Early 2008 Mac Pro Overheating in Windows XP from SWTOR?

atthesing said:

So I'm playing Star Wars: The Old Republic on Windows XP Boot Camp. I've been playing it without issue for a few weeks. After about 6 hours of playing, the game crashes. I simply run the game again, and it begins loading after I log in, but then boom: my computer restarts itself. I restart again to boot into Windows, and after about 2.5 hours of playing it happens again when I alt-tab. I left it turned off for a few hours.

My PC buddies explain to me that something was overheating. How can I figure out what was overheating? .

Click to expand...

I'm a PC guy and I'm not certain it's overheating. You could've been hitting a blue screen of death but Windows was set to reboot instead of putting up a BSOD. BSOD is analogous to a kernel panic in Mac OS X.

Uncheck the checkbox described at http://support.microsoft.com/kb/320299 to prevent the auto restart.

A possible (but not the only cause of BSODs) is a driver bug.

Your temps after boot aren't that meaningful. It would be more interesting to know them while under load, such as when you're hitting the above problems. If they aren't "high" for your hardware (CPU, video card, etc.), then it's likely nothing to do w/overheating.

Can you run the Mac Pro w/the cover off and have a large desk fan blowing into it? That could help eliminate overheating as a cause. (I don't know as I don't work on Mac software anymore. I think my old Mac Pro at work wouldn't complain if the cover was off. The old Power Mac G5s would run the fans at full blast if the cover was off.)

I am pretty sure it isn't overheating. I can run my early 2008 Mac Pro at full blast for hours on end while rendering with no effect.

The ram modules are getting hotter than the processors, that's for sure.

Also, in the place where the KB article directs you to, if "Write an event to the system log" is checked, then there should be an entry in the event log at the time it happened. It should be something along the lines of "The computer has rebooted from a bugcheck..."

The KB article also points you to http://support.microsoft.com/kb/308427 which is how to view the event logs.

GeForce 8800? I'd bet the GPU is crashing. The 8800s become unstable and then die a little later. It happens under high load when the GPU heats up.

In spite of the graphics, those temperatures would not be considered overheating. As mentioned, it could be a dying gpu.

I vote for faulty 8800. Go to any computer forum and you just trip over stories about how those fail. It's astonishing how many are resurrected by baking them though.

OP, I'd try a different card if you have access to one, and see if the problem persists.

It might be worth it to disassemble the video card to clean the cooling fins and fan. While you're in there I'd replace the thermal paste. Youtube has a couple of videos on how to do it. It's not half as hard as you think.

8800GT is old and smells of wee.

The Nvidia 8800GT was a kicking card in 2008 but that was 4 years ago. Today a decent GFX card kicks the 8800GT to the kerb and most modern games expect a lot more pixel pushing power to run well.
Consider the following replacements:
http://www.ebay.co.uk/itm/AMD-ATI-Radeon-6870-Apple-Mac-Pro-1-Gig-4-Displays-/230681534596?pt=PCC_Video_TV_Cards&hash=item35b5b0e084
http://www.ebay.co.uk/itm/Mac-Nvidia-GTX285-GTX-285-Mac-Pro-Video-Card-1-8-Gig-/230727353981?pt=PCC_Video_TV_Cards&hash=item35b86c067d
http://www.ebay.co.uk/itm/ATI-Radeon-HD-4870-Graphics-Kit-Apple-Mac-Pro-/370573023691?pt=UK_Computing_Computer_Components_Graphics_Video_TV_Cards_TW&hash=item5647df99cb
http://www.ebay.co.uk/itm/ATI-Radeon-HD-5870-Graphics-Upgrade-Kit-Mac-Pro-Mid-2010-Early-2009-/370574544091?pt=UK_Computing_Computer_Components_Graphics_Video_TV_Cards_TW&hash=item5647f6ccdb
For £180 to £280 you can make Noobs of the Old Republic run like it is on glass.

Temps are low on your Mac Pro.

My Octo MP 2008 RAM runs around 70-80ºc and my 5870 GFX card sits at 68ºc while caning WoW in a 25 man raid, your temps are cool by comparison.

atthesing said:

Doesn't visual noise usually accompany a dying GPU? Would the 8800 be failing after only 10 months of use? I'm pretty sure AppleCare put a new card in, not a refurb.

