Is Prime95 GPU intensive?
 

This is actually a larger question than this, but bear with me, haha.


About five months ago, I put together my first custom-built PC. Everything worked perfectly....for three days. And then it started hanging during games, during video rendering, and during Prime95. The hanging was a total PC hang, not an app crash. The screen froze and the speakers emitted a horrible EERRDDDKKKSSSDD. It was pretty bad.

So I sent back the ram, thinking that was the problem. The new ram didn't fix anything. Same with the PSU. And, most recently, same with the CPU. I had really hoped it was the CPU...everything seemed to lead to it. But it was not. That really leads me with the motherboard and the GPU (it's not a sofware problem and it's not harddrive). So my idea was that it MUST be the motherboard, because the hang occurs during Prime95, and Prime95 shouldn't touch the GPU because it's a CPU stress test.


But.....maybe it does touch the GPU? If it does, then my argument is invalid, and my problem could be either the mobo OR the GPU. In fact, the GPU is probably most likely in that case, since it's hanging during gaming and video rendering, which are very GPU intensive.


So my question is....is Prime95 intensive on the GPU as well?

Another question is...is there a way I can stress test the PC without touching the GPU? If I can get it to hang while the GPU is just idling, then I'll know it's the mobo, because it's not the GPU. Does this all make sense?

Thanks so much for any help I might get.

The short answer, which doesn't address the root question of "why is there this slowdown", is: no, Prime95 doesn't use the GPU.

Prime95 is *very* non GPU-intensive; it likes to send text messages to the terminal (Linux) or a cute dialog window (Windoze). It will run without graphics at all, if your OS supports it. But it will run your CPU hard.

If you want to test the GPU, look at CUDAlucas or mfaktc, which use the CUDA capabilities of the NVIDIA graphics cards. I'm not sure where msieve or NFS@home stand on this at the moment.

But, as with the unregistered gentleman named Brian in the lounge, let's start with some specifics. What CPU? CPU Cooler? Mobo? RAM? PS? What temperature is the CPU running? Can you touch all the heatsinks on the motherboard when it's hot? The RAM heatsinks? Does it care if you have the air conditioning on? (it's just started getting truly warm here in VA) And what is your overclocking situation? Does the motherboard still wake up when cold? What happens if you run memtest from an ubuntu boot CD?

And, though this will be obvious if you are in the know, is there good thermal contact between the CPU and it's cooler? Generally, this is with a white to gray grease, slightly messy, specially formulated to conduct heat reasonably well and fill all the microscopic gaps in between the top surface of the CPU and the bottom surface of the heatsink. The mounting system will place this interface under considerable mechanical pressure, generally with some kind of bending spring system. I also ASSUME your fans are turning.

Finally, I'm going through my second mersenneforum-community-assisted PC build right now, and it's left some threads here this year that might interest you, discussing component selection. The first build was rock-solid after the bent pin 997 on the CPU got straightened out.

Hey! Just noticed this seems to be a continuation of THIS thread: [url]http://mersenneforum.org/showthread.php?t=15312[/url]
In that thread, it appeared that the heatsink wasn't properly on the CPU, and there was a drop of condensation water scattered somewhere in the machine. So, how much of that advice did you follow? Has the system been sitting in a box for two months?

Thank you so much for taking interest and helping me!


It's been a long and arduous journey, haha. I've tried to explain the timeline of events so often and on so many different forum boards that I've even gotten it mixed up myself a bit :/ And a lot has happened since that last post of mine.


But the shortest story is....built a PC [URL="http://i.imgur.com/VENtk.png"](Here's the parts list again)[/URL]


It worked fine for three days. Then it started having problems.

It began hanging during games. At first it wasn't so bad, because it only happened rarely....maybe once a day. While this should have been alarming to me, I was so excited to have a "working" gaming PC that I kinda ignored it. I think what triggered it to get a million times worse was when I tried rendering something in Sony Vegas (a video editing software, if you're unfamiliar). It hung during the render, and then after that it would hang on almost every occasion that involved something more intensive than Youtube. It would hang in basically EVERY game, and Prime95 could run for maybe about 5 seconds, sometimes as much as 10 seconds, before it would hang.


