GX740 way too hot! (yes I've updated VBIOS)

M

matogl0396

New member
PRIVATE FIRST CLASS
Joined
Sep 26, 2011
Messages
6
Hello,

Just picked up a GX740-434us. Flashed VBIOS to 035757.031 and also updated to latest 5870 drivers (both from MSI website for this product).

A test of MSI Kombuster spikes my GPU temps over 100 C within a few minutes. Same result from playing Metro 2033 for ~5-10 minutes. At this rate my card is going to fry...

Would the new board BIOS and/or EC firmware address this at all? I've read a ton of posts on this forum as well as others about the VBIOS to fix GPU heat issues, but that doesn't seem to have solved my problem. The fan does appear to be working, and the card's running with 1.05v.

My only other thought is poor thermal paste...Assume it'll void my warranty to re-apply that.

Was just hoping someone here might be willing to offer some advice. I did read one post about flashing to VBIOS from the GX660R, but that was for the "dancing pixels" issue, which I am not experiencing.

Really sad right now...was so looking forward to this fine machine. :( Hope I don't have to RMA...

Thank you very much in advance for reading/replying!
 
Well just to update I flashed the BIOS and EC just for good measure, but that didn't change anything. Not sure what to do now except maybe test different VBIOS, or underclock the GPU...
 
Hmm I'm not even sure here. I bought this item off eBay, but it was advertised as NEW/UNUSED. There had been a Newegg sale on this model (they were retiring the SKU I believe) just days before and I suspect this seller was flipping it for a bit of profit.

The item came in original packaging, fresh Windows, near as I can tell the serial is unclaimed. Is my eBay receipt acceptable proof of purchase for a warranty?

Also...looking at the underside, the "void if removed" warranty sticker is just kind of placed in the middle of the bay that opens to CPU/GPU. I think there might be a screw under there, just from feeling over the area...is that how it's supposed to be?

And at any rate (especially if MSI won't accept my proof of purchase)...can you offer any insight into the heat issue? Would you agree a possible bad TIMM application? It doesn't seem to be software-related at this point, but I've certainly been wrong about things from time to time in my life. :)

Does anyone know what the TMAX spec is for this card? I can't seem to find that info anywhere, but I'm afraid that the 104 C it hit in Furmark (very temporarily, turned this off as soon as I noticed) could have even damaged it already...

Thanks a lot,
 
I've been having the same problem for months now.

I bought it brand new from newegg. Updated the vBIOS but nothing has changed... Most games will skyrocket my VGA to 90 degrees. Sometimes even sitting at the desktop takes it well over 50. I've experienced overheating shutdowns at least 20 times now, it's getting extremely frustrating!
 
And as a side note the vbios that you can download from the website is only available in x86, so even if I wanted to re-flash I can't because there's no x64 version available.
 
Colino said:
And as a side note the vbios that you can download from the website is only available in x86, so even if I wanted to re-flash I can't because there's no x64 version available.
Click to expand...

VBIOS/BIOS flash is done in pure 'DOS' mode, not in Windows, so 32bit or 64bit OS is irrelevant.
 
Well I flashed it again just in case, but I'm having the same issues.

I'm getting up to 50 degrees idle and close to 90 when load gets to 99% (sometimes a flash game is enough to get there)
 
Still looking for an answer about the eBay proof-of-purchase, if anyone would be so kind as to reply?

And as an update, in my case I am honestly starting to suspect a bad temp sensor. I know all your eyes will pop reading this, but I decided to go ahead and stress the system and see what happens...The GPU temp stabilizes at 108 C max, that's after an hour of Furmark and many hours in a variety of games including: Metro 2033, Witcher 2, BFBC2. Obviously Metro and Furmark hit it the worst, Witcher and BFBC2 seem to max around 105.

THE THING IS THOUGH, the vents continue to be at most warm to the touch, the games continue to run smoothly. There is no indication of glitching or slowdown, there is no artifacting, no system shutdown/bluescreen/whatever, no nothing.

ALSO my CPU (i5-460m) has hit a max of 86 C on Core 1, which doesn't seem bad at all with Intel's TJunction of 105 C (http://ark.intel.com/products/50179). I do believe these components share a pipe right? Seems if one was literally frying that it would heat everything else up pretty good too...

I really don't believe that GPU is running at 108 C. Anyone have thoughts on this?
 
Oh and Colino, just to give you my opinion...50 C idle is not terribly bad, and in a demanding game neither is 90ish + a bit full load. It's a tiny case with a major power-hungry videocard in there. Your numbers don't seem too crazy to me.

But then again, you said you are getting lockups/shutdown when gaming? Mind sharing what games, settings, stuff like that?
 
It can happen with any game, really. I mostly use it to play Final Fantasy XIV (which is really GPU intensive) but even a simple flash game, if it's on for long enough, will cause the laptop to just shut down.

The thing is I've updated the vBIOS but noticed absolutely no difference at all... Should I invest in a cooling panel or something like that?
 
Colino said:
It can happen with any game, really. I mostly use it to play Final Fantasy XIV (which is really GPU intensive) but even a simple flash game, if it's on for long enough, will cause the laptop to just shut down.

The thing is I've updated the vBIOS but noticed absolutely no difference at all... Should I invest in a cooling panel or something like that?
Click to expand...

Are you sure your shutdowns are related to the videocard? Like I said, 50 C idle and 90 C load are really not that unbelievable.

FFXIV would definitely push the card pretty hard. How is that game these days? I only played for about a month after release and got bored. I LOOOOVED FFXI so I'm still holding hope for XIV to turn itself around...

Anyway, you might try running ATITool (http://www.techpowerup.com/atitool/) and do the test to check for artifacts. You can also try Furmark (http://www.ozone3d.net/benchmarks/fur/) burn-in test or Heaven DX11 bench (http://unigine.com/products/heaven/). All of these will stress your videocard to its limits. Watch closely and see if you notice any artifacting in the visuals (weird colors cropping up and spreading around). Also monitor closely and see exactly what kind of temps you are reaching while the benchmarks are running.

If you see any artifacts, or the computer does indeed shut down, try to note what temps it had reached when the symptoms occurred, to see if it really was due to heat. If the card is failing to stay stable through say 30 mins - 1 hr of those tests, it probably is faulty. Would be a good idea to RMA if you're warrantied.

Hope that helps...Let us know and good luck! :)
 
You must log in or register to reply.
Back
Top