Click to expand...

Not necessarily. My 8800 did something similar in Windows where it would just crash before it died.

I don't think there have been any new 8800s made in years. They're likely all refurbs.

You set the warning parameters in Speed Fan so having the "fire" image does not mean anything. Your good on an iMac till at least 80ºC. You could have bad RAM. Or most likely bad video drivers.
gpzjock:
70º-80º is so high for a Mac Pro you may have thermal paste issues/ heat sink/ fan. My 2010 idles at 31ºC and never goes over 50ºC on the CPU and no higher than 65º on GPU, ever. That's playing Metro 2033 in DX11 with everything cranked. Or BF3 everything cranked cept only 2x AA because it is an AMD card

It is most likely the 8800GT. Had the same thing happen to me. Those cards are just waiting to die. I replaced mine with a 6870 which runs in OSX with no flashing (10.6.8 or above). You can get these cards for around $200 and it's more than twice as fast as the old 8800.

I've got the same machine but with more ram.

My several year old 4890 died while playing swtor. Similar symptoms, crashing (reboot) in game at first, after a few attempts it wouldn't even boot.

Replaced the 4890 with the base ati 2600 card the machine came with and problems went away.

Always been that temp with any number of modules

derbothaus said:

gpzjock:
70º-80º is so high for a Mac Pro you may have thermal paste issues/ heat sink/ fan. My 2010 idles at 31ºC and never goes over 50ºC on the CPU and no higher than 65º on GPU, ever. That's playing Metro 2033 in DX11 with everything cranked. Or BF3 everything cranked cept only 2x AA because it is an AMD card

Click to expand...

2010 Mac Pros have completely different (DDR3) RAM modules and a different fan/memory riser configuration to a Mac Pro 3,1.
My ECC Fully Buffered DDR2 RAM modules have always run hot since day one, the modules come with heat spreaders fitted and are worked very hard.
I have had 2 modules fail in the last 3 years both RMAed by Memoryamerica.com but they seem to work fine day to day otherwise. Memoryamerica replaced all 4 modules in each set both times to retain the matched pairs, so I have monitored temps on a total of 16 modules over 3 years, all of them give similar readings, as did the original Apple 1 GB sticks. At idle they range from 53ºc to 69ºc and under load they rise about 10-15ºc.

Recent monitor reading attached. As you can see the CPU diodes get nowhere near the temp of the RAM modules.

Did you ever clean your Mac since you have it? Or do you leave in a dusty environment?
Mac pros are not likely to overheat, but massive amounts of dust on the heatsinks can raise the temps. Especially GPUs fans are very sensitive to dust. Today's GPUs are running close to the physical limits when under load due to the combination of huge thermal output and chap stock fans...

gpzjock said:

2010 Mac Pros have completely different (DDR3) RAM modules and a different fan/memory riser configuration to a Mac Pro 3,1.
My ECC Fully Buffered DDR2 RAM modules have always run hot since day one, the modules come with heat spreaders fitted and are worked very hard.

Click to expand...

Sorry man. I thought the "My Octo MP 2008 RAM" was a label you gave the Mac for some reason and not that you were only referring to memory heat. Those heat spreaders looked awesome the first time I got a machine in to build. I was passing the risers around and everyone was like "Damn!". Same response for the price. I thought your CPU heat was 70-80ºC

If you want to be sure. Make an ordinary cleaning maintenance of your mac pro
Copied right from my other post.
(Ignore some sentence about the problems the original guy had and focus on the maintenance steps, I'll also delete and change some unnecessary and old text).
https://forums.macrumors.com/showthread.php?p=12764644#post12764644

What mac pro 2008 do you have?
What processors is in it? If you got the 3.2 GHz ones, there are some hazard special cooling compound applied to the heatsinks.
This need to be removed with care if needed. Use latex gloves!!
All this is written in the service manual, only available to Staff members at apple service centers.
The 3.0 GHz or 2.8 uses a piece of silver pads, this will melt and act as cooling compound.
Be careful when remove!

