• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Prime95 giving errors

S

SRabbelier

Junior Member
Hiya,

I've been having some CPU-troubles, and I thought I worked everything out now (got a brand-new CPU even), so I decided to try some tests.
The first one, Prime95, came up with these errors:
[Mon Mar 13 22:09:11 2006]
FATAL ERROR: Resulting sum was 2.249374916780728e+016, expected: 2.244170472289848e+016
Hardware failure detected, consult stress.txt file.
[Mon Mar 13 22:14:59 2006]
FATAL ERROR: Resulting sum was 429310920838547.8, expected: 488321881137236.3
Hardware failure detected, consult stress.txt file.
[Mon Mar 13 22:33:54 2006]
FATAL ERROR: Resulting sum was 9211895489460844, expected: 9172738711806656
Hardware failure detected, consult stress.txt file.
Click to expand...

My hardware (summary from SiSoft Sandra):
Mainboard
Manufacturer : ASUSTeK Computer INC.
MP Support : No
Model : A7V8X-X
Version : REV 1.xx
Serial Number : xxxxxxxxxxx
System BIOS : 06/23/2005-VT8377/VT8235-A7V8X-X

System Memory Controller
Location : Mainboard
Error Correction Capability : None
Number of Memory Slots : 3
Maximum Installable Memory : 3GB
Bank 1 - DDR 1 : DIMM Synchronous DRAM 256MB/64
Bank 2 - DDR 2 : DIMM Synchronous DRAM 256MB/64
Bank 3 - DDR 3 : Empty

Processor
Model : AMD Sempron(TM) 2800+
Speed : 1.99GHz
Model Number : 2800 (estimated)
Performance Rating : PR2883 (estimated)

Video Adapter
Model : NVIDIA GeForce FX 5200
Chipset : GeForce FX 5200
RAMDAC : Integrated RAMDAC
Video BIOS : Version 4.34.20.42.00
Click to expand...

I'm not overclocking (at least, not untill everything works 100%).

Where should I start looking?
What benchmarks / tests should I do?
If this is not the correct place to ask, where should I ask?

Any help would be greatly appreciated,
Thanks in advance,

Sverre Rabbelier.
 
Your CPU or RAM is unstable.

Try running with one DIMM at a time to see if you can isolate the problem to one of them. If both seem OK by themselves, try one at a time in different memory slots, to see if maybe the problem is with one of the slots in the MB.

If both seem to fail, try underclocking the CPU. If that fixes it, the CPU is not working properly.
 
Try a torture test with small FFT's to try and limit any problems your RAM might be causing.

Also, check your temps as you run it to make sure you're not getting any overheating. A 2800+ should run <40c I would think. Prime95 with Large FFT's (creating max heat) doesn't push my Athlon64 3000+ over 35c.
 
Looks like you're not overclocking, so here's the three most probable causes.

1. Bad CPU
2. Bad RAM
3. Overheating CPU

1. Very doubtful, but possible. If #2 and #3 don't turn out, it's probably a bad CPU.

2. Use Memtest86 to test ram seperately: http://www.memtest86.com/

3. While running any windows based benchmarking/stress tools, make sure you keep an eye on your temperatures. Speedfan is good for this: http://www.almico.com/speedfan.php

-z
 
Talking about speed-reply!
I'll try to isolate the problem, although I had a memtest on my memmory only yesterday, although I didn't see the results myself..

I'm sortof hoping it's my memory, it'd be cheaper to repair, and I'm in need for a good excuse to upgrade it 😛.
 
Where do I find my vdimm setting? In the BIOS I'd guess somewhere?

Motherboard Monitor 5 tells me this about my CPU temps:
|Sensor~|~Current~|~Low~~|~High~~|~Average~|
|CPU~~~|~73° C~~|~65° C~|~84° C~|~76° C~~~|

I think that timeperiod included Prime95's torture test.
It's running at 54c currently, with Prime95 running it's self-test, having removed 1 RAM-block.
 
