I think it's the graphics card...

[Rostum]

Well, a problem I am having as of late is that whenever I try to run Final Fantasy XI I'll get a random 15 minutes to 1 hour game time and then my computer suddenly shuts off -- except that the fans are still running and I have to turn it off from the back to start it up again!

I'm not really sure what could cause this... I have a relatively decent computer and Final Fantasy XI is by no means a very CPU or GPU intensive game.

My specifications:
Intel Core2Duo 2.2GHZ 4MB Cache
nVidia 7800GS 256MB
3GB DDR400 RAM
450w Power Supply (I forget the model, but it's a CoolMaster)

I am thinking that it might have to do with the settings in the bios, and it doesn't seem to be an overheating issue (I've checked, there's also plenty of cooling and airflow).

Please, any one have any ideas? T_T

[Akaria]

That happened to my desktop when I would play The Sims 2. I'd be able to play for a bit, then the screen went black and into stand by, and when I'd try to turn it on again, it would go straight into stand by again; I'd have to restart the computer to get it back on. My dad and I had deduced that either my graphics card was going bum, or my motherboard was burning out. Turns out, the fan on my graphics card was burned out and coming loose, which...Obviously isn't good. I'd say take it out, make sure nothing is physically wrong with the card itself, and of course check it in detail since there are so many tiny components. If it looks fine, if you have another computer, try switching the cards and play FFXI with the other one. If it works fine, it's your card. If not, you have bigger issues.

Hope that helps a little!:bigsmile:

[Rostum]

Hmm, well I have checked the card and there doesn't seem to be anything wrong with it visually. Unfortunately I don't have another computer or graphics card to test it on...

I checked the nVidia monitor while I was playing and I had the GPU at 47-50 degrees (Celsius I think?). CPU was at 17% and memory was at something like 10% (this is with the 2048x2048 background texture resolution and 15 drawdistance at 1680x1050).

Hmm, it's really baffling me. This sometimes happened (very rarely, like twice out of a month) when I played WoW mid-last year.

Edit:

Did a 10 minute stress test, not sure what it means:
CPU Stress Test status: 55 passes completed
Memory Stress Test status: 34 passes completed
Disk Stress Test status: 40 passes completed
GPU Stress Test status: 57 passes completed

By the way, in the last week it wasn't doing this and has just started recently. I am really baffled as to what is wrong...

[o_O]

It's likely that the powering off is caused by your CPU overheating. It's a defense mechanism built into chips to stop them from blowing.

[Rostum]

I don't know why it would be doing that. I have installed one of these (http://www.geek.com/gigabyte-g-power-lite/) on the CPU...

Any suggestions on what steps I should take from here?

[rubah]

have you installed any applications to tell you what temperature your cpu is at?

[Momiji]

If his computer was overheating, wouldn't there be artifacts all over his screen? My desktop overheated once and it damaged the video card, now it leaves artifacts everywhere no matter what temperature it is.

[Rostum]

have you installed any applications to tell you what temperature your cpu is at?

Ah, no. I've just tested the GPU temperature... Any suggestions for a program that will do this?


If his computer was overheating, wouldn't there be artifacts all over his screen? My desktop overheated once and it damaged the video card, now it leaves artifacts everywhere no matter what temperature it is.

Not necessarily. My graphics card isn't damaged.

[Citizen Bleys]

have you installed any applications to tell you what temperature your cpu is at?

Ah, no. I've just tested the GPU temperature... Any suggestions for a program that will do this?

SpeedFan (http://www.almico.com/speedfan.php). It's not always completely accurate, but it doesn't require you to install any new physical hardware, so it's better than nothing.

[Rostum]

Okay, well this is the test normally.
http://img293.imageshack.us/img293/7273/speedfannf4.gif

About 5 seconds before my computer crashed (about 5 minutes into FFXI this time!) my temperatures that were changed were:
Core: 50C (green tick?)
GPU: 50C (green tick?)

Everything else was the same. Hmm.

Note: Also I thought it might be Windower 3.3, but I tried uninstalling that and just using max default settings in the FFXI settings and it still does it (which it shouldn't).