Today is 2012. The cooling compound is 3 or 4 years old or even older, and pretty likely dry as old bubblegum, this should be changed at least once a year.
Get you some Artic silver S5 or Arctic silver ceramique at a PC hardware store, costs about ten bucks, also available in 20 gram tubes.
If they don't have it in stock. Ask for Artic silver Céramique 2 or visit some other PC hardware store.
Artic silver S5 and Arctic silver ceramique can move high temperature as 8,9 W/(m*K).
Other cooling compounds only moves 2 W/(m*K).

Get your hand on a small brush for painting, this you will clean internal parts with.
Don't use compressed air ever.

Before doing anything, try to look and understand how to disassemble your mac pro, to get it clean.
http://www.youtube.com/watch?v=xOF435jCg8E
I know it's in german. I understand..kind of nothing of what he is talking about. But hey, just fallow he's steps in the video and you will be fine.

"Why not use compressed air?"
Because if you apply compressed air, you will loosen the dust.
When the fans spin again, the dust just a moment ago was flying around now flying deeper in to the heatsink. Making the heatsink to clog even more and more for each time you use your mac.

With your brush, clean all the internal parts. Including fans, heatsinks (if they are clogged, blow from the back to forth until you can see trough clearly).

Now grab some rubbing alcohol. Take the CPU's out of their sockets.
BE CAREFUL!, don't touch the contacts underneath.
Rub of the old cooling compound of as good as you can.
Please do the same to the heatsinks.
You might find your paper now being black. This is good.
This means dirt is being removed.


Now look at all your black heatsinks as covering chipsets and voltage regulators.
Remove the clips on both sides and clean the surface with acetone, making contact with the heatsink and the heatspreader on the chipset.
Do this on all chipsets.
The heatsinks to remove is:
* in front of the left side CPU and the right side CPU there are two rectangle black colored heatsinks.
* To the top left behind the cpu's there is a bigger rectangle heatsink.
* Right behind the heatsinks there is a black heatsink with a vertical aluminum heatsink mounted on heatpipes.
When cleaned all those, applied cooling paste and Re-assembled.

Now re-seat the CPU's carefully in to their sockets. Close the latch.
Now grab your Artic silver compound and apply just a little at the CPU's
Try to spread the compound to the whole surface, you
might need to apply some more. Now apply some on the heatsinks.
Spread it just like on the CPU's.
Attach the heatsinks one by one. Remove each to check that compund has
spread to the whole surface before screwing the heatsinks to it's place.
Don't forget to put the Temperature sensors back in to place!!
And your done!

Now disassemble your GPU card, on these there are only cooling pads.
What they are made of is simply wax. And all this has with the time dried out.
Clean the heatsink and use again, rubbing alcohol cleaning the old compound of.
When taken a part, if there are differences in levels on memory parts and the cooling compound wont reach the heatsink when putting it all together.
Cut some small pieces of copper plate and apply all around as many until the level is just perfect to to make contact to the surface of the heatsink. Now apply cooling compound on both sides of the copper plate. Make sure that no part has no cooling compound.
WARNING!! When cutting the copper plate. Make sure, not to make the pieces to big to short other components. The copper plate should only cover memory components and voltage regulators where heat comes of.
Try to make them as identical as the old cooling pads, this will make sure that the copper plate is seated where it should be.
Now put the GPU card together.

Now on to the next thing. Your PSU.
Grab a flashlight. Take out the PSU and look in to the grill.
See if you notice any loose internal cables.
Anything not attached properly.
If it is just clogged with dust, take it apart and clean it.

This part was about a guy who had problems with hes PSU but it might happen to any PSU without you knowing it. Its worth checking.


(When taken apart, and you discover something not securely is sitting tight.
Are your soldering skills good, unsolder the component and reattach it again.
Other ways visit an electrician at any repair center. Don't need to be an apple repair center because the PSU has nothing with apple to do.
It's just supplies the mac with power.)

Please ask if you need help. This is a mainly service needed to be made on every computer. Computers are not machines which work forever.

Clean machine regime

I do clean my Mac Pro annually:
I open the side door, remove the GPU and RAM risers, hoover the whole board and all cavities.
Then airgun the RAM risers making sure I blow down every heat spreader channel.
Then I air blast the fans from front and rear. Finally I hoover then airgun the Vapor-X cooling fins/pipes and fan blades on the GPU (Sapphire HD5870). When I can't see anymore dust I reassemble.
I'm a non smoker and the only risk of vent blockage is fluff molted from our domesticated pet rabbit who tears around the room for about 30 minutes a day getting his exercise. I hoover the front and rear cheesegrater grills monthly.

