Is slight over voltage an issue?

[setscrew]

I had a weird upgrade experience detailed in this thread, but I may have asked too many questions as there have been no responses. So under that assumption, I will pose a detailed question here.

My system will not pass memtest86 nor prime95 (small FTT, large FTT and blend), but runs windows and folding@home. The voltages reported via software (will verifiy with DMM soon) seems a tad high, but assuming the readings are accurate, would they lead to failures in the above testing programs?

Thank you for you time.

Voltages from MBM 5 (forgot my DMM at work):
Core 0: 1.34V
Core 1: 3.42V
3.3: 3.47V
5.00: 5.18V
12: 12.35V

P4 3.0E 478 pin, stock cooler (no overclocking)
corsair value 2 x 512 (also tried Ballistix 2 x 1GB this weekend)2.5,3,3,6 (2.62V. I think, no means for adjustment in BIOS)
XP Home SP2
Enermax EG425P-VE SFMA
Abit SG-72
eVGA 6800GT with VF700 (no OCing; lite artifacts in UT2004 @ 1280x960)

 

[burningrave101]

Their mostly within spec but your 3.3V rail is a little higher then it should be unless you raised it on purpose. Some people overclock their 3.3V rail to get more juice to their RAM. As long as their within +/- 5% i wouldn't worry too much but i would take a DMM reading to see whats going on.

 

[[Spectre]]

Your system won't prime but it will fold without error? That is odd.

And let's wait and see what the multimeter says but burningrave is right within spec it should be golden.

 

[setscrew]

I do not know if it is folding without error.

All I know is that it is folding, windows has no issues with it and it does not crash to desktop. I am sure you know this, but memtest86 and prime95 both have a value that they are looking for. Thus when the genereated value is compared to the database and there is a disparity, you have an error (in memtest86) or a halted program (prime95).

It is very confusing. Before I obtain a replacement processor or RAM, I want to be sure that the underlying issue is no longer present. I did not raise any of the voltages, nor is there an option to do so in my BIOS (unfortunately). I am happy to hear that once I verify the voltages with my DMM, I could safely rule out the Enermax power supply as the cause for memtest86 and prime95 failing.

I guess RAM and processors go bad eventually.

Thank you for the input to date.

 

[[Spectre]]

I do not know if it is folding without error.

All I know is that it is folding, windows has no issues with it and it does not crash to desktop. I am sure you know this, but memtest86 and prime95 both have a value that they are looking for. Thus when the genereated value is compared to the database and there is a disparity, you have an error (in memtest86) or a halted program (prime95).

I am aware how they work....folding is just seems to be a lot more stressful if you do it full tilt.

It is very confusing. Before I obtain a replacement processor or RAM, I want to be sure that the underlying issue is no longer present. I did not raise any of the voltages, nor is there an option to do so in my BIOS (unfortunately). I am happy to hear that once I verify the voltages with my DMM, I could safely rule out the Enermax power supply as the cause for memtest86 and prime95 failing.

I guess RAM and processors go bad eventually.

Thank you for the input to date.

Well strip down to the minimum config needed to boot and go from there. Borrow a PSU or use a back up to isolate that....but bad memtest and bad prime but not bad folding really has me baffled.......and yeah of course RAM and CPU's can go bad but I doubt(but it could happen) both would have gone. Hmmmm.....

 

[setscrew]

No offense intended Spectre.

Would failing small FTT under prime95 be a positive test for a failing CPU? Or if we assume the RAM is bad, would this casue the small FTT to fail?

I am stumped. I will attempt to obtain another PS and start from there I guess. I just hate RMAing the CPU and RAM if there is another cause.

Thanks to all so far.

 

[[Spectre]]

No offense intended Spectre.

Would failing small FTT under prime95 be a positive test for a failing CPU? Or if we assume the RAM is bad, would this casue the small FTT to fail?

I am stumped. I will attempt to obtain another PS and start from there I guess. I just hate RMAing the CPU and RAM if there is another cause.

Thanks to all so far.

So the way I remember it is failing small FTT is CPU while the belnd test is either the CPU or RAM.

 

[burningrave101]

So the way I remember it is failing small FTT is CPU while the belnd test is either the CPU or RAM.

Small FFTs and large FFTs are going to be primarily CPU. Usually blend test failure is more because of the RAM because your not running FFTs in-place so its stressing the CPU about 80%. If you really want to stress the CPU you should choose custom and set the Min FFT size to 8K and Max FFT size to 4096K and put a tick in "Run FFTs in-place". You should then leaving it running for at least 8 hours.

 

[[Spectre]]

Small FFTs and large FFTs are going to be primarily CPU. Usually blend test failure is more because of the RAM because your not running FFTs in-place so its stressing the CPU about 80%. If you really want to stress the CPU you should choose custom and set the Min FFT size to 8K and Max FFT size to 4096K and put a tick in "Run FFTs in-place". You should then leaving it running for at least 8 hours.

Thanks burningrave at least my memory is still partialy functional.

 

[setscrew]

Verified MBM readings with DMM at 20 pin connector
3.30 = 3.424V.
5.00 = 5.170V.
12.00=12.180V.

Thank you Spectre and burningrave101 for the advice concerning prime95. I will run that prime95 set-up tonight. If it fails the eight hour test, I am taking processor (and perhaps mobo) back to the local vendor for an RMA. I also ordered some replacement RAM so that I can RMA the modules I currently possess.

Oh, just as an additional test, I ran the Windows MemTest to augment memtest86 and prime95; the test produced 300+ errors in six minutes.

I hear some talk of super PI. Is that program better than or equal to prime95?

 

[[Spectre]]

Verified MBM readings with DMM at 20 pin connector
3.30 = 3.424V.
5.00 = 5.170V.
12.00=12.180V.

Those voltages are fine. So that shouldn't be an issue.

Thank you Spectre and burningrave101 for the advice concerning prime95. I will run that prime95 set-up tonight. If it fails the eight hour test, I am taking processor (and perhaps mobo) back to the local vendor for an RMA. I also ordered some replacement RAM so that I can RMA the modules I currently possess.

Oh, just as an additional test, I ran the Windows MemTest to augment memtest86 and prime95; the test produced 300+ errors in six minutes.

I hear some talk of super PI. Is that program better than or equal to prime95?

Yeah that sounds like RAM then. I only ever use superpi for benchmarking not diagnostic. It may be useful but I am not sure how..........hopefully someone else can answer that for you more completely.