As said, I removed one RAM-block, and got these results (it's still running):

[Tue Mar 14 09:16:09 2006]
Self-test 768K passed!
[Tue Mar 14 13:12:35 2006]
Iteration: 198119/13820117, ERROR: SUM(INPUTS) != SUM(OUTPUTS), 7845155585041819 != 7808694244350789
Possible hardware failure, consult the readme.txt file.
Continuing from last save file.
Click to expand...

Can I conclude it's not this RAM-block, or perhaps it's "less damadged" than the other one?
Should I swap them over and see what the other one gives, or should I let Prime95 run longer?
 
Originally posted by: SRabbelier
Where do I find my vdimm setting? In the BIOS I'd guess somewhere?

Motherboard Monitor 5 tells me this about my CPU temps:
|Sensor~|~Current~|~Low~~|~High~~|~Average~|
|CPU~~~|~73° C~~|~65° C~|~84° C~|~76° C~~~|

I think that timeperiod included Prime95's torture test.
It's running at 54c currently, with Prime95 running it's self-test, having removed 1 RAM-block.
Click to expand...

Either MBM5 is borked, or your CPU definitely IS overheating!
 
start of with RAM
then CPU factor

at least eliminate one module at a time and u will find out which part is faulty
 
Originally posted by: StrangerGuy
Either MBM5 is borked, or your CPU definitely IS overheating!
Click to expand...

To be more precise, it wasn't MBM5, but the log-setting in it. I apparently disabled it, so I was looking at last-years results, at the time I indeed was having overheating problems.
Currently, these are my actual results:

+-------------------------------------------------------------------------+
|Total number of readouts: 2848 CPU Speed: 1987 MHz |
|Running from: 14-3-2006 8:19:25 until: 14-3-2006 16:12:46 |
+-----------------------------+----------+----------+----------+----------+
|Sensor | Current | Low | High | Average |
+-----------------------------+----------+----------+----------+----------+
|Case | 31° C | 25° C | 31° C | 29° C |
|CPU | 59° C | 48° C | 59° C | 57° C |
|Core 0 | 1,78 V | 1,76 V | 1,78 V | 1,78 V |
|Core 1 | 0,00 V | 0,00 V | 0,00 V | 0,00 V |
|+3.3 | 3,28 V | 3,25 V | 3,31 V | 3,28 V |
|+5.00 | 4,64 V | 4,61 V | 4,82 V | 4,65 V |
|+12.00 | 12,95 V | 12,69 V | 13,01 V | 12,96 V |
|-12.00 | 0,00 V | 0,00 V | 0,00 V | 0,00 V |
|-5.00 | 0,00 V | 0,00 V | 0,00 V | 0,00 V |
|Fan 1 | 3552 RPM | 3479 RPM | 3590 RPM | 3551 RPM |
|Fan 2 | 0 RPM | 0 RPM | 0 RPM | 0 RPM |
|Fan 3 | 0 RPM | 0 RPM | 0 RPM | 0 RPM |
|CPU0 | 0,00 % | 0,00 % | 0,00 % | 0,00 % |
+-----------------------------+----------+----------+----------+----------+
Click to expand...

The Torture test is currently running, and tells me (up till now):
[Tue Mar 14 14:50:04 2006]
Self-test 1024K passed!
[Tue Mar 14 15:10:44 2006]
Self-test 896K passed!
[Tue Mar 14 15:31:25 2006]
Self-test 768K passed!
[Tue Mar 14 15:46:30 2006]
Self-test 640K passed!
Click to expand...

So I'm wondering what to do with
[Tue Mar 14 13:12:35 2006]
Iteration: 198119/13820117, ERROR: SUM(INPUTS) != SUM(OUTPUTS), 7845155585041819 != 7808694244350789
Possible hardware failure, consult the readme.txt file.
Click to expand...
 
Have you run Memtest yet? That's THE best way to confirm/eliminate a RAM problem.

Now if those temps are correct, you have problems. 84C is waaaay too hot and would definitely be causing errors.
Did you happen to install the HSF and not add any thermal paste?
 
Originally posted by: Ike0069
Have you run Memtest yet? That's THE best way to confirm/eliminate a RAM problem.

