Prime95 crashing my computer :( Help!

YamahaAlex37

[H]ard|Gawd
Joined
Jun 23, 2005
Messages
1,775
I am building a computer for my uncle, and the computer seems to lock up after a few hours. Last night I left Prime95 running and when I woke up in the morning the computer was frozen. Now I tried to run it again and the computer rebooted after a few minutes. Please help me solve this problem, I'd really like to ship it out today and I can't send my uncle a problematic computer :(

All the parts are brand new: Athlon II X3 440, Biostar MCP6P3, GeIL Value PLUS 4GB DDR3 1600, SAPPHIRE Radeon HD 4650 512MB 128-bit GDDR2. All this is housed in this Logisys case with the 480W PSU:
http://www.newegg.com/Product/Produ...044&cm_re=logisys_case-_-11-148-044-_-Product

Would the cheap PSU most likely be the problem? I have a more powerful PSU I could test with. Or should I re-seat the CPU / Ram?

Thanks for any help!
 
Did you change any setting in the BIOS? And that question definitely includes memory settings since running any current CPU with 1600 MHz ram today is overclocking the ram controller since no CPU from Intel or AMD officially supports ram that is that fast.
 
Sorry, the Athlon X3 440 was actually pulled from my computer, but it was working perfect OC'd at 3.3Ghz. Since I put it in this computer, I have not touched a single BIOS setting. Should I be down-clocking the RAM?

Temps are 22c idle, and 40c during Prime95.

If the PSU boots the computer fine, and gets through 3dmark06 fine, is it most likely this is a CPU/Ram issue?
 
Do you have a known good power supply to test with? Have a multimeter handy? Does the motherboard display voltages in the bios? You could also use a software program to monitor voltages, speedfan and everest are able to. The reviews page at newegg is full of people complaining about the power supply so i wouldn't be surprised. You also need to check the voltages under load, which you cant do from the bios. after the psu is ruled out then check temperatures, and run memtest.
 
Do you have a known good power supply to test with? Have a multimeter handy? Does the motherboard display voltages in the bios? You could also use a software program to monitor voltages, speedfan and everest are able to. The reviews page at newegg is full of people complaining about the power supply so i wouldn't be surprised. You also need to check the voltages under load, which you cant do from the bios. after the psu is ruled out then check temperatures, and run memtest.

I do have a known good PSU to test, but no multimeter. I got speedfan installed and I'm running Prime95 again, what do I need to let you guys know?
 
Need to know the voltage fluctuation. What is it at when you start the pc, and does it vary with stress or over time. Need to know voltage values, for 12v, 5v, and 3.3v or whichever ones are listed. if it varies much, then you should try the other power supply.
 
Need to know the voltage fluctuation. What is it at when you start the pc, and does it vary with stress or over time. Need to know voltage values, for 12v, 5v, and 3.3v or whichever ones are listed. if it varies much, then you should try the other power supply.

I'm running Prime95 right now so I'll check later what it is at start, but so far those 3 seem to stay very straight. I'm surprised Prime95 has been running for at least 20 minutes, and last time it crashed in a few. Still, I really don't want to send my uncle a computer that may freeze up on him because it's being shipped across the country, and it since he has no one over there to work on computers, he would have to ship it all the way back if it has issues.
 
If you don't want to provide much support, the first component you should start with in the future is a decent power supply. I have replaced many cheap power supplies, and sometimes all the other components due to the power supply taking everything with it. If voltages stay steady, I'd open bios and load setup defaults and then run memtest.
 
So 30 minutes into Prime 95 and this time it was a blue screen, with a physical memory dump... does this indicate the problem is most likely RAM? I'll go ahead and run the memtest.
 
Voltages stay pretty stable? Were temps reasonable? Memtest should turn up errors within an hour if you are crashing that fast. Make sure everything in bios is set correctly, or at least set to defaults. Check what voltage is required for ram. A newer bios may also improve memory compatibility.
 
Voltages stay pretty stable? Were temps reasonable? Memtest should turn up errors within an hour if you are crashing that fast. Make sure everything in bios is set correctly, or at least set to defaults. Check what voltage is required for ram. A newer bios may also improve memory compatibility.

I should have checked my bios settings beforehand, but I'm already running memtest. I'll let you know if it crashes, and make sure the voltage is correct for the ram afterward.

Thanks a lot for helping.
 
Memtest has been running for 40 minutes, and it says Pass Complete, no errors, press Esc to exit. Pass reads 1 and Errors read 0. The Pass percentage is still at 39% and climbing however. Is it safe to exit? Does the ram seem ok?

I'm wondering if I should just go ahead and switch the PSU to be safe and run Prim95 again.
 
sure, you can also try to specify test number 7 under memtest, usually finds errors fairly quickly for me. I would check bios, then try prime again if you changed anything. Otherwise try the other power supply. Just wondering, When you swapped the cpu you cleaned off the old thermal goop and put on new right? guess that goes along with temperatures.
 
sure, you can also try to specify test number 7 under memtest, usually finds errors fairly quickly for me. I would check bios, then try prime again if you changed anything. Otherwise try the other power supply. Just wondering, When you swapped the cpu you cleaned off the old thermal goop and put on new right? guess that goes along with temperatures.

ya I cleaned the paste and put new, and it idles 20-22, so should that should mean I did alright?
 
idle temps sound good. diddn't see you mention any thing specific before so i had to ask.

edit sorry.. i see now.. not as awake as i should be. Temps look fine.
 
