Post Your Overclock!

[Lordbob75]

I would argue for longer testing times for a few reasons:
1) It is a tougher test, and therefor more likely to show errors
2) You know it will always be stable no matter what you do (mostly anyways) and if you suddenly DO need to run some encoding (or whatever) for 10 hours, you won't need to worry.
3) Running it for 10 minutes to an hour may not show errors you may get running a game for an hour or encoding for an hour.

Again, it is more opinion than anything else, but I see no reason NOT to run it for 12 hours as a check.

~Lordbob

 

[sygnus21]

To clearify my position...

I understand the need to run 12 hour stress test on an overclocked. However I don't think everyday usage will come near to what a program like Prime 95 does with your system, even in half the time.

In short, using everyday apps won't stress your system like running Prime 95 for 12+ hours; On the other hand, running a program like Prime 95 for 12+ hours and passing it says you have a very stable system, especially if it's overclocked.

I've said it a few times, it's one thing to say you're overclocked, it's another to run stable on a Prime 95 load with a temp monitoring program!

I can OC to 4gig... can I run it under load within temps! A program like Prime 95 answers that question

 

[Lordbob75]

I understand the need to run 12 hour stress test on an overclocked. However I don't think everyday usage will come near to what a program like Prime 95 does with your system, even in half the time.

In short, using everyday apps won't stress your system like running Prime 95 for 12+ hours; On the other hand, running a program like Prime 95 for 12+ hours and passing it says you have a very stable system, especially if it's overclocked.

I've said it a few times, it's one thing to say you're overclocked, it's another to run stable on a Prime 95 load with a temp monitoring program.

That is exactly what I was thinking.

~Lordbob

 

[sygnus21]

We both agree there

It kills me when people say I'm OC to 4gig and then show a shot of a system with no load :huh:

My .02 cents

 

[madtownidiot]

Makes sense to me.. if hardware is going to fail, it's usually under a heavy load.

I still think its much less boring to test your OC while playing a game (or letting your kids play) at the highest possible settings for several days or even a full week between adjustments. Some BSODs only happen under constantly changing conditions

 

[sygnus21]

You have a point there as well, but running a stress test program like prime 95 for a constant 12+ hours is torture on a system, especailly with regards to heat.

While games can be taxing, their not as taxing as Prime 95 is overall.

Also, prime 95 can and will fail a system that's not overclocked!

 

[madtownidiot]

I'm familiar with prime95... the last time I used it, it fried my PSU, even though the CPU temps never went above 65c

 

[sygnus21]

Hmmm....sounds like it stressed the system, which in turn shot the PS. Something to be aware of and watchful of... if the system seems to be running low on temps, check that one of the workers didn't stop due to an error. If that happens, that basically says there's a problem with your system, and if you continue to run it in that state, well...

Whenever I run Prime 95, I keep an eye on the loads and temps. If I see a load of less than 100 percent and prime is running, I know a worker failed and it's time to shut down prime 95.

 

[Dave76]

Agree with this scenario, the question is how long do you need to run Prime95 or equivalent.

For some the 12 hour test period is just fine, for others a shorter test may be adequate.

Getting a 4 GHz OC, or any OC, and not stress testing it, is just not complete. It needs to be tested for stability and temps.

 

[madtownidiot]

I guess that was the weak link in my system. To be fair the PSU was probably defective.. never seen smoke coming from a PSU without tripping a circuit breaker almost immediately, but then again if you've never smelled fried silicon or melted wires..

 

[hatehereyes]

Not bad, I guess. Still need to get another fan to go push-pull with my H50

I've got the same identical overclock as you and everything, about the same temps also with my Cooler Master V8

My question is, what are you ambient temps? The past few days it's been super hot here in Socal and I'm only about an hour and a half north of you in Riverside.

 

[dgwin7]

Here is my i5 750 overclocked to 4.1 pix below.

dg

 

[kurahk7]

Here is my i5 750 overclocked to 4.1 pix below.

dg

Wow nice, what cooling do you have?

 

[madtownidiot]

Not bad at all.. probably a lot faster than my system

 

[chev65]

In my opinion putting you're cpu under stress for 12 hours is just retarded .
When you overclock you're gpu would you leave FurMark running for 12 hours aswell?

Whether you agree or not, that is the standard practice/procedure for stress testing an overclocked computer

Perhaps you should visit some overclcking sites for tips on stress testing overclocked systems? You would see some recommend more.