[Citizen Bleys]

SpeedFan has default settings that determine what qualifies as a green check and what doesn't. Since no 2 CPUs or GPUs are alike, I pretty much guarantee you are wrong. For example, my old Pentium D was rated to withstand temps up to 69.9°C (which it regularly hit), whereas my current Core 2 Duo can't take anything over 61 (which it never approaches). My own CPU temp hovers around 30-40°C while playing FFXI, but I haven't checked my GPU. I was getting some freezing--like you, only while playing FFXI, but my screen didn't go black, just whatever was on the screen froze solid. It happens with and without windower. Since replacing my graphics card, it has not happened, but since it's random and the replacement was recent, I don't know if the issue is gone for sure.

Lately I've also been getting the same issue on my old laptop, without Windower. It could be the graphics card overheating--I keep my old laptop at work and usually keep my char in one of the beastman-controlled campaign areas and pound away at fortifications while I'm at work, only looking at the screen every 10-15mins or so, depending on how busy I am at work. As the graphics card fan blows air out the bottom of that laptop, I'm going to try a laptop cooling pad, I'll let you know how it works out. I have gone nights at work with no freezing at all, but they don't happen often, and we're only allowed to have our laptops out for 4 hours, between 2 and 6 AM. If I can go a week (4 working days) without freezing once the thermal pad is in place, I'll send you a PM. If not, assume it doesn't.

short version: you'll have to check your GPU manufacturer's website to find out what kind of temps it can handle. 50°C seems kinda high.

[Rostum]

Hmm, well on the nVidia forums they say it can handle up to a max of 60C, and run comfortably at 50C.

At the moment my CPU is 42C and GPU is 48C without running any games.

Edit: I've been going to nVidia for drivers, maybe I should go to XFX for drivers? (after all it's an XFX nVidia 7800GS). Perhaps they've put in something that nVidia don't have drivers for? I'll try it, but not sure if it's any different.

[o_O]

Nah, nVidia drivers are fine, since each graphics card model is a specification.

Can you set Speedfan to log the temperature every five seconds or something like that? You'd be able to determine the temperature at crash-time more reliably then.

[Rostum]

Hmm, I'll set that up on SpeedFan. Well after installing the XFX drivers I had a thought so I looked through the box and the manual.

There was a power cord supplied in the box, and the manual said to use it. Now I just had the card hooked straight up to the power supply... It's a split cable:

Female------\
Female-------Male

But I had to put one of the females into the card and the male into P4 of the power supply (it said to put the male into the card and the two females into separate power cords but the connections are switched round). Would it make a difference?

Anyway, I did that loaded up FFXI and so far no crashes. I'll keep it logged into a heavy area with settings up max to see if it crashes at all. If it does I'll do the SpeedFan log you suggested.

(sorry if the explanation is bad)

[Citizen Bleys]

Can you set Speedfan to log the temperature every five seconds or something like that?

Yes. The intervals are customizable. You can also set it up to email you automatically if certain conditions are met, such as the CPU or GPU exceeding the manufacturer's spec.

[Shoeberto]

Another good way to check the CPU temp after a crash is to just boot into your BIOS right after it happens and see what it reports. I had it happen with mine a while back and it showed really high temperatures.

When it happened to me it turned out there was gunk (dust and hair) all built up over the heat sink, preventing heat flow. Unscrewing the fan from it and running over it with a q-tip to clean it off helped.

[Rostum]

Hmm, well what I did didn't end up fixing it. So... I'll try cleaning the heat sink.

I could even try just putting on the cooler that came with the CPU, but it didn't come with any thermal paste (like the one I bought), not sure if it needs it but I have no thermal paste.

Hmm...

Edit: Here's the nVidia monitor while playing FFXI in Windower 3.3 with DrawDistance set to 15 and background texture resolution set to 2048x2048. The CPU is at 50% in whitegate but something like 17% in most other areas.

http://img181.imageshack.us/img181/9228/nvidiaperformancetestrm7.jpg

Edit:
Just crashed, went into the Bios straight away and CPU was at 35C. So I don't think that's it... Hmm...

[Rostum]

Okay so I formatted my computer and installed all the latest drivers for the XFX nVidia 7800GS AGP card that I have, as well as new bios and drivers for my ASRock 4CoreDual-VSTA motherboard.

