Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: 65.3ns :)

Author: Aaron Gordon

Date: 08:16:57 03/20/03

Go up one level in this thread


On March 20, 2003 at 10:48:21, Robert Hyatt wrote:

>On March 20, 2003 at 03:38:14, Aaron Gordon wrote:
>
>>On March 20, 2003 at 02:42:54, Matt Taylor wrote:
>>>The latency of a memory access on Clawhammer is at -least- 40 ns (the lowest
>>>figure I have seen reported). I saw another figure of 70-80 ns. I am not sure
>>>which one is correct. On my dual-Athlon, I get 133 ns per Sciencemark. Here are
>>>timings I obtained for an nForce 2 board:
>>>http://www.talkchess.com/forums/1/message.html?290002
>>>
>>>I do not doubt that better can be obtained. Aaron got as low as 68 ns with a 220
>>>MHz bus and fast ram.
>>
>>Just for the fun of it I retested, slightly faster fsb & timings. Here's what I
>>came up with. AthlonXP 2.46Ghz, 223fsb, 2.5-3-3-7 timings, single-channel DDR
>>
>>Memory latencies in nanoseconds - smaller is better
>>    (WARNING - may not be correct, check graphs)
>>------------------------------------------------------------------
>>Host                 OS   Mhz   L1 $   L2 $    Main mem    Guesses
>>--------- -------------   ---   ----   ----    --------    -------
>>p5        Linux 2.2.14-  2464 1.2200 8.1320        65.3
>>
>>
>>=)
>
>
>The next question:  will that clock speed pass a _strong_ memory diagnostic?
>
>:)

Yes, it's been running seti@home over night actually (very memory intensive). I
also ran 2 passes of "Memtest86" which is extremely sensitive. Luckily for me
all was perfectly stable. I checked it this morning.. still chugging along.

I'd really like to try to get below 60ns, even if it's just 58-59ns. This would
be quite an ordeal, though, as it would require me running
245-250fsb(490-500DDR). I'd need to put the peltier I have on the chipset and
bolt down a large heatsink/fan or perhaps liquid cooling ontop of that. That
alone wouldn't do it, either. I'd also have to push the VDD voltage up from
1.85v to 2.0-2.1v. Originally it was 1.58 and I pushed it up to 1.85v (still
within the Nforce2 specs). Any more though and it'd be going out of spec.



This page took 0 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.