My first guess was the RAM sticks. I sent those in and got them replaced. The new ones fixed nothing. My next guess was PSU....perhaps the wattage wasn't enough? Also, it was a "green" version, and it actually seemed to be made pretty cheaply. A new PSU was my only purchase so far--everything else has been sending parts back, which is free (until the warranties run out). Even though the new PSU didn't fix anything, I'm gonna keep it anyone because it's obviously higher quality (it's 650W and built much better).

Anyway, new RAM and new PSU fixed nothing. My next guess was CPU. It seemed to make sense.....I (thought) had put the heatsink on incorrectly the first time. It was still making contact with the CPU and was tightly attached, but not 100% tightly latched. My guess was the three days it worked fine, but then it started to burn itself out, and got worse over time. So I sent in the CPU, thinking that would fix everything.

It came back about a week ago. I installed it, and everything SEEMED fixed. Prime95 ran for five minutes, when I stopped it. It didn't hang. I just got impatient of waiting :P I figured that if it took 3 seconds to hang before, then if it lasted more than 4 minutes it MUST be fixed.


So I assumed it was fixed and started using it like normal. I played through an entire 9-hour game (Portal 2) on the highest settings. Absolutely no problems. I was convinced it was totally fixed. Another game, Battlefield Bad Company, worked perfectly as well. I experienced zero crashes. I should have known, however, that the same thing happened when I first built it: it worked fine for a few days, and then stopped working. Because on day 4, it hung while I was replaying Portal 2. I was really frustrated because I thought I had fixed it. I restarted and tried playing Bad Company 2. It hung there, too. It seems like there is just one moment that triggers the first hang, and then from there on out it will hang for ANYTHING.

For instance, on that first hang with the new CPU, I was playing Portal 2, like I said. There's a part where you have to break through glass. The glass particles are pretty graphically intensive (not more intensive than my PC should be able to handle, just more intensive than normal gameplay). It hung them moment the glass-breaking animation began. That was the trigger, and then after that it began hanging for everything. Prime95 began hanging within seconds again.


So that's where I am now, and I'm left with a few options.

1) Motherboard problem. I know it's NOT ram, it's NOT PSU, it's NOT CPU. That leaves motherboard and GPU. And if it hangs during Prime95 (which does not even really touch the GPU), then that only leaves motherboard. Also, the water thing that I mentioned before....I'm not sure how big of a deal that was. What happened was, I was assembling the PC in my basement....right below a pipe that was dripping every ten minutes or so :( I saw water land on the heatsink of the CPU. It didn't seem to spread, and I just wiped it right off. But it's possible it somehow caused damage somewhere, I guess.

2) GPU. But that seems unlikely considering what I said above

3) I did something wrong with the build. It's not impossible. But I think I did a pretty good job. I did a lot of research and performed the build along with a fantastic hour-long tutorial that used almost my exact same parts. I'm *pretty* sure it wasn't a build error on my part.



The CPU is definitely on correctly this time. I didn't make the mistake with the new CPU that I made the first time. I even scraped off the preapplied thermal paste from the heatsink and added some Arctic Silver. The CPU and heatsink should be good.



So that's where I am right now. Oh, and I filed the RMA for the motherboard today. So I guess I'm sending that in within the next couple of days. So this was just one long, giant post that I apologize for making you read, because in the end all I really want to know is....could it be GPU?

The worst thing in the world would be if I sent in my FIFTH component and got it back, only to find out that the motherboard was never the problem, and it was the GPU the whole time.

Obviously you can't know for sure just from reading all this. I'm not holding you accountable for my mistakes :P I'm just asking your opinion based on what information I was able to provide.


Sorry for making you read all that. I'm just very eager to get this problem--which has so far lasted almost five months, I think--fixed.

EVERYBODY WHO HAS EVEN MADE IT THIS FAR IS THE BEST. THANK YOU ALL SO MUCH.

OH I FORGOT TO MENTION


Ever since....sometime earlier in the whole timeline....I think it was right after I got the new ram sticks.....my motherboard began giving POST beep errors when I first booted it up.

