abc

Totally understandable to chicken out at those temps, we forgive you!

It's looking more and more like a problem with the temperature of your GFX card. You said that sometimes it freezes after 5 minutes, but I'm sure that since using furmark you're aware of how quickly the GFX card can heat up.

The first GFX card, when released, was a tad notorious for running very hot due to a slightly inadequete cooler but this was largely solved by upping the fan speed. I know for a fact that on some models you can't adjust the fan speed (i.e. you can adjust it but nothing actually happens) It's pretty easy to check whether your cards fan is working or not, because at 100% it will sound like a small hurricane!

It's something I (we've) assumed all along but I mean it may well be worth taking the side of the case off when it's turned on to check that the fan on the GFX card is actually moving. Occasionally the fan connector plug can wriggle it's way out, or if the fan has been at 70% + speed for it's life then it's entirely possible the fan has simply worn out and now goes much slower than intended if it works at all.

Did you give everything a quick dust?

__________________
"I don't like new Linkin Park, I don't like old Linkin Park... I JUST LIKE LINKIN PARK"

COD owner - COD expansion pack owner - COD2 owner - COD2: BRO player - COD3 owner - COD4 owner - COD:WaW owner - COD: MW2 owner - COD: BO owner

This is getting repetitive.

oneinchwunda

FYI, Furmark stresses GPUs by a HUGE amount.

In games, my GPU never passes 60c, usually hovers around 55c, but Furmark pushes it to over 80c

__________________

lamanga2004

I dusted it out the other night, and the fan defo works at 100% - I can modify it between 20% and 100% using CCC and you're right - it is noisy at 100%.

The interesting thing is that when I have GPU-Z running in the background, it samples the GPU temp every second and logs it to a text file.

Here's a graph of the results so far. The peaks are just before the crashes. I guess it is running rather warm!

abc

That is definitely running "dangerously/problematically hot" as opposed to "rather warm"

As I mentioned one components cooling is usually defined by other nearby components. Do you know many details about your overall cooling setup? E.g. what case do you have, what CPU cooler do you have, do you have any intake/exhaust fans, if so how many and what size etc

__________________
"I don't like new Linkin Park, I don't like old Linkin Park... I JUST LIKE LINKIN PARK"

COD owner - COD expansion pack owner - COD2 owner - COD2: BRO player - COD3 owner - COD4 owner - COD:WaW owner - COD: MW2 owner - COD: BO owner

This is getting repetitive.

lamanga2004

OK. So I took the PC up to my Dad's earlier, to have a go at it with the compressed air hose...

A couple of minutes later, and after the ejection of a surprising amount of dust and cobwebs, I put the case back together and brought it home.

This time I ran FurMark for 60 seconds in benchmark mode, at 1360x768. Max temp was 46C. Hmm. This is interesting I thought. So I then ran FurMark for 3, then 5, minutes in the same mode. The temperature climbed towards 56C, but was looking like it wouldn't go much higher.

This has to be a marvelous result, irrespective of whether the machine keeps crashing in future.

The take-away from this is:

Let's see what COD:MW2 makes of my newly cool GPU. I shall report back at some point in the next few days.

Thanks all for your help so far.

lamanga2004

Well that's got me stumped. Despite my very high expectations, I lasted 20 minutes before BOOM! Freeze and hard reboot.

Checking back over GPU-Z log, and it was 48C when it crashed.

I, therefore, rule out overheating.

Right. Back to the drawing board. Any more ideas?

oneinchwunda

As abc said (I think), that heat can cause permanent long term damage and issues with the card, even now that it is running cool again (those are VERY good temps btw).

__________________

lamanga2004

Good, glad to hear that it is now running at nice temperatures.

I genuinely don't think I've fried the card - it performs fine for every other application and game.

I just think that what we have done has ruled out overheating as a cause.

So... any other ideas?

Nibbles

Today I checked the other two minidump files you posted last week and they both show different results! In file 1 & 2 (generated late august) it seems to point towards faulty RAM.

How many sticks of RAM do you have? If more than 1, is it possible to remove one and still have enough total RAM to test the game?

