mersenneforum.org  

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

Reply
 
Thread Tools
Old 2013-06-17, 21:12   #12
kracker
ἀβουλία
 
kracker's Avatar
 
"Mr. Meeseeks"
Jan 2012
California, USA

32·241 Posts
Default

Ok, I totally lost it. You said "thoroughly stress test both cores" but you're trying to limit P95 to one core? And there's nothing wrong with running the same test on more than one worker.
kracker is offline   Reply With Quote
Old 2013-06-17, 21:23   #13
Caribou007
 
Jun 2013

10002 Posts
Default

Quote:
Originally Posted by kracker View Post
Ok, I totally lost it. You said "thoroughly stress test both cores" but you're trying to limit P95 to one core?
Yes.

Quote:
Originally Posted by kracker View Post
And there's nothing wrong with running the same test on more than one worker.
Sure there isn't. But is it better to run one Blend test alongside one Small FFTs test, or two Small FFTs tests?

Last fiddled with by Caribou007 on 2013-06-17 at 21:26
Caribou007 is offline   Reply With Quote
Old 2013-06-17, 21:37   #14
sdbardwick
 
sdbardwick's Avatar
 
Aug 2002
North San Diego County

2·11·31 Posts
Default

I don't think there will be any practical benefit to running two different types of torture tests simuntaneously.

However, if that is what one wants to do, caribou is right in that two instances of Prime95 (say prime95.exe and prime95_1.exe [renamed copy to identify in task manager] preferably from separate directories) and manual fiddling with affinity in task manager is required.
Tell each instance to run 1 worker window of whatever type and then set affinity via task manager.
sdbardwick is online now   Reply With Quote
Old 2013-06-17, 21:59   #15
Caribou007
 
Jun 2013

816 Posts
Default

Quote:
Originally Posted by sdbardwick View Post
I don't think there will be any practical benefit to running two different types of torture tests simuntaneously.
The guide I was reading said it was in order to utilize the CPU-RAM interface, which would make the CPU run hotter.

"The in-place large FFTs test is the hottest of the three ready-made tests just like it says in Prime95's "Run a Torture Test" window. And since it runs large FFTs, it's also doing a lot of memory accesses so it tests your CPU/RAM interface. It runs hotter than the small FFTs test because that test isn't accessing RAM very much so it's not heating up the CPU/RAM interface and that part of the CPU consumes quite a bit of power."

I think this may be useful because real-world applications also use the CPU-RAM interface. But I'm not sure, because they also use the video card, and other components which also generate heat. And overclocking, too, will be increasing the heat. So I'm not sure if the best test for the highest viable overclock is one that only utilizes the CPU and no memory and relies on the overclock itself to generate all the heat, or if it should be a test that also uses the CPU-RAM interface.

It's odd that the guide suggests a Blend/Small FFTs combination to stress test a CPU, because it also says, "The blend test tends to run a bit cooler than the in-place large FFTs test probably because the blend test is sitting and waiting for the memory accesses longer than the in-place large FFT test. So if you just want to run a single test, then run the in-place large FFT. It heats your CPU the most and it also tests your CPU/RAM interface when the CPU is at its hottest. It's not a thorough RAM test but then neither is the blend test and blend doesn't stress the CPU as much as in-place large FFT."

Last fiddled with by Caribou007 on 2013-06-17 at 22:05
Caribou007 is offline   Reply With Quote
Old 2013-06-18, 00:15   #16
TheMawn
 
TheMawn's Avatar
 
May 2013
East. Always East.

11×157 Posts
Default

I don't think I understand either.

Small FFT's are considered small because their data fits inside the CPU's cache which makes RAM look slow. The CPU NEVER has to wait for data and gets to work as fast as it pleases. This is the reason it generates more heat.

Large FFT's reside in the RAM so the CPU has to sit around waiting for data from the RAM every once in a while. This is the reason it says lots of RAM tested.


I personally can't see the benefit in running different kinds of torture tests at the same time. I think blend should suffice.


N.B. This idea of not wanting the RAM or GPU to compromise the test is a bit strange to me. If you wanted to underclock the RAM and GPU, I suppose you could, but it would probably be easier for you to run a GPU stress test and a RAM test (memtest86+ is excellent for the latter). I REALLY can't imagine the GPU compromising prime95... The memory can and probably will, but memtest86+ will find a lot of errors in a single pass when the computer could run fine for hours.