Derbothaus - No worries m8, easily done. I should write stuff more clearly.

When you say 'hoover' I seriously hope you aren't using a regular vacuum cleaner? Static electricity is death to electronics and vacuum's generate quite a bit of it. Your not on carpet are you?

gpzjock said:

I do clean my Mac Pro annually:
I open the side door, remove the GPU and RAM risers, hoover the whole board and all cavities.
Then airgun the RAM risers making sure I blow down every heat spreader channel.
Then I air blast the fans from front and rear. Finally I hoover then airgun the Vapor-X cooling fins/pipes and fan blades on the GPU (Sapphire HD5870). When I can't see anymore dust I reassemble.
I'm a non smoker and the only risk of vent blockage is fluff molted from our domesticated pet rabbit who tears around the room for about 30 minutes a day getting his exercise. I hoover the front and rear cheesegrater grills monthly.

Derbothaus - No worries m8, easily done. I should write stuff more clearly.

Click to expand...

Well its not just fluff from cigarettes blocking the heatsinks in long term...

Think of what we humans have on us. Skin, our skin drops to the floor every hour in tiny bits. Mixing with the fabric from your clothes and furnitures.
Particles from cars exhaust and engine.
Salt from your sweat and tears.
Particles from hair products and cleaning products.

The heat from the sun and lamps, also your body heat circulates all this particles around the room. You don't see these particles until they are so many as a pile of dust.

Where the CPU heatsinks are located is like 100% filled up with electronics or tight fit of the air duct design.
Without psychically remove the fans and duct and heatsinks and processors.
I bet parts of your heatsinks is a bit clogged because it is like a very tiny honeycomb with holes were the air is pushed through.
And by blow air in from the front and to the back clogging it even more tight.

And even before you cleaned out the computer from dust and it was getting a little hot.
The cooling paste is already gone.

nobiggiestudio said:

What mac pro 2008 do you have?
What processors is in it? If you got the 3.2 GHz ones, there are some hazard special cooling compound applied to the heatsinks.
This need to be removed with care if needed. Use gloves!!
All this is written in the service manual, only available to Staff members at apple service centers.
The 3.0 GHz or 2.8 uses a piece of silver pads, this will melt and act as cooling compound.
Be careful when remove!

Today is 2011. The cooling compound is 3 or 4 years old, and pretty likely dry as old bubblegum, this should be changed at least once a year.
Get you some Artic silver S5 at a PC hardware store, costs about ten bucks.

Click to expand...

I'm going to have to call BS on the above advice. You really think all other Mac Pro owners or anyone w/a heat sink and a pad/compound in between does that, ever, let alone yearly? The VAST majority of PC guys and I'm sure Mac owners don't do that and their CPUs are running fine.

Sure, if you want to replace whatever was there before with something better, like Arctic Silver, fine, but making it a yearly exercise is just goofy. I'm all for clearing out dust of the inside of the machine though...

As for attaching cutting copper plates and attaching them, that's pretty far out of left field, if you ask me. As for unsoldering and resoldering stuff, umm....

I started w/the original IBM PC in 1983 and have put together most of my own desktop PCs since the 386 days.

I used to test software for a professionally, for a living for >13 years at a large company, mostly on PCs but was also testing Mac software for 4 years. NEVER have I nor anyone else ever removed CPU heat sinks in order to remove the old compound and replace w/new. Some of machines got pretty old too but never had overheating problems, let alone due to old HS compound.

We still don't know the cause of the OP's issues and whether he was hitting a BSOD (which could have a LARGE number of possible causes besides heat). There are things he can do to narrow down whether it's a thermal problem, which I already suggested. IMHO, it's a foolish errand blindly do all those things suggested by your post. The issue may have nothing to do w/heat, dust, degradation (if any) of heat sink compound, etc.

I once had a Win 2K3 server (which undergoes massive file i/o activity) that started hitting BSODs, out of the blue. The issue was traced to some part of the NTFS code that was blowing up and it was suggested to run a chkdsk /f across all volumes. Sure enough, that fixed it. Never had anymore issues.

