QUOTE(lat_lonelyboyz @ Nov 11 2012, 01:32 AM)
^ No need optional RAM/VRM heatsink if oc gpu ram to max limit?
VRAM = memory/ram
VRM = voltage regulator module which is delivering the power to GPU
VRAM OC on stock/ref cooling should be fine because stock cooler usually comes with VRAM as well as VRM thermalpads.
afaik there's no temp. sensor on the VRAM (VRM has) so you just have to try and error for the most stable VRAM clock.
Start with stock voltage to find the max clock, then only overvolt it to the max.
QUOTE(shepard @ Nov 11 2012, 12:27 PM)
^ I think the max stable clock on GPU core and memory that we achieved is depends on GPU cooling performance as well...The better the cooling solution that we use, the higher OC potential or in other word higher max OC that we can get out from our GPU.
If we use reference AMD HSF, normally due to high temps or the noise we cannot push our card further for higher OC. If we use a waWC on our GC, the max limit will be higher than using stock HSF.
OC potential for our card usually base on our gpu chip potential, temps, noise, cooling performance, gpu ram potential.
Usually
WC > Aftermarket Cooler > Stock HSF for overclocking. Maybe Kevin can dclarify more as I'm also newbie here.
Anything above 1300MHz on air is basically a golden card.
You should definitely not be going above 1.3v unless you're under water.
The higher the ASIC score (recorded by GPU-Z), the better the card performs on air.
The lower the ASIC score (recorded by GPU-Z), the better the card performs under water.
This post has been edited by owikh84: Nov 11 2012, 01:26 PM