Prime 95 error ??

[R4-NV]

Hi all I am new hear

Dos anyone what is this error in Prime 95 is ???

Prime 95 was ran 18 hours then got this error


Part of the test results

[Wed Dec 19 05:29:18 2012]
Self-test 96K passed!
Self-test 96K passed!
Self-test 96K passed!
Self-test 96K passed!
Self-test 1280K passed!
Self-test 96K passed!
Self-test 96K passed!
Self-test 96K passed!
Self-test 96K passed!
Self-test 96K passed!
[Wed Dec 19 05:38:43 2012]
Self-test 96K passed!
Self-test 1280K passed!
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
[Wed Dec 19 05:45:28 2012]
Self-test 1280K passed!
Self-test 1280K passed!
Self-test 1280K passed!
Self-test 1536K passed!
Self-test 1280K passed!
Self-test 1280K passed!
Self-test 1280K passed!
Self-test 1280K passed!
Self-test 1280K passed!
Self-test 1280K passed!
[Wed Dec 19 05:54:45 2012]
Self-test 1280K passed!
[Wed Dec 19 06:01:19 2012]


My rig is at stock settings apart from I have set my memory profile to X.M.P though to only unlock the full Protestant of the ram
than you for any help!

[Agent]

It means that the calculation wasn't as expected. If everything is at stock like you say, it's almost certain there is a failing part (or something overheating - although a less likely).

Give Memtest86 a run next and see if you get any errors from that

[R4-NV]

Thanks i will try that

[R4-NV]

I have ran Memtest86 off my spare thumb drive 3 times no errors

[Agent]

Now it gets a bit more complex sadly

Double check everything is at stock: RAM timings are correct for instance. It's unlikely to be causing the issue, but I'd also update the BIOS if there is a newer version.

Prime shouldn't really error at all.

I'd recommend giving the Intel burn test a shot: http://www.xtremesystems.org/forums/...esting-program

It will probably also cause errors faster than Prime.

The other obvious culprits are mobo and PSU (it could be dropping the voltage when under sustained heavy load). Sadly temperamental issues like this are a nightmare to solve

[R4-NV]

Hi thank you for the reply
Sorry I should have mentioned it before but I am running the latest Bios from Asus I have used OCCT and Intel burn test and Intel processor diagnostic tool Prime 95 is the only test that shows a fault I was wondering if a previous bent pin on the CPU socket would cause this fault?
Ram timings is 9-9-24-2N- 1.50v – 1.20v
Thank you for all you help so far and I hope you have a good Christmas

[technodean]

do you mean 3 separate tests or did you let it run for 3 passes consecutively? I would run it for a good number of hours. also are you running mem test 86 or 86+, mine passed using 86 (see first image) then with 86+ passed 2 passes then with nearly 2 hours in boom 2000 errors(second image).

[R4-NV]

hi thanks for the reply technodean
I ran Memtest86 for 3 passes consecutively i will try to see if Memtest86 will do 5 passes and see if any think happens

thanks technodean i hope you have a nice Christmas

[technodean]

thanks and a Merry Christmas to you. But make sure you use Memtest86+ as Corsiar themselves tell there customers to only use the plus test.

[R4-NV]

Hope every one has had a good Christmas

I am using Memtest86+ for four passes and have no fault

[R4-NV]

Now it gets a bit more complex sadly

Double check everything is at stock: RAM timings are correct for instance. It's unlikely to be causing the issue, but I'd also update the BIOS if there is a newer version.

Prime shouldn't really error at all.



I'd recommend giving the Intel burn test a shot: http://www.xtremesystems.org/forums/...esting-program

It will probably also cause errors faster than Prime.

The other obvious culprits are mobo and PSU (it could be dropping the voltage when under sustained heavy load). Sadly temperamental issues like this are a nightmare to solve

thanks and a Merry Christmas to you. But make sure you use Memtest86+ as Corsiar themselves tell there customers to only use the plus test.

Any ideas on this?

[=assassin=]

I always thought these Prime test stability was down to user preference from what I've read before. 18 hours is pretty decent tbh for a continuous highly stressful test, depending on what you use your PC for, you may never get any issues at all in normal use. It is possible to be too OCD fine tuning things.

[kalniel]

Any ideas on this?

Well I'd echo what Agent says really. Prime95 should never give an error - you are asking your computer to do a calculation and checking the answer against the known correct result. A failure means it got the sum wrong, this can happen if the signal being passed around the inside of the computer is 'misheard' slightly as it is shuttled back and forth. RAM is the cheapest and most obvious source of this, but memtest86 is a thorough RAM-only test for this kind of thing and if you pass that after a long period you can probably rule out the RAM.

So the fault is either a one-off (powerline surge, system interupt?) or it lies somewhere else like the CPU, possibly in the memory controller when hot. Was a lot of other stuff running on your PC when you did the prime95 run? Are you having any other stability problems?

With the number of possible factors I wouldn't be worried about one failure in a lifetime, unless I was using the computer for something mission critical. If it is mission critical then perform a clean boot (http://support.microsoft.com/kb/929135) and run prime95 again. If you can repeat the error then disable XMP and manually set RAM to the specified timings/voltages supported by ALL of the RAM, motherboard and CPU.

[NorBdelta]

Any Error such as the one you have experienced is a hardware error, if your CPU is not over-clocked then it is faulty get another.

If it is overclocked then revert back to stock and run it again.