This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Improve integer division on IA-64
- From: Steven Bosscher <stevenb at suse dot de>
- To: gcc-patches at gcc dot gnu dot org
- Cc: Roger Sayle <roger at eyesopen dot com>
- Date: Tue, 20 Dec 2005 07:37:13 +0100
- Subject: Re: [PATCH] Improve integer division on IA-64
- References: <Pine.LNX.4.44.0512191637050.7457-100000@www.eyesopen.com>
On Tuesday 20 December 2005 01:19, Roger Sayle wrote:
> On an Itanium2 SGI Prism system, with a model 2 revision 1 processor
> running at 1.5GHz, the timings for 10^9 calls to __div?i3 are given
> below:
>
> before after before after
> operands __divsi3 __divsi3 __divdi3 __divdi3
>
> 81/1 22.69s 8.01s 27.36s 7.34s
> 81/2 22.69s 8.01s 27.36s 7.34s
> 81/3 22.69s 22.69s 27.36s 27.36s
> 81/81 22.69s 4.67s 27.36s 4.00s
> 81/100 22.69s 7.34s 27.36s 6.67s
>
Just curious -- how does this compare to the two inline integer divide
algorithms we have in ia64.md (-minline-int-divide-min-latency)?
Gr.
Steven