2 overclocking questions

[Circaflex]

Ok my first question is how many of you guys use prime95 to determine stablility? Im a little iffy on it this weekend. A few overclocks i tried failed within 30 mins of prime but i can run every benchmark and loop 3dmark2001 and 2003 without any lockups and play my games fine but prime crashes. I can also do the sandra burnin with prioty 10 and run it 15 times without a lockup im not sure if these dont stress it enough or if prime is just harder? or buggy? I dont know.

Now my next question is deciding 5:4 or 1:1 I can run a fsb of 223 with pc3200 and 1:1 ratio with 2.5-3-3-7 timings giving me 3.12 or i can run a 235 fsb with 5:4 with 2-2-3-6 timings giving me 3.3 should i just use the 1:1 or get more mhz and go with 5:4? Thanks for any help you provide.

[KrisMCool]

Your situation sounds a lot like mine. I get errors in Prime95 (blend torture test only) when everything else works (except FAH). But most would probably agree that Prime95 is an excellent test of stability, even though bugs have to be fixed often in the program.

I have similar findings with ratios. Prime95 stable up to FSB 223 @ 1:1 and FSB 238 @ 5:4 with different sets of memory and CPUs. But I am trying to determine if I have a board problem, as others with similar systems don't have a problem going higher with either ratio.

[brakezone]

I think prime 95 is pretty reliable. If u can clock it down significantly and run it without errors than there must be a weak spot somewhere in ur configuration. Use ur best judgement, some people discard prime95 because they play games and are not folding or anything... most of the time I think people want a 100% stable system- the true test of 100% stability is for the machine to run something like prime 95 indefinitely without errors... so even 24 hours without errors is not 100% stability. Thats why processors are sold slower than overclockers run them, and amd or intel wouldn't sell a processor that they didn't think would be able to pass a lifetime of operation with no mistakes- or at least i'm sure that this is what they aim for because some people actually use these chips for important stuff!

[tio]

I use it to test for stability. It's the most stressfull program that I run, although I hear some people say there's a Folding @ Home test that is a bit more stressfull. In any case, I use it because I know that if my system isn't stable, it will let me know. I don't run anything that isn't 100% stable, but some people don't mind running a slightly unstable system. If the most stressfull thing you normally do is play games and you're not getting errors while playing them, then there's probably no reason to back down your overclock. However, when it gets hotter in the summer time, you may have to back down a bit. Just bear in mind that if you ever decide to join the Folding @ Home project or Seti or any of those distributed computing projects, you absolutely need to ensure your machine is 110% stable, as sending back erronious data is only going hurt the project in a big way! :eek:

[#18]

Prime95 is a very good program to test your system's stablility, but if all the programs you use are running fine then the system is stable for you. I've run into the situation where all the programs were stable but Prime95 will fail in a few seconds every time even at stock speed. Then just ignore it.

For your second questions, you can run some benchmarks on the two different settings you were talking about. See for yourself. Then pick the better one.

[batboy]

For a quick stability check, I've been using PCmark2004 lately. Not as good as Prime95, but it's pretty good and lots faster. If a new rig can pass PCmark2004, then I know it'll be stable in all my games and apps.

Yeah, benchmark that sucker at both ratios and see for yourself. I bet it's a close toss up. Probably the 1:1 will edge the 5:4 out a little.