[RC5] VIA C3
enojon at ATTGLOBAL.NET
Mon Sep 3 18:13:37 EDT 2001
at 600mhz, the VIA Cx III performs at 50% of the Intel and AMD equivalent
It may not be due to ooo exec alone, but VIA C3 lacking equivalent number of
parallel microcode units. Nevertheless, on microbenching, the core picked the
"SS Ath" for best performance albeit performing at far slower keys/sec rate.
Maybe L2 cache size of the C3?
----- Original Message -----
From: "Peter Cordes" <peter at llama.nslug.ns.ca>
To: <rc5 at lists.distributed.net>
Sent: Wednesday, August 29, 2001 7:23 PM
Subject: Re: [RC5] VIA C3
> On Wed, Aug 29, 2001 at 01:02:53AM +0800, Jin-Wei Tioh wrote:
> > Would the lack of out-of-order execution on the VIA C3 processor
> > significantly impact its RC5-64 performance?
> Probably not. Out of order execution gives you better performance
> with code that isn't perfectly tuned, and gives you more flexibility
> when tuning. For something like RC5, it's probably still possible to
> keep all the (applicable) execution units running almost all the time.
> Intel's P5 core was a superscalar in-order execution design. (U pipe
> and V pipe...) The RC5 core optimized for it kept both pipes full,
> > Would any other aspects of its design affect it?
> How many instructions per clock it can issue (how many pipes there
> are) is important, as well as the latency and throughput of the important
> instructions like rotate, and MMX operations.
> #define X(x,y) x##y
> Peter Cordes ; e-mail: X(peter at llama.nslug. , ns.ca)
> "The gods confound the man who first found out how to distinguish the hours!
> Confound him, too, who in this place set up a sundial, to cut and hack
> my day so wretchedly into small pieces!" -- Plautus, 200 BCE
> To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
> rc5-digest subscribers replace rc5 with rc5-digest
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5