It goes like this: I turn on my computer. The fans spin, and the lights light up. Nothing is outputted on the screen. After about twelve seconds of this, the motherboard starts beeping. Two short beeps, then a pause, then five short beeps. And it repeats over and over. However, if I hit the reset button, it will boot just fine. This happens every time I turn it on.


However, sometimes it boots okay if it's "warm." If I had it running for hours, and then turned it off, but then turned it back on only five minutes later, it might not do the beeping. If it's off overnight and I want to boot it "cold," however, the beeping is sure to happen.

You know how I said that when I installed the new CPU, at first it seemed that everything was fixed? I was wrong--it still did this beeping. Even though it PERFORMED perfectly for the first few days, the beeping still occurred. And the only way to turn on my computer was to hit power, give it a few seconds, let it start error-beeping, and then hit reset.

Don't even bother trying to look up what the 2 beeps 5 beeps means. I've tried. I've looked in the manual. I've googled. I've called Asus to ask them. But they don't tell me. When I told the Asus guy, he just told me to take out my RAM sticks and put them back in. So I did, but it fixed nothing and he still couldn't tell me what the beeping meant. You'd think they'd have a code for that :/

So basically, the motherboard knew the problem was still going on even though there were no symptoms yet. I guess that sounds like a motherboard problem to me.

Hazarding a few guesses here:

-I'm guessing the core issue is with the mobo, given how your second CPU worked fine at first and then started to go (just like the first one) and then continued to not work after that. Surely you didn't get two bad CPUs in a row; hence my guess of the mobo being the issue. This could be due to water damage, as you mentioned. (The mobo beeping, as described in your most recent post before this, would also seem to lend credence to this theory.)

-Since the issue seemed to be cumulative, and "reset" when you got the new CPU, I have to wonder if the CPU was in some way damaged by the probable bad mobo. When you get your new mobo in, I would not be entirely surprised if you still had issues--in which case I'd try replacing the CPU [i]again[/i] and see if that fixes it. If you want to play it completely safe (to make sure a potentialy already-damaged CPU doesn't fry a good replacement mobo), you could RMA the CPU as well before even trying the old one with the new mobo.

As others have mentioned, Prime95 is not at all GPU-intensive--essentially all it does with graphics is display its little window on the screen and update its status on there every few seconds. :smile: The fact that heavy graphics-intensive gaming seems to be the catalyst for the problem revealing itself would make me suspect, again, mobo damage, possibly somewhere to do with the data transfer lanes and/or mobo bridges (which are, from what I understand, heavily stressed during GPU-intensive tasks).

Once you get the system finally working, I would recommend letting a Prime95 stress test run for AT LEAST a few days, given its track record of problems thusfar. If it holds up after all that, try something GPU-intensive for an extended period of time as you did before and see how long it lasts. :smile:

I looked up your Mobo on newegg. Lots of bad recent reviews on there with stories like yours. And stories of problems that don't reset with simple power offs, too. Similar for the video card, along with rather vague power specs.

[QUOTE=wontstoptalking;259619]OH I FORGOT TO MENTION


Ever since....sometime earlier in the whole timeline....I think it was right after I got the new ram sticks.....my motherboard began giving POST beep errors when I first booted it up.

It goes like this: I turn on my computer. The fans spin, and the lights light up. Nothing is outputted on the screen. After about twelve seconds of this, the motherboard starts beeping. Two short beeps, then a pause, then five short beeps. And it repeats over and over. However, if I hit the reset button, it will boot just fine. This happens every time I turn it on.


However, sometimes it boots okay if it's "warm." If I had it running for hours, and then turned it off, but then turned it back on only five minutes later, it might not do the beeping. If it's off overnight and I want to boot it "cold," however, the beeping is sure to happen.

You know how I said that when I installed the new CPU, at first it seemed that everything was fixed? I was wrong--it still did this beeping. Even though it PERFORMED perfectly for the first few days, the beeping still occurred. And the only way to turn on my computer was to hit power, give it a few seconds, let it start error-beeping, and then hit reset.

