What is Prime95?
- Thread starter Nitrojan
- Start date
Interesting opinion?
actually apart from just being a mathmatically complex task, Prime screens are a good way for overclockers to show that their work is stable.
There's a big rush to claim some massive overclock but few people bother to test that overclock and publish screenies?
I always appreciate peoples effeorts and seeing a good set of shared results gives me confidence that product XYZ is actually doing the business!
Nitrojan dont worry about some of these posters as they just like winding people upLike you I am a noob but for your own safety do a search amongst the forums and you will more than likely find your answers
Take care and enjoy your PC
Thanks to all those being supportive ;P
Last edited:
Soldato
- Joined
- 10 Apr 2004
- Posts
- 13,497
70C is nothing when its designed for 90+C.
And a few minutes really is not long enough.
An hour at-least and encoding can quite easily get you to the same CPU usage levels.
ran a quick test of prime vs snm for temp testing to see which one gets the core hotter.
first pic shows prime run for 11 mins. used small fft since its supposed to be the most intensive cpu test prime can do.
as you can see peak temps were 52c for the hottest core.
i then stopped prime and let the pc rest back to idle for a few mins as shown here:
as you can see realtemp records the peak temps and shows what they were during the prime run.
then after the rest i clicked on the run in snm:
as you can see just a little over 3 mins and the remps are already much higher than what prime manages.
also take into account that i have smartfan enabled so the fan ramped up even more in speed during the snm run, so if i had set both tests to a fixed fan speed there would have been an even larger difference in temps between prime and snm.
tested this on cpus since the old socket A durons and prime has always has been the worst test of stability for me since prime would be 50+ hours stable yet crash when playing a cpu intensive rts game.
never used prime since for stability testing.
testing goes to show how a program used to find prime numbers cant compete in stressing the cpu more than a program which has been designed from the ground up to do just 1 task which is stress the cpu.
edit: some additional infor, prime version used was 25.6 build 6. snm version 1.9.1 was used.
first pic shows prime run for 11 mins. used small fft since its supposed to be the most intensive cpu test prime can do.
as you can see peak temps were 52c for the hottest core.
i then stopped prime and let the pc rest back to idle for a few mins as shown here:
as you can see realtemp records the peak temps and shows what they were during the prime run.
then after the rest i clicked on the run in snm:
as you can see just a little over 3 mins and the remps are already much higher than what prime manages.
also take into account that i have smartfan enabled so the fan ramped up even more in speed during the snm run, so if i had set both tests to a fixed fan speed there would have been an even larger difference in temps between prime and snm.
tested this on cpus since the old socket A durons and prime has always has been the worst test of stability for me since prime would be 50+ hours stable yet crash when playing a cpu intensive rts game.
never used prime since for stability testing.
testing goes to show how a program used to find prime numbers cant compete in stressing the cpu more than a program which has been designed from the ground up to do just 1 task which is stress the cpu.
edit: some additional infor, prime version used was 25.6 build 6. snm version 1.9.1 was used.
Last edited:
not sure on what method snm uses but other forums say its very similar to what intel TAT uses.
iv had snm miss some errors before but those errors were memory related, on my opteron setup command rate 1t was unstable in prime but passed in snm.
lots of hunting around though pointed to forum threads telling me to run memtest test number 8 which is very good at picking out command rate issues.
and what do you know 2nd pass of test 8 clamped the errors in the memory, test 5 in memtest missed it all even after it had been running test 5 overnight over 10 hours. lol
best to use isolation tests i guess, i now let memtest run all its tests overnight before i use the system after an overclocking session.
iv had snm miss some errors before but those errors were memory related, on my opteron setup command rate 1t was unstable in prime but passed in snm.
lots of hunting around though pointed to forum threads telling me to run memtest test number 8 which is very good at picking out command rate issues.
and what do you know 2nd pass of test 8 clamped the errors in the memory, test 5 in memtest missed it all even after it had been running test 5 overnight over 10 hours. lol
best to use isolation tests i guess, i now let memtest run all its tests overnight before i use the system after an overclocking session.
needed to add, worst thing that pushed me away from from was that i had system that was 56hour prime stable. yep i let it running over the weekend and checked it when i got back. 56hour stable in prime yet it crashed when running a game called dawn of war for around 1 hour.
did my noggin it that did. eventually found out about snm, used that and it errord out after running for 5 mins. never used prime since, and that on my old barton xp cpu i believe.
did my noggin it that did. eventually found out about snm, used that and it errord out after running for 5 mins. never used prime since, and that on my old barton xp cpu i believe.