To be honest, if you want help, I think you should stop trying to hide your intentions from us. I'm sensing some tension due to the advice we're providing not being what you're looking for, but I can't say I really know what you want from us in the first place. It feels a bit like "grrr just answer my questions and let me deal with the way I want to stress test," to me anyway.

Last fiddled with by TheMawn on 2013-06-18 at 00:17
TheMawn is offline   Reply With Quote
Old 2013-06-18, 00:32   #17
Batalov
 
Batalov's Avatar
 
"Serge"
Mar 2008
Phi(4,2^7658614+1)/2

32·17·61 Posts
Default

Quote:
Originally Posted by Caribou007 View Post
I have a dual core, socket 939 CPU, the AMD Opteron 180.

Can someone post instructions on how to thoroughly stress test both cores please?
So, help us understand something here, too. Your goal is to overclock (this is what you put in the title) a 7-year old 90 nm technology CPU that spends more in electric cost over a small period of time than a whole shiny new system that will use less electricity and will be 10x more productive even before overclock?

You are pulling our legs, aren't you?
Attached Thumbnails
Click image for larger version

Name:	O180.png
Views:	101
Size:	25.6 KB
ID:	9904  

Last fiddled with by Batalov on 2013-06-18 at 00:45
Batalov is offline   Reply With Quote
Old 2013-06-18, 02:24   #18
TheMawn
 
TheMawn's Avatar
 
May 2013
East. Always East.

6BF16 Posts
Default

Quote:
Originally Posted by Batalov View Post
So, help us understand something here, too. Your goal is to overclock (this is what you put in the title) a 7-year old 90 nm technology CPU that spends more in electric cost over a small period of time than a whole shiny new system that will use less electricity and will be 10x more productive even before overclock?
Holy crap a 110W TDP no less. I didn't realize it was that old. I thought it was a younger processor that was just smaller à la Intel Atom.

I agree with Batalov on this. Just throw your arms into the air and give up on this entire chip. My overclocked Ivy Bridge uses 70% of the power that thing will, is clocked twice as high, has twice as many cores and at least four times the computing power clock per clock.

That give me 23 times more computing per Watt if I did the math correctly...
TheMawn is offline   Reply With Quote
Old 2013-06-18, 19:57   #19
VBCurtis
 
VBCurtis's Avatar
 
"Curtis"
Feb 2005
Riverside, CA

2×2,339 Posts
Default

Quote:
Originally Posted by TheMawn View Post
Holy crap a 110W TDP no less. I didn't realize it was that old. I thought it was a younger processor that was just smaller à la Intel Atom.

I agree with Batalov on this. Just throw your arms into the air and give up on this entire chip. My overclocked Ivy Bridge uses 70% of the power that thing will, is clocked twice as high, has twice as many cores and at least four times the computing power clock per clock.

That give me 23 times more computing per Watt if I did the math correctly...
The OP never suggested he'd be doing Prime searching, or any other 24/7 CPU-using, endeavor with this system. Seems to me he just wants to figure out the highest OC setting possible in an attempt to make an old system usable in a day-to-day context.
OP- Run two copies of Prime95, period. It doesn't matter if both copies keep switching from core to core, or are locked to one core. Rather than agonize over which test is the "best", run each of the combos you've considered. That best ensures stability in a range of load conditions if you do plan to use this machine for regular computing (as opposed to fulltime crunching- that's where the "massive waste of power, don't bother" comes in).
-Curtis
VBCurtis is online now   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
Can non-K Intel CPUs overclock memory? Prime95 Hardware 5 2015-12-31 22:09
Guide OmbooHankvald Operation Billion Digits 4 2009-08-26 08:18
Guide OmbooHankvald 3*2^n-1 Search 3 2005-07-13 21:03
Why can't I overclock??! db597 Hardware 11 2005-07-05 01:51
Can people overclock caches? clowns789 Hardware 3 2003-07-29 20:27

All times are UTC. The time now is 03:26.

Sun Mar 7 03:26:56 UTC 2021 up 93 days, 23:38, 0 users, load averages: 1.39, 1.44, 1.41

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.