Prime95 25.7

[VirtualLarry]

link

 

[petrusbroder]

Thanks! It was a longtime I checked the versions ... 😉

 

[LOUISSSSS]

thanks for the link, but to save some time for EVERYONE who stability tests their computers. the latest OCCT beta supposedly uses the same algorithm as P95 but better optimized to find errors quicker.

i've found this true in my own testing. i had an unstable system and p95 took 18 hours to find an error, i then ran occt and it found the error within 1 hour

 

[LOUISSSSS]

btw wrong forum? i'd put this in CPUs

 

[petrusbroder]

Not necessarily. We are very much interested in stability and reproducibility when crunching. 🙂

 

[Assimilator1]

Originally posted by: LOUISSSSS
btw wrong forum? i'd put this in CPUs

No it's fine here, P95 is a DC project afterall 🙂, & yea not to mention loads of us like to test our o/c rigs far more thorougly than the average o/c'er 😉.

Btw worth knowing OCCT, I'll try that next time :thumbsup:, I wonder if it's quicker than F@H?

 

[Drsignguy]

Originally posted by: Assimilator1

Originally posted by: LOUISSSSS
btw wrong forum? i'd put this in CPUs

No it's fine here, P95 is a DC project afterall 🙂, & yea not to mention loads of us like to test our o/c rigs far more thorougly than the average o/c'er 😉.

Btw worth knowing OCCT, I'll try that next time :thumbsup:, I wonder if it's qukcker than F@H?

It is, and Intel burn test is even faster. But have to warn everyone who doesn't know about this program, it will raise the temps higher than OCCT or prime.

 

[LOUISSSSS]

Originally posted by: Drsignguy

Originally posted by: Assimilator1

Originally posted by: LOUISSSSS
btw wrong forum? i'd put this in CPUs

No it's fine here, P95 is a DC project afterall 🙂, & yea not to mention loads of us like to test our o/c rigs far more thorougly than the average o/c'er 😉.

Btw worth knowing OCCT, I'll try that next time :thumbsup:, I wonder if it's qukcker than F@H?

It is, and Intel burn test is even faster. But have to warn everyone who doesn't know about this program, it will raise the temps higher than OCCT or prime.

i've heard that the intel burn test doesn't actually find errors faster, its just coded to produce more heat from the CPU, that doesn't necessarily = find errors faster

 

[Drsignguy]

Originally posted by: LOUISSSSS

Originally posted by: Drsignguy

Originally posted by: Assimilator1

Originally posted by: LOUISSSSS
btw wrong forum? i'd put this in CPUs

No it's fine here, P95 is a DC project afterall 🙂, & yea not to mention loads of us like to test our o/c rigs far more thorougly than the average o/c'er 😉.

Btw worth knowing OCCT, I'll try that next time :thumbsup:, I wonder if it's qukcker than F@H?

It is, and Intel burn test is even faster. But have to warn everyone who doesn't know about this program, it will raise the temps higher than OCCT or prime.

i've heard that the intel burn test doesn't actually find errors faster, its just coded to produce more heat from the CPU, that doesn't necessarily = find errors faster

I think it does. I have used all "tests" and Intels finds them pretty quick. You just have to watch for it on the first run. OCCT will find them very quick too. All depends, if very unstable, OCCT will BSOD, and or bring up the temp and voltage charts. Preferance really?

 

[petrusbroder]

Hmm, very interesting. I do not like the Intel-burn test due to the heat it generates, but OCCT sounds nice.
I have only used Prime95 until now ... 24H and no errors is good enough for me.

 

[LOUISSSSS]

Originally posted by: petrusbroder
Hmm, very interesting. I do not like the Intel-burn test due to the heat it generates, but OCCT sounds nice.
I have only used Prime95 until now ... 24H and no errors is good enough for me.

what does the amount of heat generated have to do with stability???
if the intel test can bring your CPU to 90C and find the same amount of errors as OCCT in the same amount of time, what does the temp mean? you're never going to reach that temp under normal usage...

all that matters is that the program finds no errors in "x" time period.

 

[Drsignguy]

Originally posted by: LOUISSSSS

Originally posted by: petrusbroder
Hmm, very interesting. I do not like the Intel-burn test due to the heat it generates, but OCCT sounds nice.
I have only used Prime95 until now ... 24H and no errors is good enough for me.

what does the amount of heat generated have to do with stability???
if the intel test can bring your CPU to 90C and find the same amount of errors as OCCT in the same amount of time, what does the temp mean? you're never going to reach that temp under normal usage...

all that matters is that the program finds no errors in "x" time period.

Proper cooling! F@H is a demanding program and it does require good cooling as well as these stress tests. If you are going to run 24/7, good airflow and good cooling is essential. So, it isn't just stability, its the heat generated by it. These tests can bring out the worst case senerio not only for stability but your temps.🙂

 

[Assimilator1]

Originally posted by: petrusbroder
Hmm, very interesting. I do not like the Intel-burn test due to the heat it generates, but OCCT sounds nice.
I have only used Prime95 until now ... 24H and no errors is good enough for me.

I & others have found that you can pass P95 for 24hrs but when you run F@H it can fail in 10s of mins!:Q, I guess it depends how stable you want it 😉.

Btw what does OCCT show when it finds an error? (it hasn't yet found one, currently testing my Q6600 on my new mbrd @2.9GHz. On my old mbrd it topped out at 3GHz).