Yes, i know all about it being the norm
ive got my own tests that prove my cpu is stable
occt for an hour / crysis / gta 4 / Re-encode 10 gb file

i just dont need to run prime for 12 hours

when i have used prime ive only ever used it for 1 hour
seriously is you're pc ever going to be more stressed out than an hour of prime running?

Yes easily in fact. 10 hours "complete loop" of not just prime but the prime blend test is the standard defacto best all around test for stability there is. The small fft test only loads the CPU and doesn't stress the memory controller at all, for this reason it's better to run the Blend test for overall stablity.

In fact there are 37 different K tests in prime blend and each one of them applies stress factors that are slightly different from each other. To run them all it usually takes around 10 hours of testing,.

#1) 4,000 @ 1024k / 0:15
#2) 800,000 @ 8k / 0:30
#3) 560,000 @ 10k / 00:45
#4) 4,500 @ 896k / 1:00
#5) 5,300 @ 768k / 1:15
#6) 460,000 @ 12k / 1:30
#7) 380,000 @ 14k / 1:45
#8) 6,500 @ 640k / 2:00
#9) 7,800 @ 512k / 2:15
#10) 340,000 @ 16k / 2:30
#11) 270,000 @ 20k / 2:45
#12) 9,000 @ 448k / 3:00
#13) 11,000 @ 384k / 3:15
#14) 210,000 @ 24k / 3:30
#15) 180,000 @ 28k / 3:45
#16) 13,000 @ 320k / 4:00
#17) 17,000 @ 256k / 4:15
#18) 160,000 @ 32k / 4:30
#19) 120,000 @ 40k / 4:45
#20) 19,000 @ 224k / 5:00
#21) 22,000 @ 192k / 5:15
#22) 100,000 @ 48k / 5:30
#23) 84,000 @ 56k / 5:45
#24) 27,000 @ 160k / 6:00
#25) 34,000 @ 128k / 6:15
#26) 75,000 @ 65k / 6:30
#27) 56,000 @ 80k / 6:45
#28) 39,000 @ 112k / 7:00
#29) 46,000 @ 96k / 7:15
#30) 3,100 @ 1280k / 7:30
#31) 2,500 @ 1536k / 7:45
#32) 2,100 @ 1792k / 8:00
#33) 1,900 @ 2048k / 8:15
#34) 1,500 @ 2560k / 8:30
#35) 1,200 @ 3072k / 8:45
#36) 1,000 @ 3584k / 9:00
#37) 900 @ 4096k / 9:15

1 hour isn't even close to being stable.

It's always best to prove your overclock is stable not just guess at it.
I defy anyone here to run the Prime Blend test for the complete 10 hour loop then post the results here.

I wouldn't consider running an overclock with anything less.

 

[madtownidiot]

Ever see those idiots who use liquid nitrogen cooling to get some ridiculously high overclock then post it on youtube and think they're accomplishing something?

 

[chev65]

Ever see those idiots who use liquid nitrogen cooling to get some ridiculously high overclock then post it on youtube and think they're accomplishing something?

Exactly the reason for posting this and thanks for helping me prove my point.

I did it with some attitude I'll admit but there are far too many unusable overclocks with unproven stablity being posted everywhere and I just couldn't take it anymore.

 

[madtownidiot]

Ever see those idiots who use liquid nitrogen cooling to get some ridiculously high overclock then post it on youtube and think they're accomplishing something?

Exactly the reason for posting this and thanks for helping me prove my point.

I did it with some attitude I'll admit but there are far too many unusable overclocks with unproven stablity being posted everywhere and I just couldn't take it anymore.

I agree, it's kind of stupid to overclock beyond a setting that can be permanently maintained, especially when you're risking a couple thousand worth of hardware. I usually try to find the highest stable OC possible then back it off 10-15% and don't mess with it any more.

 

[dgwin7]

Here is my i5 750 overclocked to 4.1 pix below.

dg

Wow nice, what cooling do you have?

I just bought this Arctic Cooling Freezer 7 Pro Rev.2 heatsink with fan not bad after all.

dg

 

[Windows i7 920]

My very first overclock. I think my 920 did pretty well. Well, really, the ASUS P6T Deluxe V2 motherboard with ASUS TurboV helped too. I can't believe I'm posting at 4GHz right now.

Also I thought the 920 had a multiplier limit of 20. Mine went up to 22!



Overclock attempt #2. This one ended up being more stable.