Unfortunately, it looks like it's getting to that stage where you may have to start swapping out hardware anyway - do you have a friend who can lend you a different graphics card that will play nice with your PSU? Try swapping out RAM first if possible.

Check the RAM clock settings too in the BIOS. I know you reduced the frequency but check the CAS / RAS timings & voltage (the sticker on each stick of RAM should detail the timings).

lamanga2004

Will do that Nibbles, and post the results here. Thanks for analysing those minidumps.

lamanga2004

201009140045 - freeze and hrb
201009141000 - freeze and hrb
201009141003 - removed one 2GB stick of RAM and trying again
201009141033 - changed BIOS settings: tCL 6, tRCD 6, tRP 6, tRAS 18, voltage was already at 1.9V
201009141045 - crashed to desktop - iw4mp.exe
201009141335 - freeze and hrb

That didn't seem to work, so I am about to swap them over.

For the record, it says this on the stickers:

Crucial CT25664AA800.16FE
2GB 2RX8 PC2-6400U-666-12-E0
2GB, DDR2, 800, CL6

Darn it. Swapping them didn't work either. Lasted only a few minutes.

201009151910 - swapped RAM sticks over
201009151915 - freeze and hrb

oneinchwunda

Try booting with 1 stick of RAM, and make sure you have it at "stock" speeds and timings.

I would also try raising the North Bridge (NB) voltage a little too, probably about 0.1Volts.

__________________

lamanga2004

Thanks, will raise the NB voltage by 0.1V and edit this post when I have done so and rebooted.

Quick question though - for the RAM I described in the post above, what are stock speeds and timings? Thanks in advance.

oneinchwunda

I'm not sure.

Most common RAM like yours, i.e. DDR2 800MHz, has timings of 5-5-5-15. Now yours I see, has CL6 on the sticker. This means CAS timings (like 5-5-5... etc.) of 6-6-6-18 (approx.).

Downloading CPU-Z, will tell you the default "profiles" if you like that your memory has. Go into the "SPD" tab, and it will show you everything about your RAM, along with these "profiles" (actually called "JEDEC", now I look at them) which show default combinations for speed, timings, and voltages. Setting it to the one closest to 800MHz, with timings of 6-6-6-18 would be your best bet (as that is what is shown on the sticker). And, FYI, 400MHz in CPU-Z actually means 800MHz. I could see that question coming.

Fuck. Scratch that. BEFORE YOU DO ALL THAT... There should be something in your BIOS called "Restore BIOS defaults". Do this before you fiddle with RAM timings and stuff. Additionally, you can, If you are comfortable with it, take the side off of your case (after removing the power lead etc. etc.) and take out the CMOS battery for a minute or two. This also does a reset of the BIOS and puts everything back to factory settings. Hence, everything should be the same speed etc., as when you bought/built it.

That was a long post (for me) and I have a few other ideas, but see how this goes for now.

__________________

lamanga2004

Cheers oneinch, but resetting BIOS had no effect. I didn't remove the battery (the one that's bigger than a 1p but smaller than 2p right?) but I did restore factory defaults.

CPU-Z tells me I should use 6-6-6-18, so I have also changed the timings to this. Unfortunately this had no effect either.

So, any more ideas? I am still getting freezes.

oneinchwunda

The next step, unfortunately, is probably to start swapping out hardware.

The main problem with this isn't just sourcing parts, however, but the fact that this issue could be any of either the RAM, GPU, Motherboard or even the PSU.

Start with the easiest to replace, which is either the GPU or RAM.

__________________

lamanga2004

OK, so I figured swapping out stuff would be an expensive, and probably time-consuming, exercise so I skipped it for now.

I've stopped playing COD as much because it's annoying.

So tonight I start watching Iron Man 2. About 60 minutes or so in I get a stripy screen followed by the squeal of death. I was using VLC to watch the film.

Nothing for it but to hard reboot.

Does that alter anyone's diagnosis?

lamanga2004

After many months of not playing COD, I decided to install Windows 7 thinking this might fix my issue.

5 minutes into playing COD, it became apparent that even that hasn't helped.

I am so angry and disappointed right now. Please help.

lamanga2004

Is it an IRQ thing? PCE IDE controller and HD Audio drivers both on IRQ20...