Problems with AMD Processors?

Pope On A Rope

Pope On A Rope

CAGiversary!
Feedback
21 (100%)
Has anybody had any problems with PCs with AMD processors resetting?
I've had 3 computers (1 Duron, 1 Athlon, 1 Athlon Thunderbird) all with this problem; out of nowhere they will reset. It doesn't matter whether the CPU usage is high or low, the temperature of the processors is fine, the machines will do this after a hard drive wipe on a fresh install of Windows or with plenty of software loaded, and there don't appear to be any RAM issues. I'm stumped. Any ideas or similar frustrating experiences?
 
I have never had any issues with AMD processors. Did you check the RAM using that one program that checks it for bad sectors? And how big is you power supply? I have learned that in rebooting cases where everyting seems fine that it tends to be the PSU.
 
First you should disable the automatic boot feature to see the STOP error message when the systems crash. Right now you have it set to automatically reboot which is why they "reset". Press Windowskey + Break, click Advanced tab, click Settings under the Startup and Recovery heading, uncheck Automatically Restart.

There are two main reasons for AMD systems to crash: Heat and chipset/RAM errors. Heat is the #1 enemy as AMD processors run extremely hot. They are designed to withstand up to 80C, but once you start approaching 55C the chance of a crash is likely and anything 60C or up will undoubtedly cause a crash. Cooling of the CPU and chipset is critical.

You should also determine the chipset your motherboard uses and install the latest chipset drivers. There are several chipsets out there that are just error prone (older VIA in particular) and no drivers will completely alleviate the problem, but having the latest drivers is the best you can do.

AMD chipsets appear to be very sensitive to RAM timings and the quality of RAM. I suggest running memtest for a few hours to detect any RAM errors. I downloaded and burned the memtest ISO which makes it easy to test any system. Simply insert the CD and boot from it. Also, make sure you are using the RAM's default SPD timings (set this in the computer's BIOS).

Also, as was mentioned, AMD systems consume a lot of power (hence the high temperatures) and should have relatively large power supplies.
 
My older PC would randomly shut down. I clean it out completely and reseated everything and re wired it. It worked great for about a week but then it died again and smelled really bad. I went to clean it again and this time I took apart the power supply. The fan was locked solid and caked with dirt.

Did you have 3 PCs at the same time or did you upgrade one a lot? Were you using the same PSU?
 
Did you build them yourself? If so, it sounds like you tend to either provide inadequate cooling or inadequate power.
 
Sorry for the slow response, went out of town for a few days.
Two of the computers are shop-bulit and I built the third myself. The one I built has a 300w power supply and adequate cooling; the mobo and processor both run very cool. Three physically seperate machines.

Thanks Zing and dtarasev for the RAM and Windows settings tips, hadn't run memtest before, and it found errors on one of my machines. Everything seems to be running well (except for the machine I need new RAM for, which it's at least good to know what was wrong) now, thanks everyone. :)
 
Update:
Getting the same error message again and having the computer reset again; problems with sysdata.xml and the minidump. Have tried numerous "solutions", anybody found a way to fix this?
 
[quote name='Zing']
There are two main reasons for AMD systems to crash: Heat and chipset/RAM errors. Heat is the #1 enemy as AMD processors run extremely hot. They are designed to withstand up to 80C, but once you start approaching 55C the chance of a crash is likely and anything 60C or up will undoubtedly cause a crash. Cooling of the CPU and chipset is critical.[/QUOTE]

Just wanted to say that this is no longer the case. This may have been true a few years back but in the past couple of years it has flipped, Intel is the one with the heat issues now. As I type this my CPU is at 54c and that is an idle temperature, when I start playing a game like Doom 3 it jumps in to the mid 60s.
 
[quote name='Pope On A Rope']Update:
Getting the same error message again and having the computer reset again; problems with sysdata.xml and the minidump. Have tried numerous "solutions", anybody found a way to fix this?[/QUOTE]

I checked around and it looks like those are caused by bad RAM. I would take all the RAM out and use RAM from a working PC.

Also try askin around on this site : computing.net

I go there whenever I have some terrible problem.
 
[quote name='dtarasev']Just wanted to say that this is no longer the case. This may have been true a few years back but in the past couple of years it has flipped, Intel is the one with the heat issues now. As I type this my CPU is at 54c and that is an idle temperature, when I start playing a game like Doom 3 it jumps in to the mid 60s.[/QUOTE]
agreed, intels are currently the heat champs.......we had issues similar to this at work.....ended up being the ram. replaced with a better quality stick and all is now fine.
 
It's probably RAM or your Power Supply. Do you know what brand power supply you have? 300w is a little low for some of the newer CPU's, especially if you have a decent video card installed. If the psu came with the case, it's probably a cheap one. The included ones may have a high wattage, but it's likely most of that power is on the 3.3 volt rail that you're not actually going to use for much more than PCI cards. You want a supply with a good rating on the 12v rail, or dual 12v rails, as that's where your drive motors and CPU / Motherboard pull their power from.

Still, if you're getting error messages, it's probably not that. Most Power issue's I've experienced don't give you error messages, they just reboot the system at random. I have heard some low power issues can cause random instabilities that are hard to narrow down.

In your case this likely may be the RAM. If you have multiple sticks, pull them both, blow out the ram slots with canned air. Then try running the machine with one stick or the other, and see if the problem goes away. Run Memtest or another memory benchmark on the system with each stick of RAM seperately. I did this on my in-law's machine and it turned out one of their DIMMs had died of old age.
 
I had this problem before. I cleaned out the PC got rid of all of the dust. It really helps. However that was not the problem. The problem seemed to be that my CPU was overheating. It finally died after a few months.
 
[quote name='dtarasev']Just wanted to say that this is no longer the case. This may have been true a few years back but in the past couple of years it has flipped, Intel is the one with the heat issues now. As I type this my CPU is at 54c and that is an idle temperature, when I start playing a game like Doom 3 it jumps in to the mid 60s.[/QUOTE]


Beat me to it, P4s are the new Thunderbird Socket As. Watch your fingers.

It's also probably NOT the ram. RAM would give you a hard lock in most cases. Reboot is PS!

There I said it.
 
Oooh, I just thought of something else. I used to have a video card that did this. When it got too hot, the card would lock up and when it locked up, it locked up the whole system via the AGP bus.
I added an aftermarket fan to replace the stock one and it fixed the problem temporarily until I could get a better card.

What card do you have in there?
 
Power Supply is insufficient. 300W is BARE minimum for an AMD/Intel proc nowadays. If you are using a '300W' powersupply that came with your case, it most likely not a true 300W power supply. Most power ratings are peak power & not RMS (read:average) power. Plus as the PC/Power Supply heat up, the power efficiency decreases quite drastically.
 
You're running the BARE minimum for a powersupply. 300W does not cut the mustard nowadays. 450W is the baseline for stability. What most people don;t realize is that the '300W's is the peak power rating or RMS rating in LOWER temperature enviroments. As the PC/Powersupply heats up, power efficiency drops signifigantly meaning more power is needed to maintain the same throughput.

EDIT: Double post. My bad
 
You must log in or register to reply here.
bread's done
Back
Top