HD5850 overclocking? how soon can it be done?

[sanchaz]

I have a hd 5850, ive had it for about 4 days 2 of those days it was running at full load.

 

My question is how soon can i overclock?

Will overclocking it now make me have weaker results than from a months time?

 

Does the card need some sort of mileage before it will achieve its maximum OC performance?

[imm0rt41]

wasn't aware you needed to wait so long lol. I took my 6970 out of the box, installed it, played games on it for a week, then I OCed the hell out of it. Was able to break 1GHz/1.5GHz on it, so that was nice.

[Overclocker16]

Yeah I don't think you really need a break in type period. I think for most people, they play on it stock and then OC the hell out of it just to see for themselves how much better games are.

[gotdamojo06]

 

ive never waited to OC any video card...

 

I usually do run benchmarks on it first at stock speeds and then clock the crap out of it!

[sanchaz]

 

ive never waited to OC any video card...

 

I usually do run benchmarks on it first at stock speeds and then clock the crap out of it!

 

ok i was worried cause i started overclocking and then i remember this.

[Black64]

Why would you need to wait to OC? That doesn't make any sense to me at all. I always OC right out the box.

[Starbreaker]

Indeed fun to read. I never ever would have had come to the concluion that i would need to wait a certain time before i can start to OC a piece. From my view it only makes sense to test out the pieces first at stock speed in order to find out of there is any manufacturer failure. So you can bring it back for warranty in term there is something wrong or unstable, because a piece should never have any flaws at stock speed.

 

But as soon as the initial control is done, then i see no reason why not to start OC. In term there is a failure, then it wouldnt OC at all, and even at low OC it would stop working. So its indeed proof of quality when it can OC high with as less volt as possible. I guess thats the fun every OCer is experiencing. Testing out how much further they can push the pieces without increase on volt. And then many will start to go even further, while others stop at that point already because its enough for them.

 

Anyway, i see no point why to wait a certain time before OCing because if you do enough of stress tests at stock speed, then any failure surely will be revealed quick. I dont know other reasons for delaying it.

[IVIYTH0S]

I usually OC when i come to a game I can't max or one that has frequent frame drops but I could OC right from the first use.

[wevsspot]

Sanchaz, there isn't any magical amount of time for when you can start overclocking a new cpu or a new gpu. I usually wait a few days to make sure that everything is stable at stock settings and that it doesn't appear that I'll have to RMA something. But once that period is done, if I'm gonna o/c I start to o/c

 

I guess there might be a valid argument for slowly increasing clock speeds and voltages (like a break in ramp up) - but I've never personally experienced it. I figure that all the stress testing that I'm doing while I'm ramping up the frequencies and voltages counts for a break in period.

[Starbreaker]

Well, in term of CPU, new boards got a dynamic OC option, so it will only OC when the CPU is at very high load. So whats the point of it? Its here so the CPU will have lesser work in usual terms and it will live longer. In case it does dynamic OC i see no point why to wait with OC because for the CPU the difference is very low, since it will only get leeched at high loads.

 

And for the GPU, i dont think anyone have to worry as long as it is a good double slot cooler design. There isnt much that can truly go wrong, but OCers do well know that you always have to OC step by step and never huge steps, OCing have to be made carefully and always lot of tests between. In order to slowly move upward and find the optimal clock, some may say "its enough, i dont need even more" and stop before the true limit is reached (such as me). Apart from that not to much to worry as long as everything is done carefully and sufficient cooling.

 

Ah yes: I suggest to make stress testing with Prime 95 and Ungine Engine, But not both at once because when Prime is taking away all the CPU power, then you cant anymore put high stress on the GPU. So obviously do both tests on its own and do it after every single step. Try to run Prime 95 and Ungine Engine for 15 min at least. Ungine can be run countless times, dont stop between, just let it run and go dinner break, same for Prime. Please be aware to control the temperatures manually the first 5 min, dont go away until that is done. if there is any critical value, stop it imidately. For CPU 80° should never be exceeded, the GPU can handle up to 100° but i wouldnt go above 90°. Ofc thats just max value where it starts to get critical, i would try to stay much lower than that. I suggest to get GPU-Z and REAL TEMP GT, so the GPU will heat up a lot and more likely to cause instability if there is any, same for CPU. Good luck.

 

Btw:; I do not suggest Furmark, because its not truly making a balanced test of any parts of the GPU architecture, Ungine is much better at that, including tesselation. Apart from that, Furmark is pretty much driver busted because the driver may detect it and reduce the power going to the GPU as a counter measure.

Edited June 14, 2011 by Starbreaker

[sanchaz]

ok guys thx for all ur answers i have OC'ed my GPU to 850 cant go, any futher or drivers crash.

I was however using furmark ill have to give Uengine a try dont know if it will make a difference...

 

my other card a Powercolor 5850 go it up to 900 stock voltage, this sapphire xtreme 5850 will only go to 850, but i guess it really depends on the brand and batch of the cards.

[IVIYTH0S]

You got it!! Every card is different

I'd try some games on it, that is the best way to gauge stability