SG Broadband Community

I've noticed that with v22 I seem to have very good results from my system ....


When I run v23 and it runs the large fft's ( 1024 specifically) It always seems to fail even with 1.80v


I was able to run at 1.70v stable for like 6-8 hours at 210 X 11.0 when I first built the system ( on v22 ) perfectly stable and 1024 is the first set it always runs on that version ....



For instance tonight ... I installed v23 and it fails after 2min ... ( on both 1.70v and 1.80v )

I uninstalled v23 and reinstalled v22 and it's been running for the past 30 minutes no hiccups on 1.70


I had read from several other people the exact same issue on the exact same fft setting sometime back ( when v23 first came out ) I was just wondering what some of your experiences might be. Is one version more trustworthy than the other ?

Fact:
Both versions were running at 500 MB memory usage with screen displays every 5000 iterations.

Also on a second note , I've never had a single issue whatsoever with stability in everyday usage , Gaming, Music , picture editing, DVD watching etc ...


So real world stability seems fine ....


I'm wondering ... Would folding work as a more reliable , more realistic stability tester ? Would it tell me if it found any errors in real time ?


I just want to know if I can push it back up to 2420 again ( It seemed to have lost some stability at that speed some time back. If I need to raise the CPU Fan speed then I'm going to have to place my PC in another location not quite so close to my sensitive ( and rather large ) ears.

I don't use Prime at all. I'll use memtest86 to test Ram. To test an OC I usually use future mark 2k3, folding and every day use. IMO Prime sucks.

So when you run those super high OC's of yours ( wasn't it you that his 3.6 with an AMD a few months ago? ) you don't use prime or anything other than folding to test your stability ?

Dam I wish I hit 3.6 on an AMD, that would be awesome.
Nah I just do what I said above, it works well. Prime isn't that great a gauge in my opinion.

Originally posted by cyPHer_138
......... Prime isn't that great a gauge in my opinion.

unless you want to make sure your system is doing calculations correctly, if it gives you errors, then it is not doing any good for F@H.

that is what i have gathered in my reading of the whole prime95 debate.

just my 2 cents worth

Originally posted by cyPHer_138
Dam I wish I hit 3.6 on an AMD, that would be awesome.
Nah I just do what I said above, it works well. Prime isn't that great a gauge in my opinion.

eerrr ummm I meant 2.6 lmao

i've noticed the same thing with prime95 v23

although the system is stable real world and other benchies, prime would require extra voltage...

haven't tried prime95 v22 but I suspect I would get the same results as you