Don't even bother trying to look up what the 2 beeps 5 beeps means. I've tried. I've looked in the manual. I've googled. I've called Asus to ask them. But they don't tell me. When I told the Asus guy, he just told me to take out my RAM sticks and put them back in. So I did, but it fixed nothing and he still couldn't tell me what the beeping meant. You'd think they'd have a code for that :/

So basically, the motherboard knew the problem was still going on even though there were no symptoms yet. I guess that sounds like a motherboard problem to me.[/QUOTE]

I'm not a computer person myself but have you checked out:

[url]http://en.wikipedia.org/wiki/Power-on_self-test[/url]

in this it says for POST AMI BIOS beep codes:

2 Parity error in base memory (first 64 KiB block)
5 Processor failure


for the mac codes( pre 1999) it's :

2 Incompatible RAM type installed (for example, EDO)
5 Bad checksum for the ROM boot block

for the mac codes(1999 and onward) it's :

2 incompatible RAM types
5 processor is not usable

of course there's more on that page.

you might also look at :

[url]http://www.computerhope.com/beep.htm[/url]

[QUOTE=science_man_88;259646]I'm not a computer person myself but have you checked out:

[url]http://en.wikipedia.org/wiki/Power-on_self-test[/url]

in this it says for POST AMI BIOS beep codes:

2 Parity error in base memory (first 64 KiB block)
5 Processor failure


for the mac codes( pre 1999) it's :

2 Incompatible RAM type installed (for example, EDO)
5 Bad checksum for the ROM boot block

for the mac codes(1999 and onward) it's :

2 incompatible RAM types
5 processor is not usable

of course there's more on that page.

you might also look at :

[url]http://www.computerhope.com/beep.htm[/url][/QUOTE]

oh and if it's a winbios I found a [URL="http://www.technick.net/public/code/cp_dpage.php?aiocp_dp=guide_beep_codes"]site[/URL] with resolutions to problems.

[QUOTE=Christenson;259634]I looked up your Mobo on newegg. Lots of bad recent reviews on there with stories like yours. And stories of problems that don't reset with simple power offs, too. Similar for the video card, along with rather vague power specs.[/QUOTE]

The delphic oracle says that
1) You have a bad motherboard, and
2) We have no information on whether the video card is good or not.
so
3) You have bought your last ASUS mobo
and
4) You have bought your last ATI/AMD video card. I really didn't like that AMD supports all these cool features (crossfire, physics computations) without letting ordinary mortals know the interfaces. NVIDIA simply publishes the CUDA APIs, thus we have mfaktc running circles around CPUs for trial factoring.

[QUOTE=Christenson;259664]The delphic oracle says that
1) You have a bad motherboard, and
2) We have no information on whether the video card is good or not.
so
3) You have bought your last ASUS mobo
and
4) You have bought your last ATI/AMD video card. I really didn't like that AMD supports all these cool features (crossfire, physics computations) without letting ordinary mortals know the interfaces. NVIDIA simply publishes the CUDA APIs, thus we have mfaktc running circles around CPUs for trial factoring.[/QUOTE]



Haha, I suppose those are all true.

I filed the RMA for the mobo yesterday, but didn't send it in yet. I'll do that tomorrow. As for sending in the GPU or CPU? Well....it would certainly be VERY safe to do so, but that seems very overwhelming to me...I don't like the idea of having the three most expensive parts scattered around the country all at once. Maybe that's just me being OCD :P

So I'm living on a hope that the mobo is the only problem. I HAVE considered what you said--that the mobo might actually be a CPU shredder, and it's ruining all my CPU's. That would very much suck.

But since I've waited about five months so far, I can wait the week it takes to send in the motherboard. And if the new one fixes nothing, I can wait another week for a new CPU (if it seems like it's a problem.) And another for GPU, if there's even a problem after THAT.


Right now I'd say it's most likely that the mobo is the ONLY component with a problem, seeing as how the CPU is brand new (unless what you said about the mobo ruining CPU's is true). I'll assume the new mobo will fix everything. If not, I'll just keep going down the line again...... I wish I had a friend who had a computer I could pop the CPU into to make sure at least that works.