how long should stress tests be.and what is engine x

[greybing]

hi ,overclocking palit gtx 460,got 850 core,1690 shader and 1'100 memory.{.stable. overclock.}...load temperature in call of pripyat[game of choice for stress tests is averaging around 75c.....my stress tests have not been long most were about 15 mins long...some lasted longer.double..i have not touched voltage as i am a little reticent about this as i have heard it is more potentially harmful to your component if goes wrong.....anyway are my stress tests long enough......also on my final change to factory settings ,during the stress test the game stopped and a message came up saying engine x.1.10 [not sure of the exact term,but very similar,sorry]had stopped workimg,windows 7 was looking for a solution in a little message box but it appeared to just be stuck.....i am assuming i reached my limit ,one way or the other...........thanks

 

[Zen_]

Furmark is good but the only test I've found for GPU overclocks that is foolproof is actual gaming. For instance my 5770 is 100% stable all night @ 925/1250 in Furmark, but if I overclock at all BC2 has stuttering and eventually crashes.

 

[arroyo]

Zen has right. I've had a bad time overclocking my HD4850 to the limit. You think that after 8 hours of Furmark, 8 hours of Crysis and 8 hours of 3dMark your card is stable and than you run "Plants vs Zombies" and BSOD.
Just test your card at games, that you play. It there would be no BSOD's than everything is ok. Don't believe Furmark and other stuff. Do you enjoy playing Furmark? No? So why bother?

 

[Fatal]

Furmark is good but the only test I've found for GPU overclocks that is foolproof is actual gaming. For instance my 5770 is 100% stable all night @ 925/1250 in Furmark, but if I overclock at all BC2 has stuttering and eventually crashes.

I agree with you Zen gaming is the way I test as well. My MSI 4850 would run benchmarks then bam!! once I would try games on max the driver would fail. I would test with max setting while gaming.

 

[greybing]

thanks ...and yes,i agree ,games are a very good test and call of pripyat is the one i have been using for this particular overclock,my question was and is how long should i be doing these benchmark stress tests....or what would people say is a good general average per test.obviously if i am increasing each core individually in increments of 20 then stress testing a game for an hour then that is going to take a long time....i have been stress testing 15-20 mins and i have been overclocking each core individually ,increasing in increments of 20.....i have not touched voltage.......so to sum,what is the ideal average game stress test legnth

 

[sneekypeet]

*waves hand like a jedi* These are not the trolls you are looking for.

Carry on

 

[Thrackan]

If your stresstest stopped working suddenly, I wouldn't call your OC stable. And the exact error message will help alot as there is a lot of pages with "engine x" in it according to Google.

Stable is a variable term. Some people believe in 10+ hours, some days, some think an hour is enough and other believe that if it can start a game, it's stable. Use your own definition there.

 

[greybing]

ok ...thanks for your words,

 

[@RaXxaa@]

i would go for the time you would play at most 3 hours of gaming too much for you then use that as a target to get

 

[kenkickr]

I'm a fan of running 3 loops of 3dMark06 or Vantage to test a OC out or a game that has built-in benchmark capabilities.

 

[greybing]

thanks for your thoughts everyone