libmvec.so enabled by default for x86_64.

On the whole I think it would be best to separate the 
architecture-independent support for a new library, from the 
architecture-specific code to enable it.

> +# Rules for the test suite are in math directory

'.' at end of comment.

> diff --git a/sysdeps/x86_64/configure.ac b/sysdeps/x86_64/configure.ac
> index c9f9a51..349d6ca 100644
> --- a/sysdeps/x86_64/configure.ac
> +++ b/sysdeps/x86_64/configure.ac
> @@ -99,6 +99,11 @@ if test $libc_cv_cc_avx2 = yes; then
>  fi
>  LIBC_CONFIG_VAR([config-cflags-avx2], [$libc_cv_cc_avx2])
> 
> +dnl Set build_mathvec
> +if test x"$build_mathvec" = xnotset; then
> +build_mathvec=yes

There's no point in the comment here that just repeats the plain meaning 
of the code.

-- 
Joseph S. Myers
joseph@codesourcery.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]