Utility for testing a cpu?

[exeterdad]

Something is going on with this computer.  I had faulty ram a while back.  And just recently got a decent stick so I could run without errors.  Well I've got errors.  Memtest86 shows nothing is wrong after letting it cycle 10 times.  I'm not anywhere near overheating.  I tried slowing the ram down and thought it might be running more stable, but that didn't last long.  I suspect the cpu.  It's a 2.4 ghz P4 800 mhz fsb prescott.

I do have a 3 ghz P4 800 mhz fsb prescott that "should" run on this board, but it doesn't so a cpu swap isn't going to help narrow this down.

I get freeze ups.  Sometimes segfaults.  Sometimes I can't decompress any kind of packages or archives without errors.  But sometimes this machine is as fast and stable as can be without any problems at all.  I don't get it.  This cpu has never been overclocked.

Any suggestions on some sort of utility?  Things are tight for us for a while.  If and when I buy the part that will resolve this, it has to be right the first time.

Looks like I have to move back in to the Ol' Athlon 1ghz Thunderbird computer for a while.  It's slow, but it's steady.

[lagagnon]

Any suggestions on some sort of utility? 

Prime95 : http://www.mersenne.org/freesoft.htm

[exeterdad]

After much interesting reading.  I'm not clear how software designed to search for prime numbers will help diagnose my issue? 

[wcs]

After much interesting reading.  I'm not clear how software designed to search for prime numbers will help diagnose my issue?  Huh

Oh, it will... Searching for prime numbers is apparently a crazily intensive process. You will be notified of any little error in the computations, which can be caused by faulty processors. You should leave it running for quite a while.

So far as I remember, there is an option in mprime that is specifically for CPU testing.

Check http://en.wikipedia.org/wiki/MPrime

[wcs]

The wikipedia page for Prime95 (the windows version of mprime) has some more information.

http://en.wikipedia.org/wiki/Prime95

I've just downloaded mprime and tried it myself. The compressed archive contains a file called stress.txt that will give you more information.

For the "torture test", run it as

Code:

mprime -t

For many, many hours...

Although if your cpu really is faulty, maybe you'll find an error without having to wait for 20 hours or so.

[exeterdad]

Thanks for the responses.  I stumbled upon the wikipedia entry before I came back to see your posts.  I got mprime going and it puked in less than a minute.

So for the heck of it, I ran two instances of it since this processor has hyper-threading.  I made it thirteen minutes. Lucky me.  I wish I could figure out what is causing the problem.  At least it's confirmed that this box is junk.

[lagagnon]

In my experience, and since you have already tested the RAM I would do the following (unless your box is under warranty in which case you're lucky):

1) remove CPU fan and renew the heat sink compound. Remove the old junk with rubbing alcohol. If it is a small metal pad instead of HS compound remove the old pad with a razor knife (carefully - no knicking the fan base) and add quality thermal compound, enough to cover the CPU die in a thin smear. Remove old gunk from the CPU die also.

2) remove the CPU and reseat it - sometimes that alone may fix some problems

3) replace heat sink fan carefully and be sure to re-hook-up the CPU fan's power cord

4) retest using Prime95. If still no luck it could still possibly be the motherboard components rather than the CPU

5) if the machine does not actually lock up or spontaneously reboot then it is unlikely to be a power supply issue.

6) If you can get hold of another motherboard try the same CPU to test situation 4 above.

Good luck!

[nubcnubdo]

The Prescott P4 runs very hot, and the standard P4 cooler is really inadequate to keep it cool. May I recommend the Zalman CNPS7000B-ALCU cooler for $29.99 at newegg.

http://www.newegg.com/Product/Product.asp?Item=N82E16835118112

For the P4 cooler there is a weight limitation of 480 grams. According to my research, coolers that weigh more than 480 grams create a potentially destructive moment on the socket and processor. If the tower is dropped from a height of more than, say, 6 inches, then the weight of the cooler causes a torque on the socket and processor. The Zalman Aluminum-Copper cooler linked above weighs in at around 470 grams. In comparison, the pure copper version of this design weighs 700+ grams, and would therefore be more likely to cause damage from accidental dropping.

I am just saying the Zalman CNPS7000B-ALCU is well-designed. One advisory, this cooler is large and you must consider the exact measurements of the cooler with respect to the space around the processor socket, especially height of adjacent capacitors. Also, the unprotected, propeller-style fan is vulnerable to catastrophic breakage from ribbons and connectors, so you must be careful, and use ties to prevent accidental contacts.