Since a possible cause is bad RAM (already mentioned earlier), one could try running http://oca.microsoft.com/en/windiag.asp and/or http://support.apple.com/kb/ht1509 (to test more than RAM). RAM tests and hardware diagnostics aren't necessarily foolproof though.

cwerdna said:

I'm going to have to call BS on the above advice. You really think all other Mac Pro owners or anyone w/a heat sink and a pad/compound in between does that, ever, let alone yearly? The VAST majority of PC guys and I'm sure Mac owners don't do that and their CPUs are running fine.

Sure, if you want to replace whatever was there before with something better, like Arctic Silver, fine, but making it a yearly exercise is just goofy. I'm all for clearing out dust of the inside of the machine though...

As for attaching cutting copper plates and attaching them, that's pretty far out of left field, if you ask me. As for unsoldering and resoldering stuff, umm....

I started w/the original IBM PC in 1983 and have put together most of my own desktop PCs since the 386 days.

I used to test software for a professionally, for a living for >13 years at a large company, mostly on PCs but was also testing Mac software for 4 years. NEVER have I nor anyone else ever removed CPU heat sinks in order to remove the old compound and replace w/new. Some of machines got pretty old too but never had overheating problems, let alone due to old HS compound.

We still don't know the cause of the OP's issues and whether he was hitting a BSOD (which could have a LARGE number of possible causes besides heat). There are things he can do to narrow down whether it's a thermal problem, which I already suggested. IMHO, it's a foolish errand blindly do all those things suggested by your post. The issue may have nothing to do w/heat, dust, degradation (if any) of heat sink compound, etc.

I once had a Win 2K3 server (which undergoes massive file i/o activity) that started hitting BSODs, out of the blue. The issue was traced to some part of the NTFS code that was blowing up and it was suggested to run a chkdsk /f across all volumes. Sure enough, that fixed it. Never had anymore issues.

Since a possible cause is bad RAM (already mentioned earlier), one could try running http://oca.microsoft.com/en/windiag.asp and/or http://support.apple.com/kb/ht1509 (to test more than RAM). RAM tests and hardware diagnostics aren't necessarily foolproof though.

Click to expand...

Well, ..Its time to go back to school.
Read a little physics and chemistry and you'll know why I mentioned all this about changing out the cooling paste.
I'm only 18 years old and I'm been proving every old engineer at school, work and private that our generation knows allot more than them in modern computers. Its not just a bunch of wires and cards sticking out here and there.
Computers ain't the same as in the 80's and 90's anymore, there's allot more going on in todays and our future computer's.

nobiggiestudio said:

Well, ..Its time to go back to school.
Read a little physics and chemistry and you'll know why I mentioned all this about changing out the cooling paste.
I'm only 18 years old and I'm been proving every old engineer at school, work and private that our generation knows allot more than them in modern computers. Its not just a bunch of wires and cards sticking out here and there.
Computers ain't the same as in the 80's and 90's anymore, there's allot more going on in todays and our future computer's.

Click to expand...

Great... an 18 year old who thinks he knows it all.

FWIW, I've got a computer science degree and had to take a chem class and three physics classes, which were of the same series that all engineering and physics majors had to take.

You seriously don't think that the engineers at major hardware companies such as Apple didn't do proper validation testing, testing for longevity and didn't understand the properties of the thermal compound they applied?

As for adding copper plates, if OEM components running at or under rated specifications (not overclocked) needed heat spreaders, they'd have put them on. Major hardware companies would be negligent to not test in a variety of environmental conditions (esp. those they list in their specs).

Go work professionally in a large hardware or software companies for 5+ years (for me, it was over 13 years) w/a large # of people w/a large population of PCs and/or Macs and see how many replace their HS compound at all (let alone annually) and have power hungry components such as CPUs or GPUs die due to the lack of such "maintenance".

I still stand by my statement and my earlier advice for the OP. Your advice on changing out heatsink compound every year is ridiculous. We need to understand what is happening first before blindly mucking w/hardware beyond getting rid of dust (esp. when we don't know the skill level of the OP or anyone who may read your posts). It may have nothing to do w/heat at all.