Now if those temps are correct, you have problems. 84C is waaaay too hot and would definitely be causing errors.
Did you happen to install the HSF and not add any thermal paste?
Click to expand...

Lol, when reading your post I though "Huh? Didn't he read my reply?", *mutters*, I probably copied the wrong log or something :S, anyway, the _proper_ log is in place now.
Is 59 degrees to high, while running Stress test and working normally?
 
I've just downloaded the image, planning to burn it and run a test in a hour or so.
How many runs should I do?

While Prime95 stopped error-ing (on this module) I think I'll run the test on both modules (one at a time), just to be sure.
Also, I'm thinking about buying a new cooler, but I'm not sure if I really need one... do I?
 
Hopefully you can isolate the problem using the methods already described, but I'll throw another possibility into the mix.. the motherboard. My last system was a P4-2.53 using a Soyo Motherboard. I had 3 different sets of ram in it and could never get it 100% stable when I would try and run my ram at rated speeds. It would fail MemTest and Prime95. I had to run my PC2700 RAM at PC2100 speeds to be stable. I bought a new motherboard (Shuttle), put all the same components into it, re-installed my windows, and the system is rock solid.. even with a little overclock.

I see you have an ASUS motherboard, so you're probably safe. Every ASUS I've owned has been solid.
 
Originally posted by: SRabbelier
I've just downloaded the image, planning to burn it and run a test in a hour or so.
How many runs should I do?
Click to expand...

Let it run overnight. Try to get about 12hrs of testing done.

About your vdimm (voltage to ram) setting, go into BIOS and have a look around. Likely there will be "PC Health" or similar section that will show what the vdimm setting is. If you have the manual for the motherboard, check it out. It will tell you what section of the BIOS to goto to change the vdimm if needed.

IIRC, MBM5 will also show the vdimm

BTW: I've heard the Asus skt A mobo would undervolt ram a bit, and overvolt the CPU a bit (opposite of an Abit like mine). That's why I keep mentioning the vdimm

Fern
 
I scouted the BIOS and MBM5, but I could nowhere find vdimm.
All I found in the bios was:
VCORE 1.71V
+3.3V 3.28V
+5V 4.83V
+12V 12.54V

MDM5 results are in a previous post...

I ran memtest86 for some seven cycles on both modules, apart, and once on them together, with no errors.
I'll be running it on both tonight, but I wanted to see what Prime95'd say.
Currently, it's just started on "Test 2" on a continuous self-test.

What should I do if it stops giving errors now... (while having changed nothing at all)?
 
Memtest need to run thru 8 individual tests for one complete pass. I would let it run thru two complete passes with both sticks installed. If no errors are seen, the problem is most likely not the RAM.

If you get no errors in Memtest or Prime, then just go on as normal unless you are getting random BSOD's or restarts or other problems.
 
I've done 7 complete passes for both separate, and one pass with the two apart.
Judging from "GO SP2004", I think it's overheating, you can see the temperature going up, gradually.

From 50c to 60c in 20 minutes, with
Test 1, 800000 Lucas-Lehmer iterations of M172031 using 8K FFT length.
Click to expand...

After starting with
Test 2, 560000 Lucas-Lehmer iterations of M210415 using 10K FFT length.
Click to expand...
it started declining, even more gradually.

How much would you give me on the "get-a-new-cooler" scale (from 1 to 10)?


[EDIT]
According to the guys I had my pc repaired at it should run at 1.75V...
Where could this be double checked?
[/EDIT]
 
Yis saying I'm cooking my CPU? 0.o
Strange thing is, I didn't set it to this, the BIOS, when set to "auto", automagicly picks 1.75V...?

[EDIT]
And sorry, what's a "barton"? 0.o
[/EDIT]
 
Originally posted by: SRabbelier

[EDIT]
According to the guys I had my pc repaired at it should run at 1.75V...
Where could this be double checked?
[/EDIT]
Click to expand...

Socket A semprons stock vcore is 1.60v. See HERE. Some skt A athlons would exhibit instability with temps at 60c and above.

Fern
 
Can your computer discover the next mersenne prime just by stress testing?

Or is it just like running on a treadmill, using power but going nowhere? 😉
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top