Zalman CNPS7000B-ALCU (retail, no blue LED)

[Vanger]

You can try superpi for testing too.

Are you sure there is no overheating? Have you tried monitoring temp with lm_sensors?

[exeterdad]

Okay here's "sensors" output at start of mprime test.  The mprime -t output.  Then "sensors" output at mprime failure.

I made a simple bash script to fire off all three commands so the sensor outputs are immediately before and after mprime test.

During the test, temps went up to 48.5 C and held steady.  Now after test temp is back down to 34 C.

I'm not clear what the "ALARM" lines are all about?

w83697hf-isa-0290
Adapter: ISA adapter
VCore:     +1.50 V  (min =  +0.03 V, max =  +0.02 V)       ALARM 
+3.3V:     +3.28 V  (min =  +0.90 V, max =  +0.00 V)       ALARM 
+5V:       +5.00 V  (min =  +0.00 V, max =  +0.97 V)       ALARM 
+12V:     +11.98 V  (min = +10.28 V, max =  +0.49 V)       ALARM 
-12V:      -7.18 V  (min =  -4.38 V, max = -14.25 V)       ALARM 
-5V:       +0.48 V  (min =  -1.28 V, max =  -4.49 V)       ALARM 
V5SB:      +5.54 V  (min =  +0.00 V, max =  +0.00 V)       ALARM 
VBat:      +3.22 V  (min =  +0.51 V, max =  +0.00 V)       ALARM 
fan1:     2556 RPM  (min = 10546 RPM, div = 16)              ALARM 
fan2:     2721 RPM  (min =  585 RPM, div = 16)                     
temp1:       +34°C  (high =    +0°C, hyst =   +99°C)   sensor = thermistor           
temp2:     +32.0°C  (high =   +70°C, hyst =   +65°C)   sensor = diode           
alarms:   
beep_enable:
          Sound alarm enabled


Beginning a continuous self-test to check your computer.
Please read stress.txt.  Hit ^C to end this test.
Test 1, 4000 Lucas-Lehmer iterations of M19922945 using 1024K FFT length.
Test 2, 4000 Lucas-Lehmer iterations of M19922943 using 1024K FFT length.
Test 3, 4000 Lucas-Lehmer iterations of M19374367 using 1024K FFT length.
Test 4, 4000 Lucas-Lehmer iterations of M19174369 using 1024K FFT length.
Test 5, 4000 Lucas-Lehmer iterations of M18874369 using 1024K FFT length.
Test 6, 4000 Lucas-Lehmer iterations of M18874367 using 1024K FFT length.
Self-test 1024K passed!
Test 1, 800000 Lucas-Lehmer iterations of M172031 using 8K FFT length.
Test 2, 800000 Lucas-Lehmer iterations of M163839 using 8K FFT length.
Test 3, 800000 Lucas-Lehmer iterations of M159745 using 8K FFT length.
Test 4, 800000 Lucas-Lehmer iterations of M157695 using 8K FFT length.
Test 5, 800000 Lucas-Lehmer iterations of M155649 using 8K FFT length.
Self-test 8K passed!
Test 1, 560000 Lucas-Lehmer iterations of M212991 using 10K FFT length.
Test 2, 560000 Lucas-Lehmer iterations of M210415 using 10K FFT length.
Test 3, 560000 Lucas-Lehmer iterations of M208897 using 10K FFT length.
Test 4, 560000 Lucas-Lehmer iterations of M204799 using 10K FFT length.
Test 5, 560000 Lucas-Lehmer iterations of M200705 using 10K FFT length.
Test 6, 560000 Lucas-Lehmer iterations of M196607 using 10K FFT length.
Self-test 10K passed!
Test 1, 4500 Lucas-Lehmer iterations of M17432577 using 896K FFT length.
Test 2, 4500 Lucas-Lehmer iterations of M17432575 using 896K FFT length.
Test 3, 4500 Lucas-Lehmer iterations of M17115073 using 896K FFT length.
Test 4, 4500 Lucas-Lehmer iterations of M16815071 using 896K FFT length.
Test 5, 4500 Lucas-Lehmer iterations of M16515073 using 896K FFT length.
Self-test 896K passed!
Test 1, 5300 Lucas-Lehmer iterations of M14942209 using 768K FFT length.
Test 2, 5300 Lucas-Lehmer iterations of M14942207 using 768K FFT length.
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Torture Test ran 1 hours, 8 minutes - 1 errors, 0 warnings.


