Go Back   Hardware Canucks > PC BUILDERS & TWEAKERS CORNER > Overclocking, Tweaking and Benchmarking

    
Reply
 
LinkBack Thread Tools Display Modes
  #21 (permalink)  
Old June 16, 2011, 02:04 PM
Rookie
 
Join Date: Jun 2011
Location: USA, Maine
Posts: 3

My System Specs

Default

Just thought I would let you know it doesn't work on my AMD Phenom ii x4 955BE lol. That would be sweet if you did end up developing it for AMD based computers though. If it did the tests and was relatively accurate I would be willing to pay $20 for such a tool. I dont know if you plan on marketing it but that would be pretty sweet.
Reply With Quote
  #22 (permalink)  
Old June 17, 2011, 02:16 AM
Rookie
 
Join Date: Jun 2011
Posts: 9
Default

Thanks a lot for your interest!:)

But it is a research and my departement is interested not on money but on the results...:)

My program use specific instruction of i7 950 (but I had good results also with microprocessors of the same family,e.g. 920 or 930).

The goal of my thesis is creating a tool which focalizes on a specific microarchitecture. Unlike the other programs it takes few time (max 5 minutes) and it produces very little heat....
Reply With Quote
  #23 (permalink)  
Old June 17, 2011, 03:36 AM
Hall Of Fame
F@H
 
Join Date: Nov 2008
Location: Ottawa, ON
Posts: 1,226

My System Specs

Default

Too bad I don't have any Nehalem CPUs around here to try, but this seems interesting to me. That's a nice concept behind the test though - rather than trying to burn the CPU you target areas that are the must vulnerable to a frequency increase, if I understand correctly?

When you say "evolutionary approach", do you mean that you "grew" your program with genetic algorithms? If so, that sounds a lot like you could feed the generator for your program any instruction set and make (the beginnings of) a stress tester for it. It may be interesting if you can have results that way, if I understand correctly.

Have you also checked if your program can find errors faster if the CPU is at a different temperature during the test? For example, say your overclock is JUST BARELY unstable but the only instability is that it cannot run instruction X at a temperature greater than or equal to Y. If your program keeps the CPU a lot cooler than temperature Y, then you'll potentially miss the problem and generate a pass.
__________________
"The computer programmer says they should drive the car around the block and see if the tire fixes itself." [src]
Reply With Quote
  #24 (permalink)  
Old June 17, 2011, 04:42 AM
Rookie
 
Join Date: Jun 2011
Posts: 9
Default

Quote:
Originally Posted by frontier204 View Post
Too bad I don't have any Nehalem CPUs around here to try, but this seems interesting to me. That's a nice concept behind the test though - rather than trying to burn the CPU you target areas that are the must vulnerable to a frequency increase, if I understand correctly?

When you say "evolutionary approach", do you mean that you "grew" your program with genetic algorithms?
Yes! Your assumption is corrected!
The software has been developed through an evolutionary approach using a tool of my department which is based on genetic algorithms.
This tool is "ugp" and it is distributed under The GNU public license (MicroGP | Download MicroGP software for free at SourceForge.net)
In addition,if you want further information about it you can read this page
Research: MicroGP
or this White paper
http://www.cad.polito.it/downloads/W...onary_tool.pdf

Quote:
Have you also checked if your program can find errors faster if the CPU is at a different temperature during the test? For example....
the temperatures do not affect the test ... in fact the objective of the program is not producing heat to
test the CPU, but stressing it with precise calculations using special instructions.
Reply With Quote
Reply


Thread Tools
Display Modes

Similar Threads
Thread Thread Starter Forum Replies Last Post
HDD fails S.M.A.R.T. short test, but passes long test? frontier204 Storage 10 May 19, 2011 04:10 PM
System Stability Test PuB_Evo Overclocking, Tweaking and Benchmarking 9 January 12, 2010 02:01 PM
Programs that can test GPU stability? Silvgearx Video Cards 19 March 11, 2009 08:43 PM
OS Stability Question CORE O/S's, Drivers & General Software 11 January 4, 2009 11:04 PM
4850 Stability Oversized Rooster Video Cards 42 July 30, 2008 08:30 AM