From: Andrew Beals (bandy_at_cinnamon.com)
Date: Fri Aug 09 2002 - 16:57:22 EDT
Exactly the reason why Apple/Motorola/IBM has been spanking Intel's behind on
benchmarks for years.
On Fri, 09 Aug 2002 17:04:28 -0400, "Allan J. Ballard" <aballard_at_ix.netcom.com>
wrote:
> If you could increase the cache memory up to 1 Mb the improvement would be
> amazing.
>
> I have done many benchmark tests fooling around with L1 memory and both OS2
> and Windoze starve for it, at least up to a meg.
>
> The newer systems may bes equipped to handle this problem but the older ones
> with 256K maybe less are not.
>
> ----- Original Message -----
> From: "Andrew Beals" <bandy_at_cinnamon.com>
> To: <AnimePhreak_at_aol.com>
> Cc: <thinkpad_at_cs.utk.edu>
> Sent: Friday, August 09, 2002 2:29 PM
> Subject: Re: old thinkpad horsepower increase
>
>
> >
> > AnimePhreak_at_aol.com said:
> > > It would be better if instead of you timing how long it takes a
> > > program to start then to run an actual benchmarkprogram. Like PCMark.
> >
> > If I had such a proggie, I'd have run it . As it was, I saw a real
> > discernable difference with Planet's Visibility and decided to document
> it.
> > If anything, it was a good example as to why having a math co-processor is
> a
> > Very Good Thing if you're going to be crunching lots of numbers.
> >
> >
> > ross_at_math.hawaii.edu said:
> > > Most likely this is a computation-bound program, so would probably
> > > respond very well to this change in CPU.
> >
> > And fewer things are more computation-bound than I had previously thought,
> or
> > they're just poorly optimized.
> >
> > Windoze seemed a little bit snappier, and may actually be, 133MHz versus
> > 33MHz, but it's definitely not four times faster.
> >
> > I do realize that I haven't changed the actual bus bandwidth, video
> > performance, or i/o.
> >
> > > I saw similarly nice results when moving a desktop to an AMD 586-133
> > > (see http://www.math.hawaii.edu/~ross/cpu.shtml), and the same is true
> > > for some TP701 owners here who changed the CPU to the AMD chip.
> >
> > Thanks for the pointer to your page.
> >
>
This archive was generated by hypermail 2.1.3 : Thu Jan 23 2003 - 09:59:13 EST