Wow I got a bad gpu in my notebook...

I happen to be the "proud" owner of one of the hottest 8600gt's on the forums and probably in the public period. Even out of the Asus C90 crew where alot of us have much higher than normal tempatures I am at the top of the list.

Stock my card would idle mid to high 80's and under just a 3dmark load could go over 100c!!!

So that lead me to modding my c90 to help it cool, and it helped alot, I got over a 10c drop in temps in both idle and load to levels that no longer scare the crud out of me. Not cool by any means but not terminal to the hardware.

So with that mod in place I really wanted to give it a shot and see how well it can overclock now that I can handle more heat.

From the 475mhz core stock and 400mhz memory most people are hitting numbers like 500+ on memory and into the 600s even on core speeds with no problems (and no mods)

Well I just ordered a cooling pad so that I can keep an overclock permanent and safe, but did the testing today to see what to expect.

My MAX overclock is 550core 400 memory! Seriously!!!

At even 430mhz the memory will cause a crash mid way thru 3dmark, at something like 450-500mhz it will freeze the computer the moment 3dmark starts. With the core it can go abit higher than 550 but I can see artifacts sometimes much higher. I tried a few "suicide" runs with higher clocks and noticed my mods really did the temp side some good. I was only 2c higher with the highest speeds I tried, and the safe speeds didnt raise the heat at all.

So that means my cooling pad is going to do me no good (the 50$ zalmans no less) and that my overclock dreams are shattered beyond belife.

I cant say that its anybodies fault, there is no guarentee a card can overclock to any level, some people get lucky and get super cards, others just get normal cards that do decent. I was just very very unlucky and got the very very bad card that less than 1% get that cant overclock at all, and runs hotter than normal...

I was thinking of upgrading to the GDDR3 card, this definitly gives me some incentive, my machine is more than ready for the extra heat it brings, and infact it may run cooler than my DDR2 card considering its borderline defective.

 

Sorry to hear that.

Probably stupid question, I suppose you took proper care of the thermal pad/paste btw/ the sink and the GPU? I get the feeling that the 'defaults' that one gets in a notebook are not so hot (or rather the other way around... if you get the pun).

 

Yeah in adition to the mod to the vents I replaced the stock TIM on the core and heat spreader with AS5 wich gave about a 3c drop in temps.

Unfortunatly tho, this is totally hardware level limitation and not a tempature one. Because the card was running just fine at temps a full 20c higher than it is now with the stock clocks, and things like it freezing instantly with a mild overclock also show this because there is not even a fraction of a second for tempatures to build up.

 

curiously though,

I think im experiencing the same problems... but only in overclocking mode. @500/450 ( standard turbogear enhanced VGA driver... ), im getting artifacts at overclocking mode. When i checked the temps, i was still at the high 70's so definitely not the heat. Lemon GPU as well ? talk about bad luck.

 

Like mine. Core OC is arround 530-550 max (Rivaturner reports that the 475 core is in fact 513 mhz though). Memory isn't stable higher than 470.

The asus guy told me I can RMA it for overheating but I'm gonna wait a bit, or my father will drive crazy. (I TOLD YOU YYOU SHOULD HAVE BOUGHT IT LOCALLY !!!!!!!!!!!!!)

 

Hey vicious,

Im not sure but doesnt the Turbogear VGA program overclock the gpu to over 400 mem....if it does then I would test it that way and try to make it crash or show artifacts. Because the turbogear vga driver is part of the standard setup, the gpu would be considered faulty if it cant run at the speeds the vga driver runs it at......valid rma. just a though...

 

No it overclocks it to the stock 475/400 if you take them off its even lower numbers.

 

DDR2 Cough Cough, stone aged video memory, pre-mainstream OCing. GDD3's calling you.. very close to 8700 speeds.... yeah baby.

 

Does it overheat at all? That would be a good reason for RMA also.

 

overheat would mean that it gets so hot that it wont function under normal conditions. So no, its hot as hell, it was over 100c load before my mods, but it still didnt crash or anything.

 

Do you think there is a problem with your heatsink mechanism Vicious? It just seems as though you have had issues with every single part of the computer...I had similar problems for a while after playing with the GPU, but after re-fixating my heat sink and tightening the bolts it was better. (Not to say that those things will fix your problems, but may be indicative that there was a flaw in your heatsink since day 1)

 

I think I already asked him a related question and his issues aren't caused by temperature, but hardware limitations... (the GPU doesn't run at those frequencies).

 

I just meant overall, considering all of his problems to this point, aggregate, with the high stock clocked temperatures considered. By the way, what mods did you do Vicious?

 

Damn 90º-100º is a lot...

My 7950 gtx with core/mem at 620/815 in Colin DiRT i got 93º...

I try a Cooler Master infinite and it works fine (i got 8º to 10º less).

I use a Infinite Cooler Master and Logitech Alto at same time... and i got like 12º - 15º less... Besides the computer at same level of my head...

regards.

 

vicious put 9 holes on it, and re did the thermal compounds ...

hmm might really be a bad part *hmm*, send it to asus, the videocard's warrenty is seperate from the laptop... well, also u can try putting the laptop in the freezer and try everything again o.o...

 

Mine does the same exact thing w/ a laptop cooler. At idle it's at 83c-86c. In game upwards of 100c. I think better fan control would improve things but I definitely think something is up.

 

idle = 80c in mine. Even with good airflow underneath there's only a few degrees of difference. I actually thought of an improvement to the "Holes in case" cooling idea but it requires a few more things than just a drill bit and some patience. I'm sketching out some rough ideas as to how it works but keep in mind that I'm planning something a little more extreme than just holes and I'm also calling in some help from otehrs to get this done right cause it's a little permanent...

 

My gddr3 card idles at 76 or 77 with a lapcool2 underneath, about 81 with nothing. I've seen it at mid nineties under load. But the cooler was a good idea 3 years ago.

 

Everybody wants me to send it back (the card) but in theory its not defective, there is no operating temperature in the warranty, if it doesn't melt or crash its considered "ok" I am upgrading to the GDDR3 card probably.

I am borrowing one for a test run, my only concern with that is that only the stock drivers are giving the right performance (4000 points 3dmark06) but updating to new drivers drops it down to 2000 points and thats below the DDR2 card!!!

So with new games like bioshock that required a driver update, the orange box had several fixes in the 163.75 drivers, plus who knows how many more games to come. You may get totally shafted being unable to play the game due to old drivers, or update the drivers then suddenly be playing with the power of like a 8400gs or worse!

Its really hard for me to make that call. Right now I am looking at the HD2600 as being the best card in reality but there is virtually no information on it and I dont even know where to buy it.

 

I'm going for an ATI card as soon as I can get one if it stands up to the one I have or is at least close.