linux-mips
[Top] [All Lists]

On Sat, 13 Nov 2004, macrohat wrote...

To: macrohat <emblinux@macrohat.com>
Subject: On Sat, 13 Nov 2004, macrohat wrote...
From: peter fuerst <pf@net.alphadv.de>
Date: Mon, 15 Nov 2004 23:43:40 +0100 (CET)
Cc: linux-mips <linux-mips@linux-mips.org>
In-reply-to: <20041113134735Z8224907-1751+1490@linux-mips.org>
Original-recipient: rfc822;linux-mips@linux-mips.org
Reply-to: pf@net.alphadv.de
Sender: linux-mips-bounce@linux-mips.org

Hello !

BogoMips is most useful as a benchmark, if the main purpose of your
machine is to calculate BogoMips' ... (see BogoMips Mini-HOWTO :))
However - since there seems to be such a strong desire to see large
BogoMips values - here is some help:

Your BogoMips factor of 0.66, instead of the usual 0.99..., indicates
that the delay loop is misaligned, i.e. there's a instruction-cache-block
boundary inmidst the loop. (Recently i managed somehow to achieve this on
a R10000 :)
A ".align 3\n\t" at the begin of __delay() will keep the branch and its
delay-slot together.

You should even be able to double the BogoMips value (factor 1.99...) by
unrolling the delay loop (at least on R10k):

  static __inline__ void
  __delay(unsigned long loops)
  {
      loops |= 1;
      __asm__ __volatile__ (
          ".align 4\n\t"  /* only the paranoid survive. */
          ".set\tnoreorder\n"
          "1:\n\t"
          "dsubu\t%0,1\n\t"
          "bnez\t%0,1b\n\t"
          "dsubu\t%0,1\n\t"
          ".set\treorder"
          :"=r" (loops)
            :"0" (loops));
  }

Nevertheless, despite all this trickery, your machine will run exactly
as fast (slow), as it did before.

with kind regards

pf



  "I have been a happy man ever since January 1, 1990, when I no longer
  had an email address. ..."

                                                        Donald E. Knuth
                  (http://www-cs-faculty.stanford.edu/~knuth/email.html)


On Sat, 13 Nov 2004, macrohat wrote:

> Date: Sat, 13 Nov 2004 21:47:02 +0800
> From: macrohat <emblinux@macrohat.com>
> To: linux-mips <linux-mips@linux-mips.org>
> Cc: linux-cvs <linux-cvs@linux-mips.org>
> 
> Hello linux-mips:
> 
> I have a question to ask you: why BCM1250 CPU Bogomips is so much lower than 
> CPU clock frequency,such as:
> CPU 700MHz - 465.30 Bogomips, CPU 800MHZ - 532.48 BogoMIPS.And i find out 
> that CPU Bogomips is a fixed value regardless L2 cache open or closed,
> 
> Enclosed is the log from the console
> 
> Regards!
>                               
> 
> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡macrohat
> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡emblinux@macrohat.com
> ¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡¡2004-11-13
> 



<Prev in Thread] Current Thread [Next in Thread>
  • [no subject], macrohat
    • On Sat, 13 Nov 2004, macrohat wrote..., peter fuerst <=