Re: [PATCH] New generic cosf

["H.J. Lu" <hjl dot tools at gmail dot com>]

On Tue, Dec 5, 2017 at 12:45 PM, Paul Clarke <pc@us.ibm.com> wrote:
> From 60e12f38145e4699cf231f7815cf1894d5e2d68d Mon Sep 17 00:00:00 2001
> From: "Paul A. Clarke" <pc@us.ibm.com>
> Date: Tue, 5 Dec 2017 09:32:56 -0600
> Subject: [PATCH] New generic cosf
>
> The same logic used in s_cosf.S version for x86 and powerpc
> is used to create a generic s_cosf.c, so there is no performance
> improvement in x86_64 and powerpc64.
>
> -- 8< --
> For s390, this is the improvement noted.
>
> With patch:
>   "cosf": {
>    "": {
>     "duration": 1.00479e+10,
>     "iterations": 1.53856e+08,
>     "max": 900.645,
>     "min": 4.264,
>     "mean": 65.3074
>    }
>   }
> Without patch:
>   "cosf": {
>    "": {
>     "duration": 9.93841e+09,
>     "iterations": 4.63972e+08,
>     "max": 1010.9,
>     "min": 6.593,
>     "mean": 21.4203
>    }
>   }
>
> Tested on s390x, x86_64 and powerpc64le and powerpc32.
>
> I include below a diff with recent generic s_sinf.c, as it is more
> instructive than a diff with existing s_cosf.c.  There are a fair number
> of cosmetic changes, a few hard differences because it's a different
> computation, and adoption of changes from recent patches deemed acceptable.

1. Please use the current s_sinf.c.
2. Please use "diff -up".


-- 
H.J.