New Test For CPU Stressing??

[Butch]

I'm using Prime95 to stress test a Phenom II X4 965. One of the workers keeps stopping. It started on #2 now 2 is working and #3 stops. I can't find the file that is suppose to explain the failure.

Is there another stress test that I can try?

Thanks,
Butch

[mmettin]

IBT ?

http://majorgeeks.com/IntelBurnTest_d5987.html

[Butch]

Thanks!
I downloaded installed and ran at High and Standard. Passed both!

So is there a reason I fail P95 and pass intels test?

I am going to redownload P95 and try again.

Butch

[mmettin]

Thanks!
I downloaded installed and ran at High and Standard. Passed both!

So is there a reason I fail P95 and pass intels test?

I am going to redownload P95 and try again.

Butch

I'm gonna guess an old or broken P95

[Butch]

Your guess is correct!
I was using the 266v from there site and MajorGeeks had 267v that seems to be running fine.

I'll post back saturday and let you know how it turns out.

Thanks for the help!!
Butch

[Butch]

Small FFT's ran 12 hours no errors.
Tried Blend but workers 2 and 3 stopped. The message:
"Fatal error: Rounding was 0.5, expected less than 0.4".

This happened on #2 after 1:16 running and #3 after 1:41.

Workers 1 and 4 ran 2:35 untill I stopped with no errors.

Ram Memtest 24hours no defects.

Does passing small fft test and failing blend help solving this?

Thanks, Butch

[Activate: AMD]

Small FFT's ran 12 hours no errors.
Tried Blend but workers 2 and 3 stopped. The message:
"Fatal error: Rounding was 0.5, expected less than 0.4".

This happened on #2 after 1:16 running and #3 after 1:41.

Workers 1 and 4 ran 2:35 untill I stopped with no errors.

Ram Memtest 24hours no defects.

Does passing small fft test and failing blend help solving this?

Thanks, Butch

if it passes IBT and small FFT's for 12 hours, then the CPU itself is probably stable since small FFT's really only stress the immediate CPU subsystem IIRC. what happens if you try large FFT's? Thats more of a test of the memory/IMC system

[Butch]

The memory will Pr95 at 1333 all day and night on small or large ftts and blend but starts having errors if I try to run at rated speed of 1666.

It's Corsair CML4GX3M2A1600C9. Tried settings manually and auto with same results on an Asus M4A88T-M.

Any ideas before I RMA from the Overclockers here???

Thanks,
Butch

[Sidd]

I believe that running at 1600 is overclocking the IMC on that processor you may have to loosen the timings or up the voltage to get it to run at those speeds

[Butch]

Switched to TW3X4G1333C9A memory.On the supported list. Still have the same error running Pr95. Once again Memtest shows no errors.

The error took 2hrs to show on the worker.

Am I just chasing a ghost? IBT passes, Memtest passes, it just has the same error rounding that 0.5.

This is not my PC or I'd just run it and see how it goes.

Any chance a driver issue? Had a lot of trouble getting the embeded ATI to load drivers correct?

Reaching out here but I sure would like to run Pr95 with no errors.

Thanks,
Butch

[Butch]

I took the cooler off and put AS5 on the cpu.

I found a small nick on cpu and burr on cooler. I didn't see a nick before I put the cooler on so I am guessing that the burr was covered by the stock thermal paste that was applied by AMD.

Got the burr off cooler and applied AS5 to CPU.
Been testing with OCCT 4.2.0 and for the last 2 hours no errors. Running between 55C and 57C.

Should I return the CPU or do you feel it will be fine???

Thanks,
Butch

[Butch]

Prime95 blended going for over 2.5 hours. No errors!!

[Butch]

18 hours Prime 95 and no errors! I guess it's a keeper.

After doing IBT so many times I'm lucky I finally pulled the cooler off and found that nick.

Thanks,
Butch