Originally Posted by Alastair
I said he should go Fury X. Thoughts?
Fury X IMO.RX 480 PCI-E slot power usage (Click to show)
RX 480 cooler (Click to show)
This is how I see it, RX 480 is hammering the PCI-E slot even for average figure.Collated THG power data for 390X/Nano/Fury X/RX 480 (Click to show)
Do bare in mind the data from THG is being recording at very high speed, the spikes are not as relevant as average power usage, PC perspective
's article uses a better method IMO and also has paragraphs explaining why their test method is so.
I don't believe they can curb RX 480 to be like those past cards on PCI-E slot power usage. IR3567B AFAIK does not differentiate on powering VRM from PCI-E slot/plugs. ROM PowerPlay PowerLimit does not contain values to separate PCI-E slot/plugs. To me via driver I reckon they will either lower voltages on card or exert tighter PowerTune algorithm to reduce power usage but still have higher usage compared with past cards. I reckon it is a PCB design flaw.
The RX 480 cooler is pants, if upgraded = money. In the UK Nitro RX 490 1 etailer is preordering at £250, I've seen Amazon warehouse deals on Fury X for £300 recently, I'd rather pay the £50 extra and get a Fury X open box.
Thoughts on RX 480 vs Fury X on perf.per watt (Click to show)
Edited by gupsterg - 7/3/16 at 8:58am
I usually use a UK magazine's reviews for purchases and they are sometimes on Bit Tech, in their test setup
they used later drivers for 390X & Fury X. Comparing say 1080P results only in Hitman did RX 480 gain over Fury X, I think a driver thing with Fury X possibly. Then rest of test the Fury X is faster than RX 480 by (used min FPS):-
AOTS 18% , Fallout 4 36%, Division 34%, Warhammer 11%, Witcher 3 26%
So average for those set of games = ~25% better performance. For the power usage test they take total system draw and run Valley at 1440P, Fury X is 50% faster whilst system uses 50% more power. TPU's performance per watt chart
may not be that accurate, right at the top it states:-
We used the relative performance scores and the typical gaming power consumption result.
Typical gaming power consumption result would be 163W RX 480 and 246W Fury X on page 22
of review and I think they use that wattage and apply to charts on page 24
to come up with performance per watt. As they use Metro: Last Light @ 1080P for the typical gaming power result I thought I'll use FPS of that but there is no data to calculate perf.per watt.
How I see it there could be games which don't stick to highest DPM state on Fury X = less power usage where as due to RX 480 being lower SP, etc it maybe sticking to higher state. So better performance per watt data would be measuring each games power usage/performance and coming up with figures. I'm reckoning Fiji is pretty good on performance per watt.