Abit IC7-G and Prime95

I've been having some issues running Prime95 stably. It generally generates an error within 5-10 minutes of starting to run 2 instances. (It has also errored when I've only run a single instance of Prime 95.)

I've got a P4 2.4C running at 2.4 gHz in an Abit IC7-G (Bios v. 1.3). Corsair TWINX 3200LL running at 2-3-3-7 (and DRAM at 2.7v). I have a Sony DVD-RW, a floppy drive, and a 120 gig maxtor IDE drive. All powered by a 400W Antec PS.

Relaxing the memory timings, increasing DRAM voltage to 2.8v, and increasing vcore by up to .05 have had no effect. Prime95 will still crash. Testing with each individual stick also seems to have had no effect. Sooner or later, Prime95 generates an error. The memory has no trouble passing memtest86, even at the rated 2-2-2-6 speed.

My first question is, at what temperature as reported by Winbond's Hardware Doctor should I expect heat to play a role in causing Prime95 failures? With the retail heatsink (keeping the warranty in tact for now) and running 2 instances of Prime95, the heat steadily climbs towards 68 C.

Also, how reliable is memtest86? How many passes through the 12 or so tests without an error would need to occur to give you confidence in the memory? Does Prime95 stress memory in a way that would cause it to expose hardware defects that memtest86 would miss?

Has anyone had a similar experience with the IC7-G and Prime95?

If you are using the latest version of prime95 and getting errors then there is somtething wrong !!

68c in IC7G is not very high ... i have seen about 76 while doing some high overclocking ... and it was prime95 stable with 2 primes running with affinity ...

As for the memtest86 , it's better to use only the the 3 and the test 5.
Test no3 tests the cpu ... so leave it about 1-2 hours and it should be enough...
test no5 tests the memory ... do the same for 1-2 hours ...

If you get no errors the you should be OK ... but i think that you are going to see some ...

BTW if you have the performance mode to ultra or anything else just turn it to disabled ... just to be sure ...

68c is definitely too high at stock speed even on the IC7.....try reinstalling your HS + thin layer of thermal paste. I'm running my 2.6C @ 3.25 1.675v(bios) and under prime95 temps are 65c max after several hours.

...But you don't know the temperature of his room ! ... if he lives in a hot place ????

I'm just using the intel retail heatsink and fan for now. I don't want to void the warranty. I have a copper heatsink and faster fan, along with some arctic silver just waiting for when the machine can run prime95 stably. The air temp around here has been around 27 C. Really, I just wanted to know if 68 C (as reported by Hardware Doctor) could be the cause of prime95 failures. I guess the answer is no.

I ran memtest86 test #3 for 2 hours. No errors.
I ran memtest86 test #5 for 2 hours. No errors.

Performance mode has never been enabled.

The version of Prime95 I am running is...v22.12.1

I've even tried reinstalling Windows XP, to no avail. I'll grab some more sticks of memory and see what the deal is. I appreciate the help. Thanks.

Too high temperature can definitly make Prime fail.
I have a Prometeia and Prime is stable upto FSB 250. Right now I'm burning in on air and Prime fails at 63*C everytime, and I'm only at FSB 225 right now!
Luckily I know my CPU is good, but purely based upon these air test result I would start to think otherwise.

Cooling is everything!

I guess I cant complain about my temps. Idle 40c under load prime95 56. I'm running 3.2ghz with 263fsb no problems yet.Using air cooled copper heatsink

Question, if running two instance prime and one fail at about 3 hours and the other keep going for over 24 hours, do you guy consider it as stable?

My setup now is prime(both instance)stable @ 277fsb with 1.65vcore,but if I try 280fsb then one out of two prime95 will fail at about 3 hour and the other will keep going overnight.

Any suggestion is greatly appreciated.

I am starting to think that there are some issues with prime95 ...
right now it crashes in 1 minute , but i can do mpeg2 compression for 10hours , run 2 seti with my hyperthreading all night , run 2 tasks of burnP6 for many hours (which produces about 3-4 higher temp than 2 primes) , run memtest test 3 and 5 for hours without errors....
If i give a little more voltage to my CPU ([email protected] with 1.6V right know) , like 1.625 it will crash after 9 hours ...some times after 1 ..
...I don't know ...
I also noticed that it's PSU dependant ... if you have a weak PSU , that is working at full load , prime95 is going to fail
Some guys at prime95 forum have written that prime95 may crash due to driver bugs .... who knows ???

In other words ... it's better to test your stability with many programs ...