At first Final Fantasy XI was working fine, but then it started to crash again... I really don't have the money to go spend on a new PSU, or a new GPU or a new motherboard, or new RAM, when I don't know for certain which is going to fix it.

I had a thought, reading through the motherboard's manual. The AGP slot it has is a 1.5V and they specifically say not to use a video card with 3.3V. Now, I can't find anything that says what the voltage is on my card, not even in the manual... Which is incredibly weird. Also, what if it is 3.3V? Is there a flash bios to make it 1.5V? Is that safe? Hmm.

Any help would be appreciated...

Also... It says for system requirements "An aGP 2.0 compliant motherboard, and an available hard disk drive power dongle (only for AGP)". I have no idea what they are talking about here!

[Rostum]

Hmm, well XFX Force (the manufacturer's of my nVidia graphics card) said that it sounds like I have a faulty card and are sending me a replacement.

I also thought that I could buy some new RAM (running some old cheap DDR400 RAM and can get DDRII 667 RAM). I am thinking that there are only two slots but I could get 2x2GB DDRII 667 RAM although I am not sure if the motherboard supports more than 2GB...

The manual doesn't say anything about limits, but I am still not sure. Will it be fine to get the 4GB RAM?

I was thinking of these:
Kingston 4096MB (2 x 2048MB) PC2-5300 667Mhz DDR2 RAM - CL5 (68274 | KVR667D2N5K2/4G) | Techbuy Australia (http://www.techbuy.com.au/p/68274/MEMORY_DDR-2_RAM/Kingston/KVR667D2N5K2_4G.asp)
Or
Corsair 4096MB (2 x 2048MB) PC2-5300 667MHz DDR2 RAM - CL5 - Value Select (81416 | VS4GBKIT667D2) | Techbuy Australia (http://www.techbuy.com.au/p/81416/MEMORY_DDR-2_RAM/Corsair/VS4GBKIT667D2.asp)
Or their 2GB counter parts if 4GB is a no-go...

Any help would be appreciated, cheers.

Edit: Motherboad is 4CoreDual-VSTA incase anyone wanted to double check for me or something.

[o_O]

Yes, you should be fine to get 4GB of RAM - it mostly depends on your CPU. If you only have a 32bit CPU then you won't get above 4GB, and if you're on XP you won't get above 3GB. I would personally go with Corsair RAM as it's one of the top brands, but it certainly doesn't come cheap.

All signs so far are pointing to faulty hardware, but keep in mind that it may not be the GPU. Your CPU and GPU are clearly not overheating, so my guess at the next thing in line would be your PSU. What power is your PSU, and what devices do you have that are drawing decent amounts of power (drives, graphics, processor)?

[Rostum]

But there shouldn't be a limitation from the motherboard if the manual doesn't mention it? I'll go with that Corsair one. Edit: Ah it says it's "Max. capacity of system memory: 2GB" in the memory specs on ASRock's website.

I have an eXtream 430w/350w Cooler Master that connects to a CPU, GPU, 250GB HDD, DVD/CD writer drive, and USB drivers (tablet, micro hi-fi system, and mouse) and the rest is just an ethernet port and a keyboard port.

Well I'd like to get new RAM anyway, but will have to think about the PSU. I was told that the PSU's power should be absolutely fine but maybe it's not enough... If it's not, is there anything you can suggest for me?

[Citizen Bleys]

Recap: Remember I was having what sounded like the same issue as you on both a desktop and a laptop pc.

Update: I replaced the graphics card on the PC and haven't frozen since. I suspect your RMA card will resolve your issue similarly.
Getting a cooling pad for my laptop did precisely bugger.

[Serapy]

Do you have the same problem when you run a different game?

[Rostum]

Bleys: Perhaps I should wait until I get the new card before buying a new PSU and memory.

Serapy: Yes, as I believe I have stated (if not my apologies).

[o_O]

If they're sending you a replacement graphics card under warranty then yeah, definitely wait. If you can get your hands on compatible parts for testing then you could try swapping them out while you're waiting. :p

[Rostum]

Alright well thanks for the help so far guys! :D I will tell you all how it goes once I get the replacement card and test it out.

Cheers!