Computer Chess Club Archives

Search

Terms

Messages

Subject: Re: SSE2 bit[64] * byte[64] dot product

Author: Gerd Isenberg

Date: 12:33:33 07/18/04


>I am guessing something like 50 cycles?  Really not that bad . . . probably
>close to the speed of a scan over attack tables.
>
>anthony

14.45ns on a 2.2GHz Athlon64, ~32 cycles now.

Some minor changes, byte vector values (weights) 0..63, therefore only one
psadbw, no movd but two pextrw, final add with gp. Computed bit masks in two
xmm-registers (0x02:0x01). Some better instruction scheduling.

Gerd

Re: SSE2 bit[64] * byte[64] dot product Tony Werten 23:36:05 07/19/04
Re: SSE2 bit[64] * byte[64] dot product Anthony Cozzie 07:59:05 07/19/04
- Re: SSE2 bit[64] * byte[64] dot product Russell Reagan 21:51:40 07/19/04
  - Re: SSE2 bit[64] * byte[64] dot product Tony Werten 23:25:03 07/19/04
    - Re: SSE2 bit[64] * byte[64] dot product Russell Reagan 00:05:47 07/20/04
      - Re: SSE2 bit[64] * byte[64] dot product Tony Werten 01:03:26 07/20/04
        
        Re: SSE2 bit[64] * byte[64] dot product Russell Reagan 02:05:21 07/20/04
- Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 10:21:23 07/19/04
  - Re: SSE2 bit[64] * byte[64] dot product Anthony Cozzie 08:54:44 07/20/04
    - Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 11:23:30 07/20/04
      - Re: SSE2 bit[64] * byte[64] dot product Anthony Cozzie 12:57:36 07/20/04
        
        Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 23:03:46 07/20/04
        
        Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 02:37:22 07/21/04
        
        Re: SSE2 bit[64] * byte[64] dot product Anthony Cozzie 06:32:22 07/21/04
        
        Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 05:38:48 07/22/04
        
        Re: SSE2 bit[64] * byte[64] dot product Daniel Clausen 05:49:32 07/22/04
        
        Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 08:07:04 07/22/04
        
        Re: SSE2 bit[64] * byte[64] dot product Fabien Letouzey 06:44:15 07/22/04
        
        Re: SSE2 bit[64] * byte[64] dot product Anthony Cozzie 06:56:50 07/22/04
        
        Re: SSE2 bit[64] * byte[64] dot product Fabien Letouzey 07:15:33 07/22/04
        
        Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 08:36:33 07/22/04
    - Re: SSE2 bit[64] * byte[64] dot product Anthony Cozzie 09:31:59 07/20/04
      - Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 10:54:12 07/21/04
        
        ignore previous post Gerd Isenberg 11:05:33 07/21/04
      - Re: SSE2 bit[64] * byte[64] dot product Anthony Cozzie 09:39:17 07/20/04
        
        Re: SSE2 bit[64] * byte[64] dot product Gerd Isenberg 02:16:00 07/21/04
        
        Re: SSE2 bit[64] * byte[64] dot product Anthony Cozzie 06:41:48 07/21/04
- Re: SSE2 bit[64] * byte[64] dot product Fabien Letouzey 08:35:10 07/19/04
  - Re: SSE2 bit[64] * byte[64] dot product Fabien Letouzey 08:52:29 07/19/04
    - Re: SSE2 bit[64] * byte[64] dot product Tony Werten 23:33:20 07/19/04

This page took 0.01 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.