CPU stress

[Mookow]

Which distributed cracking program puts the most stress on a CPU/

 

[Robor]

I believe SETI@Home stresses one part of the CPU while rc5 stresses another. I'm sure one of those more informed can give you the details on that. If you really want to stress it, why not run both? Just make sure you set SETI@Home to run in screensaver mode (when idle) so rc5 gets some CPU time. I'm not positive but I'm pretty sure that if rc5 and SETI@Home are both running together SETI@Home will get most of the CPU cycles.

Rob

 

[Baldy18]

I'm no genious but I belive that RC5 stresses the CPU the most because it fits inside the cache of the CPU. SETI@Home puts the most stress on the memory and the memory bus.

 

[Assimilator1]

Hmm ,maybe but SETI will still use the cache intensively .
RC5 uses mainly the integer side of the cpu whilst SETI uses mainly the FPU

 

[Viztech]

I've read where Prime 95 is more stressful than RC5, OGR and Seti.

viz

 

[BurntKooshie]

The problem with RC5 is it stresses only the integer/load/store units. It does nothing with the FPU unit. S@H, on the other hand, stresses mostly the FPU part, a little integer, and stresses the cache and memory subsystem a lot. Folding at home seems to stress the exact same things as S@H, but uses less ram (less bandwidth too?).

Gamma Flux....That stresses the FPU part more than anything else, and I don't think it is bus limited (it should fit inside the L2 cache of modern CPU's). PRIME95 has been known as a very good check for more than stability

With PRIME 95, not only does is stress the CPU and test for "stability", it also tests for accuracy! Overclocked systems can produce errors, while maintaining stability (short term). PRIME 95 makse sure it's stable (from a CPU perspective), AND accurate.

[EDIT]dangnabit, I type too slow 😛

 

[Jeepin]

Cause errors? So you're saying it's quite possible that we've gone past an rc5 key already but because it was decoded wrong we're just going on? That would suck. But when I think about it, yeah, that can't be ruled out I guess.

 

[Polo]

On one system (c2 600@900) I run both RC5-64 and SETI. When it's too hot in the house, the system isn't fully stable anymore and SETI crashes while RC5 continue cracking keys with no probs. Just a hint ?...

 

[ViRGE]

Ya, a "bad" key is one of Dnet's fears. They have an FAQ asking that you allways keep your CPU 1 step below the max speed it can handle, so that its results aren't so boarderline.😱