mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Hardware

Reply
 
Thread Tools
Old 2007-01-10, 19:25   #12
Cruelty
 
Cruelty's Avatar
 
May 2005

22×11×37 Posts
Default

Integrated sound should not be a problem, at least it is not for me
Cruelty is offline   Reply With Quote
Old 2007-01-10, 20:23   #13
rx7350
 
rx7350's Avatar
 
Feb 2006
AR, US

9016 Posts
Default

I've got two Core2 Duos - an E6400 and an E6700 - and the only time I've ever seen a perfomance hit manifested in iteration time slowdown is when one core is LL testing and the other is doing stage 2, but I believe that has occurred on every platform type (PD805, PD915, and AMDs) I have, and was in the neighborhood of 4%. Once the stage 2 was complete, the iteration times dropped back down.
rx7350 is offline   Reply With Quote
Old 2007-01-10, 20:33   #14
rx7350
 
rx7350's Avatar
 
Feb 2006
AR, US

24×32 Posts
Default

However, I have noticed that running FSB and memory at something other than 1:1 does seem to increase iteration times. Check out Newegg for some bargain basement video cards. If you don't have a problem with 'open box' items, you can pick up some nice GPUs for $30, shipping included. If you don't want to buy a video card, then crank your resolution down as low as possible, with medium or low color. I've done this on systems where I don't have a video card, and using it just for a Prime95 node, and it does decrease iteration times. If you're using the pc for productivity tasks, then a very low resolution may be onerous to use.
rx7350 is offline   Reply With Quote
Old 2007-01-11, 14:17   #15
db597
 
db597's Avatar
 
Jan 2003

7·29 Posts
Default

Ok, update - I've played with the BIOS settings for the integrated graphics and changing DVMT to fixed and less memory has made no difference in iteration times.

Have not had the change to test a discrete graphics card yet. Maybe tonight.
db597 is offline   Reply With Quote
Old 2007-04-26, 15:41   #16
MatWur-S530113
 
MatWur-S530113's Avatar
 
Apr 2007
Spessart/Germany

A216 Posts
Default

I made the same experience with my Core 2 Duo E6600
Running a LLT with 2M FFT (exponent size: 36M) with each CPU:

only 1 core running: Per iteration time 49 ms
both cores running: Per Iteration time 54 ms

I got this results while looking for the CPU temerature with 'Core Center', a proggy I got with my MSI Mainboard:
starting Windows, no Prime95 is running, Windows Taskmanager shows both CPUs @ 0%-1%: after 1 h temperature: 48°C
starting 1 prime95, Taskmanager shows one CPU @ 100%, other CPU still @ 0-1%: after 1 h temperature: 58°C Iteration time: 49 ms
starting second Prime95, TM shows 100% for both CPUs: after 1 h temperature 69°C Iteration time: 54 ms

The temperature seems ok to me, but I still don't understand the loss of Perfomance while running both CPUs. 1 core: 100% -> 2 cores: 185% ??



mfg

S530113
MatWur-S530113 is offline   Reply With Quote
Old 2007-04-26, 18:43   #17
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

116208 Posts
Default

There are two standard explanations for production not scaling linearly: memory access and the shared-cache design of a Core2. It is possible that one instance of Prime95 uses more than 50% of available memory bandwidth, which would not leave enough bandwidth for two instances to run at full speed. To test for this, try two double-checks instead of two first-time tests; you should find better scaling than 185%. I run LLR (a more general Prime95) on exponents around 1M, where memory is not a bottleneck, and suffer no performance hit with two instances.

Second, much less likely from what I grasp of Prime95: The Core2 shares the cache among both cores, allowing one core to use the entire L2 if the other core is idle. When both cores run full blast, less cache is available to each. I don't think this has a great effect at your exponent size, but it might explain a few percentage points.
-Curtis
VBCurtis is online now   Reply With Quote
Old 2007-04-26, 19:28   #18
MatWur-S530113
 
MatWur-S530113's Avatar
 
Apr 2007
Spessart/Germany

2·34 Posts
Default

Thx for your fast reply. Both possibilities you talked about seems to have some logic... I don't think that I can change the memory-bandwith somehow, but maybe I can upgrade Cache like RAM ? (I'm a hardware noob...). I will make some doublechecks next month to see what happens on smaller exponents, that's a good idea . If I have more results I will post them.

mfg

S530113
MatWur-S530113 is offline   Reply With Quote
Old 2007-04-26, 21:48   #19
dsouza123
 
dsouza123's Avatar
 
Sep 2002

12268 Posts
Default

Cache is a feature of the CPU, so it isn't separately upgradeable,
only the upgrade to a new CPU will get you more cache.
dsouza123 is offline   Reply With Quote
Old 2007-06-07, 05:38   #20
popandbob
 
May 2007

10112 Posts
Default

Quote:
Originally Posted by MatWur-S530113 View Post
I made the same experience with my Core 2 Duo E6600
Running a LLT with 2M FFT (exponent size: 36M) with each CPU:

only 1 core running: Per iteration time 49 ms
both cores running: Per Iteration time 54 ms

I got this results while looking for the CPU temerature with 'Core Center', a proggy I got with my MSI Mainboard:
starting Windows, no Prime95 is running, Windows Taskmanager shows both CPUs @ 0%-1%: after 1 h temperature: 48°C
starting 1 prime95, Taskmanager shows one CPU @ 100%, other CPU still @ 0-1%: after 1 h temperature: 58°C Iteration time: 49 ms
starting second Prime95, TM shows 100% for both CPUs: after 1 h temperature 69°C Iteration time: 54 ms

The temperature seems ok to me, but I still don't understand the loss of Perfomance while running both CPUs. 1 core: 100% -> 2 cores: 185% ??



mfg

S530113
You might want to double check the max temp for CPU, My E6400 cpu max temp is 65C. If that is max temp then take a look at your cooling, your cpu could be throttling back to lower heat levels.

BoB
popandbob is offline   Reply With Quote
Old 2007-06-07, 18:49   #21
MatWur-S530113
 
MatWur-S530113's Avatar
 
Apr 2007
Spessart/Germany

2·34 Posts
Default

Quote:
Originally Posted by popandbob View Post
You might want to double check the max temp for CPU, My E6400 cpu max temp is 65C. If that is max temp then take a look at your cooling, your cpu could be throttling back to lower heat levels.

BoB
Thx for the hint. In the meantime I got lower temperature by opening one side of my midi-tower. Now my CPU temperature is about 58°C running both cores (depends on room temperature). Next time I buy a computer I will spent some more Euros for a better chassis and coolers...

Matthias
MatWur-S530113 is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
running two copies of prime95 in the same machine ppo Information & Answers 25 2007-07-30 23:25
Dual core: three copies or just two? JuanTutors Software 10 2007-04-13 01:43
Importance of dual channel memory for dual core processors patrik Hardware 3 2007-01-07 09:26
Running multiple copies of mprime on Linux hc_grove Software 3 2004-10-10 15:34
Anyone have hourly copies of summary.txt? PrimeCruncher Data 0 2004-05-15 23:32

All times are UTC. The time now is 17:46.


Tue Oct 26 17:46:27 UTC 2021 up 95 days, 12:15, 0 users, load averages: 2.35, 2.40, 2.23

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.