w83697hf-isa-0290
Adapter: ISA adapter
VCore:     +1.46 V  (min =  +0.03 V, max =  +0.02 V)       ALARM 
+3.3V:     +3.26 V  (min =  +0.90 V, max =  +0.00 V)       ALARM 
+5V:       +5.00 V  (min =  +0.00 V, max =  +0.97 V)       ALARM 
+12V:     +11.92 V  (min = +10.28 V, max =  +0.49 V)       ALARM 
-12V:      -7.18 V  (min =  -4.38 V, max = -14.25 V)       ALARM 
-5V:       +0.43 V  (min =  -1.28 V, max =  -4.49 V)       ALARM 
V5SB:      +5.51 V  (min =  +0.00 V, max =  +0.00 V)       ALARM 
VBat:      +3.18 V  (min =  +0.51 V, max =  +0.00 V)       ALARM 
fan1:     3375 RPM  (min = 10546 RPM, div = 16)              ALARM 
fan2:     2721 RPM  (min =  585 RPM, div = 16)                     
temp1:       +38°C  (high =    +0°C, hyst =   +99°C)   sensor = thermistor           
temp2:     +48.5°C  (high =   +70°C, hyst =   +65°C)   sensor = diode           
alarms:   
beep_enable:
          Sound alarm enabled

I really wasn't concerned about the cpu cooler so much.  It's a Intel cooler from my 3 ghz P4 Prescott. I don't overclock or game.  I figured Intel should be able to keep their own stuff cool under normal conditions.  Let's hope?

About the warranty?  LOL!  This is a homebuilt box made from salvaged parts from other (mine) homebuilt box parts.  I believe the cpu is a 2 or 3 years old by now.

I'd love to throw the 3 ghz. on this board, but for some reason it won't even post.  I don't have another motherboard that will accept either processor.  Well I have one, but it can't be trusted.  It has some visible physical damage from my 4 year old standing in my open computer that was left unattended for a moment.  No....  this cpu wasn't mounted in it.

[lagagnon]

exeterdad: your CPU temperature seems fine to me so it does not appear to be an overheating issue - thus replacing the fan or CPU heatsink compound will probably not make any difference.

Your Power Supply voltages appear OK except for the -12V line which is abnormally low, but then the -12v line is not really used much anymore so I am suspecting that your PS is OK too. Also, you are not getting reboots or lock-ups during the actual test - that fact alone tends to tell us that it is probably not an overheating or power supply problem.

Thus you probably have a faulty CPU or motherboard. It is difficult to say which, although as you are running a math program (rather than say a video or games program) and it is failing, I do suspect you have a dud CPU.

[Joe1962]

A bit late, but you could also try this:
http://users.bigpond.net.au/cpuburn/

[Vanger]

ALARM means, that your normal readings don't get into absurd borders, set by lm_sensors.

So, you say other, working (have you tested your other processor?) CPU does not work on this mb.
Ok, can you examine your motherboard physically? Does it have any damage?

And can you borrow a perfectly working mb to finally localise the problem? Or go with your processor (expecially with 3Ghz one) to test them?
Looks like it's your motherboard that fails.

[exeterdad]

Thus you probably have a faulty CPU or motherboard. It is difficult to say which, although as you are running a math program (rather than say a video or games program) and it is failing, I do suspect you have a dud CPU.

I am completely comfortable with your diagnosis.  And I appreciate the help all of you have given me on this matter.

Now I'm at the crossroads.  I don't think I will replace the 2.4 ghz.  I have a known working 3 ghz.  In fact, barely used. I think dollar for dollar, a motherboard that the 3 ghz. will run on would be the best route.  Not that socket 478 motherboards are plentiful these days.

[exeterdad]

ALARM means, that your normal readings don't get into absurd borders, set by lm_sensors.

So, you say other, working (have you tested your other processor?) CPU does not work on this mb.
Ok, can you examine your motherboard physically? Does it have any damage?

And can you borrow a perfectly working mb to finally localise the problem? Or go with your processor (expecially with 3Ghz one) to test them?
Looks like it's your motherboard that fails.

The 3 ghz was working great on the motherboard my son stood on.  Even before that happened.  The 3 ghz. didn't work on this board (Abit VI7).  I could never get it to post.

As far as I can see, even with my magnifying lamp (or whatever it's called), I can't see any physical damage to this board (with the CPU in question).  Is this what you were asking?