Is the chipset heatsink seated properly? Does it wiggle(Standard disclaimer: turn off the pc and disconnect the psu, make sure you have discharged any static)? Could also be an error with speedfan. also check your video card heatsink?
 
Is the chipset heatsink seated properly? Does it wiggle(Standard disclaimer: turn off the pc and disconnect the psu, make sure you have discharged any static)? Could also be an error with speedfan. also check your video card heatsink?

Will do, also when I turn on my computer my Memory Voltage reads 1.584. This is what I see in CPU-Z however (not sure what this means):


Uploaded with ImageShack.us

Recommended for my memory is 1.6.

My vcore2 reads 1.2. NB Frequency reads 1607.4Mhz. Just giving any information I think that might help. Thanks.
 
voltage seems within spec, no memtest errors either. Do the voltages drop when prime is running? You can also try with one stick of ram removed, and use the onboard video. Try the other power supply if it isn't too much trouble.
 
Last edited:
Thanks so much Malecoda. Currently running Prime95 with the 600W PSU. Hopefully I won't be replying for another few hours *fingers crossed* (with good news)
 
I really wish I hadn't done this, but I'm building two computers at once here, so I was switching the vga, keyboard, and mouse back and forth between the computers, and when I hooked it back up to this one to see how Prime was going, the moving or pressing the keyboard/mouse wasn't bringing up anything, just stuck at a black screen. I'm assuming this means it froze :(
 
hmm, ps2 keyboard and mouse? If so that is a no-no iirc. Restart the pc? Come back up?
 
Ya restarted and it came back up. Mouse is usb but keyboard is ps2. I guess all I can do is leave it overnight and check in the morning?
 
sounds fine, make sure to turn the pc off before changing ps2 connected stuff. I have had bad luck doing it myself in the past.
 
Damnit! already got a blue screen physical memory dump with the new PSU and keeping keyboard/mouse attatched :(

Does the physical memory dump necessarily eliminate the CPU as the problem? The CPU was working fine in my gaming computer before this, however the memory is brand new (so previously untested).
 
hmm, Set defaults in bios? try taking out a stick of ram, and running on the onboard video? Any other cards installed?
 
nah, dumps memory for any error. could still be anything. You can also try underclocking the cpu to check ram, and underclocking ram to check cpu. may also try running memtest overnight. good luck, i may be on tomorrow or someone else will hopefully chime in, exams tomorrow ugh.

also you mentioned your ht link was at 1600mhz? that board only supports a 1000mhz ht(as per newegg/motherboard manual specs). check bios settings again.

Did you use the drivers off of the cd for vga/motherboard? make sure you get the newest drivers from the website instead.
 
Last edited:
nah, dumps memory for any error. could still be anything. You can also try underclocking the cpu to check ram, and underclocking ram to check cpu. may also try running memtest overnight. good luck, i may be on tomorrow or someone else will hopefully chime in, exams tomorrow ugh.

also you mentioned your ht link was at 1600mhz? that board only supports a 1000mhz ht(as per newegg/motherboard manual specs). check bios settings again.

Did you use the drivers off of the cd for vga/motherboard? make sure you get the newest drivers from the website instead.

I did use all drivers from Biostar's website. Right now I'm running Prime95 with the memory voltage up .05, so it's 1.632 instead of 1.582... so far so good.

So since the board only supports 1000mhz I need to turn a setting down? Right now in CPU-Z NB Frequency reads 1607.3 Mhz
 
That is another troubleshooting step. If the increased voltage doesn't work i'd drop the ht multi to put it within spec. good luck i have to be up in a few hours for exam. sorry i cant be of more assistance.
 
You've already helped so much Malecoda! I wish you could rep in this forum...

Anywhere after upping the memory voltage, I am experiencing something new. This is a tri-core CPU, and Worker #3 in Prime 95 reads FATAL ERROR... Worker stopped.

It looks like the other two workers are still going...
 
mmm.. exam.. anyway.. try lowering the ht multiplier. You should have options for HT link speed in bios. Drop it to the lowest one for testing. Are you running blend or one of the other tests in prime? Try running small fft to test the cpu, and blend for overall stability. After lowering the ht link and running small fft, you may also need to set the cpu voltage manually, 1.425v is the max for that processor, so i would try no higher than 1.35, but you may have more experience with that chip having overclocked it.
 
Heyy great news!!! Thanks so much for all your troubleshooting advice Malecoda, you were right about the thermal paste/steating of the CPU I think. I decided to re-apply the paste, and I"m not sure if this happened while I was removing the CPU, but it seemed to not come out nice and when it did, I noticed one of the pins was bent. Maybe I put it in wrong last time.

Anyway I straightened the pin, re-applied the paste, and got through the entire night with Prime 95 :D
 
glad you got it fixed.

stability can be hard to get. Ram causes the most problems when not clocked properly or timings set properly.

Power supplies cause problems to if the voltage drops on them expectingly.

Some times its just the motherboard (usually cooling related, gets to hot over continued full load), or a add in card, such as a sound card, network card, etc. Cpu can cause problems if over-clocked.

Heat causes problems so always make sure the Pc has enough air flow that the ambient temp doesn't get terribly high.

I had a friend i build't a pc for, he said it kept crashing. I went to his house only to find the pc stuffed in a wooden cabinet, with no air holes, other than a hole big enough to get the power cord and other wires through. It was simply overheating. I busted out the handy Jig Saw and opened up the back a good bit, no more problems.
 
Back
Top