Tony,

Check out SiSoft's Sandra for a full suite of system benchmarks.  Also if you
feel that you might have some faulty memory I'd suggest running a program
called "Prime95" for a few days.  The program is very memory as well as CPU
intensive because it tried to factor large numbers to find mersenne primes.
If your system can withstand the "torture test" feature then you know you
have a stable system as far as memory timing, CPU cooling, etc.  The torture
test works by factoring numbers of known primes and if the answers computed
doesn't correspond with the known solution then you know you have a memory
issue.  Its been used widely in the overclocking community (which I used to
be a part of for a few years).

Sandra - http://www.sisoftware.demon.co.uk/sandra/

Prime95 - http://www.mersenne.org/freesoft.htm

Hope that helps.  If you are really into memory performance you should get
out your BIOS manual (assuming you have a good BIOS) and learn about some of
the advanced memory features.  On my workstation at home I have some PC100
running at 133Mhz (Crucial RAM is top notch on the fab) and just by modifying
the CAS Latency, Bank interleaving, etc I was able to boost my memory
performance just under that of DDR266.  Not bad for a no cost upgrade.  Oh
yea... and my system will run Prime95 till the cows come home... this is
where it pays to get high quality RAM and not the cheap stuff.

http://www.crucial.com <-- arguably the best RAM on the market

matt

-----Original Message-----
From: Tony Karavidas [mailto:tony@;encoreelectronics.com]
Sent: Thursday, October 17, 2002 5:15 PM
To: 'Protel EDA Forum'
Subject: Re: [PEDA] [OT] memory testing and benchmarking


Phil,

You're the guy that mentioned the memtest the other day right? Could you
tell me/us what you get for a benchmark on your RDRAM? Do you remember
what you had for your SDRAM? I downloaded that app the other day when
you mentioned it, and got all sorts of wildly varying results.

I'd like to know the full benchmark (L1 cache, L2, etc)

Tony

> -----Original Message-----
> From: Phillip Stevens [mailto:pstevens@;capecod.net] 
> Sent: Thursday, October 17, 2002 1:49 PM
> To: Protel EDA Forum
> Subject: [PEDA] Re[2]: Protel99se and win2k fun etc etc
> 
> 
> 
> 
> TH> There is no direct scaling. Once you have enough memory to avoid 
> TH> thrashing then additional memory just avoids some disk 
> access with 
> TH> rapidly diminishing returns.
> 
>  I agree.
> 
>  I went from 256Megs to 512Megs just because I could do so
>  for (only) $45.00,  but saw no real substantial difference.
> 
>  When I was evaluating DXP I did see a big difference between 
>  autoroute times on my old PC,  vs the new system I put 
> together to  run DXP on.  The time to route my test board 
> dropped from over 6.5  hours to 57 Mins (WOW!).  There were a 
> lot of variables here (a faster CPU,  DXP vs 99SE,  512Megs 
> vs a GIG, ect.) so it is difficult to draw any  real conclusions.
> 
>  My guess though,  is that my biggest gains may have come 
> from the RDRAM  in the new system vs the SDRAM in the old 
> one.  (I'm sure the faster  CPU played some part too...)
> 
>  So,  I'm thinking there _might_ be a more worthwhile payoff 
> in changing  memory type (say SDRAM to DDR),  vs the amount 
> of RAM in a system,  assuming one is past that thrashing point.
> 
>  ---Phil
> 
> 

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
* To post a message: mailto:proteledaforum@;techservinc.com
*
* To leave this list visit:
* http://www.techservinc.com/protelusers/leave.html
*
* Contact the list manager:
* mailto:ForumAdministrator@;TechServInc.com
*
* Forum Guidelines Rules:
* http://www.techservinc.com/protelusers/forumrules.html
*
* Browse or Search previous postings:
* http://www.mail-archive.com/proteledaforum@;techservinc.com
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to