AMD Fury no 980Ti killer

[ocre]

Voltage adjustment experiments done on Fury X - abysmal results with scaling vs core clock increases:
http://www.techpowerup.com/reviews/AMD/R...ervoltage/



That's basically half the fps gain for each MHz gain, proportionally-wise!!!  It reminds me of the scaling of overclocking AMD's bullsh*coughs*dozer or pile-o'turd-driver CPU's!



Looks like even stock-clocked HBM1 isn't plentiful already for Fury X cards, poorly optimized with default 512GB/s bandwidth, despite having much larger L2 cache than Hawaii.  



Warning:  W1zzard said that +40mV would be the max for safe continuous operation, even with overclocked watercooling pump.  Anything more overheats the VRM's, especially given the fact that they're already directly cooled by a copper pipe nearby, connected to the watercooler.

OMG!!!!

Bofox, what do you mean by half a fps per each mhz?  it looks nowhere near that.  I would say 50mhz increase gave about a half a fps.  that would be .01 fps every mhz.

but the OMG i first posted........
it is in reference to the power consumption.  Its really shooting up.  You need to sight the article

As you can see, power ramps up very quickly, much faster than maximum clock or performance. From stock to +144 mV, power draw increases by 27%, while overclocking potential only went up by 5% and real-life performance increases by only 3%.

 

that is what it says, but actually he rigged the test.  Its probably even worse.  

In this graph, I'm showing full-system power draw during testing. This test has clock speeds fixed at 1100 MHz for better comparison.

You see, they locked the speed at 1100mhz!!!!
That means that the power consumption results are not showing the card with a 5% overclock.  That 27% figure is not based on the card running with a 5% higher overclock.  He fixed the clock speed at 1100mhz.  With +144mv his gpu clocks at about 1220mhz.  He talks about the extra 5% potential in his power consumption commentary but did not show the actual results of 1220mhz @ +144mv.  The results would have been worse.  Possibly, way worse.......most likely, way way worse

and they are already, really really bad

he says, "Looking at the numbers, I'm not sure if a 150W power draw increase for a mere 3 FPS increase is worth it for most gamers. Smaller voltage bumps to get a specific clock frequency 100% stable are alright, though." BUT------> the power consumption results are from his card at 1100mhz. There is no extra 3 fps, it is not accurate. At 1220mhz, his power consumption could would have been worse