yes we can do it as we like bro..i'm doing it by reducing the core slowly..any way will do

that's why I always OC one by one..first I'll find the max stable core, then find a max stable mem clock...later will combine both max OC and reduce the mem first by 10-15MHz, if not stable then reduce both by 10mHz and rerun the stability test again...normally should be okay one, be sure to monitor the temp though and you'll be fine.

measure temp on gpu surface and mem surface, or just the ambient temp inside case? or cpu sensed temp?

umm..what I mean is using a software like MSI afterburner or GPU-Z to monitor gpu temps during gaming sessions..the download link is already at first post and you've to turn on OSD in MSI afterburner setting..also click the checkbox unlock voltage monitoring on settings page if you want to monitor Vddc also



This post has been edited by shepard: Jun 23 2013, 09:59 PM

My result on Unigine Valley 1.0: so poorrr



CPU: AMD FX-8350 No OC
GPU: HD 7870 XT Core: 1020 Mhz ; Mem: 1520 Mhz

Try overclock your cpu a bit? Is that the max OC of ur graphics card?

i've try max out my GC and when i run unigine again it gave me strip line on my screen.

lower the memory clock perhaps...OC your cpu to release some more fps

will that make the picture more sharp or better?

if u seeing soem weird lines or images appearing during stress-testing, most probably your gpu oc isnt fully stable. U need to lower either core/memory clock.

how often this oc will lead to gpu failure or need to reflow the gpu?

becos ur card is not capable to be overclocked that high, it really depends on the card, chip and the silicon themselves...like a cpu lottery if u get a golden chip u can use lower vcore for higher overclock...same concept applied...hope that helps...

will it be the same gc, same model gpu, test under the same pc, will it be the same oc results?

u mean same gpu card tested under different pcs? I guess more or less u will get the about same clock..nt so sure thou

no.

meant same configuration PCs, but put the same model gc to oc, same results come out?

it will vary? one graphics card or two u mean?

1) same gc but tested on PCs(same configuration)
2) same gc model but tested on PCs(same configuration)
really hv no experience to oc gc

It wont be the same OC results bro, different chip on graphics card means different OC capability even though same GPU model and same brand being used to compare. The OC difference comes from each gpu being tested and not from other parts for your case.. for example:

AMD hd7970 reference #1 - highest oc @ stock volt - 1200MHz
AMD hd7970 reference #2 - highest oc @ stock volt - 1125MHz

both cards are same model and same brand from AMD but max oc for both card is differ because of different gpu(chip) inside..

a good reference

how about those oem gc, diffferent brand from amd / nvida but they share the same model no.

much difference?

hmm did you mean different brand GC but using same model number? I dont think this is even the case..unless same oem PC model number then yes. there will be a difference between every single graphics card whether it is same brand, same model, diff brand or diff model bro..some might have similar oc capability but it's random and or could be batch related so its more on to the luck..If we're lucky enough we might just purchase a very2 good OC card or if really unlucky we get the worst oc card

This post has been edited by shepard: Jul 2 2013, 08:19 PM

individualistic

that's the charm to oc, same item but come out different results