VirtualHogger Posted June 10, 2009 Posted June 10, 2009 I'm a newbie at overclocking and have this question about stress testing. My mb (Gigabyte EP45-UD3P) allows me to save up to eight BIOS profiles to a flash drive. I have been bumping up my FSB a little at a time to get about a 200MHz increase in CPU speed, then I stress test. After it passes MemTest86+ for a couple of passes and runs Prime95 for 24 hours, then I save the BIOS to my flash drive. If I go back to a previous profile, is there a need to run the tests on it again? I haven't run any benchmarks yet, looking back, it would have been interesting to see the performance improvement, if any. However, I would really hate to spend a lot of time for stability re-testing at each level just to see benchmark results. Do you think once it is stable at a certain setting it should be good to go again? Also, should MemTest be run every time I bump the FSB up, or am I wasting my time doing that? At stock settings, MemTest86+ ran for 24 hours without any errors. BTW, the machine I'm currently working on is the #1 in my sig. E5200 and it is a couple of hours into Prime95 as I type @ 3750MHz, 53C max temp so far, idles at 37C, all stock voltages. Looks like I should be able to get to 4GHz (fingers crossed). Quote Share this post Link to post Share on other sites More sharing options...
Crazy_Nate Posted June 11, 2009 Posted June 11, 2009 Hi Ted, You shouldn't have any trouble with a saved BIOS configuration, if it's been tested. However, as overclocking eventually takes the toll on a processor, it may require slightly higher voltages in the [far] future. I used my motherboard's saved BIOS configurations to have a stock profile and an overclocked profile for my Opteron 144. Worked great If you're looking to overclock by a large amount, you can run certain configurations to see how far your memory will go (and what settings are needed), without overclocking your processor too much. Drop your CPU multiplier and crank up the FSB. That should give you an idea of what works when, and then you can figure out what sort of dividers you'll be using at the clock speeds you're looking at (as well as voltages and timings). For intermediate steps in your overclocking process, I wouldn't worry about doing a full 24 hour Prime95, or even an 8. You could probably get away with running one of the OCCT tests for an hour or two and then move on. It's really up to you to determine what computer performance you're looking for. Synthetic benchmarks may give you an indication of performance increase, but, it may take looking at a game that you play to really get a feeling for how much more your computer can do... That's a pretty impressive overclock for stock volts Quote Share this post Link to post Share on other sites More sharing options...
VirtualHogger Posted June 12, 2009 Posted June 12, 2009 Thanks for the info. I kinda figured that once a configuration was tested, it should be stable again if reloaded, but I've seen stranger things. This will help me dial in my Q9400 a little faster than the E5200 I'm working on now. I got the Q9400 running at 3.2GHz now but want to see how far I can get it to go. The E5200 @ 3.75GHz failed the Prime95 torture test 22 hours in on core#2, so I bumped the VCore up a notch and retested and got 24 hours without an error. This will probably be my 24/7 config. I have it running at the moment at 4.0GHz with VCore maxed at Intel spec of 1.3625v, but core #2 failed Prime95 one minute in. I am going to see if I can get it to run at 3.9GHz before I quit on it. Again, thanks for the info! Quote Share this post Link to post Share on other sites More sharing options...
hardnrg Posted June 12, 2009 Posted June 12, 2009 The only two situations I can think of where a previous overclock would fail on the unique system configuration are: 1) you didn't test 3D apps and games - sometimes, if the RAM is on the bleeding edge, I've found that some games and benchmarks are an even better test than Prime95 Blend / Memtest - additional heat from graphics cards pushes CPU/RAM/NB/SB temps up into instability 2) your PSU sucks - additional power draw from the system under 3D, heavy hard drive load (if you have a lot of hard drives) causes the voltages to dip or fluctuate, causing errors. Something to consider: I've had P95 fail after 57 hours before... you might want to test longer than 24 hours for the very last time you run it... especially if it failed at 22 hours before Quote Share this post Link to post Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.