Prime95, run for how long?

Slinky1989

Slinky1989

Slinky1989

Soldato
Joined
8 Mar 2007
Posts
3,586
Location
Swindon, Wiltshire
Basically I've been dabbling in some amateur overclocking of my p4 620 (3.2ghz running at 3.5ghz).

It seems ok but I want to stress it and find out if it's stable.

I've just downloaded prime95, and I was wandering how long I should run it to see if my overclock is trully stable, and what will happen if it isn't?


I'm slightly worried as before this overclock, I tryed to get it to 3.6 but it screwed up my windows installation hence why I'm not going that far.

Any help would be appreciated :)
 
There is really no point running prime or orthos for more than the time you normally spend on the computer in one go. It is also a case of diminishing returns - running it for 12 hours isn't going to give you much more confidence than 3 hours.
 
Bomag said:
running it for 12 hours isn't going to give you much more confidence than 3 hours.
Click to expand...
Yes it will, I've seen an overclock fail after 8 hours of prime. Anyway its a personal thing as long as you let other people know what tests you've done.
 
Bomag said:
There is really no point running prime or orthos for more than the time you normally spend on the computer in one go. It is also a case of diminishing returns - running it for 12 hours isn't going to give you much more confidence than 3 hours.
Click to expand...

Sorry mate thats rubbish
 
CompMan said:
silly but my jist is one can go overboard on the stabilty test
a system that fails a test may never crash day to day ever.
Click to expand...


Its not about crashing,Its about data integrity.
 
Last edited:
You guys say "orthos 5 hours". Shall I take that as orthos on all 4 test for 5 hours each? thats 1 day.

I was clocked at 3Ghz and could run gromacs for days without a crash, yet small ffts failed in 25minutes.
 
I see this question pop up yet again, with all the usual ill informed rubbish answers. The simple fact of the matter is, unless it's Prime Blend or Orthos Blend stable for minimum of 8 hrs then it is'nt a stable clock, period. Anyone who thinks otherwise is deluding themselves and giving other people bad advice. It also isn't about the "my rig does everything i need it to do" mentality. Stability is about the integrity of data, as much as it is about playing this game or that game. If it's not Prime or Orthos stable it will corrupt your install at some point and worse it will corrupt your data as well.
 
so... the best way to test fr stability is to run orthos blend at priority 10 for 8 hours?

ps: Can some of you vet guys explain abit the different tests in orthos. If small fft fails etc what is means? cause gromacs might go on for hours and then fail mall fft in minutes or something similar with other tests.
 
Priority 10 is MUCH more harsh than 1. Sometimes if Orthos is stable for 24 hours at priority 1, it may fail priority 10 after only 1 hour.
 
Surely you need to run it so that has run all the different tests, with the different iterations to make sure the CPU can handle it?
 
kitfit1 said:
I see this question pop up yet again, with all the usual ill informed rubbish answers. The simple fact of the matter is, unless it's Prime Blend or Orthos Blend stable for minimum of 8 hrs then it is'nt a stable clock, period. Anyone who thinks otherwise is deluding themselves and giving other people bad advice. It also isn't about the "my rig does everything i need it to do" mentality. Stability is about the integrity of data, as much as it is about playing this game or that game. If it's not Prime or Orthos stable it will corrupt your install at some point and worse it will corrupt your data as well.
Click to expand...

There is no such thing as a guaranteed 'stable' clock. After about 5-10 hours the difference in probability of an error with a stock clock vs overclocked are very much the same (in terms of Human perception anyway). Also the probability of Windows corrupting data after a stable 3 hour run just because it is overclocked is going to lower than windows underlying propensity to corrupt data.
 
Personally I don't put that much trust in Prime or Orthos. I have pushed clocks with said stress tests for 12 hours plus and then had my system fail within 1 hour of playing a game. Nasty if you loose you game save. Even worse if you corrupt your OS.

Here's what I do. Run orthos for a couple of hours and then back off the clocks by 100-200 and then play some games. Playing games is the ultimate test for stability.

Use Orthos as an indication of stability and nothing else.

IMHO anyway :)
 
Ok I can see this thread going on and on (and on) . . .

In short, a good overclock will do everything a stock processor can do, only faster. If you can run Prime/Orthos for 12 hours while the CPU is at stock but not when its overclocked something is wrong.

As long as people inform others what kinda stability tests they have done then all is good i.e: "Yo I have my e6600 at 3.8GHz.Battlefield 2 stable"

I think there are too many impatient overclockers, in a rush to produce a stellar overclock, not because they actually need the extra speed but more so that they can come to the forums and make themselves look smart and collect their medals :D

In the cold light of day it means little, your DVD-encodes will take a few minutes less and your games play at 147FPS instead of 139FPS lol ;)

I encourage people to perform at least one round [1/1] of Orthos until they state an overclock is stable, CPU-z screenies, save it for xtreme :o
 
You must log in or register to reply here.
Back
Top Bottom