AMD Announces Q4 2016 Earnings [Anandtech]

[Bacon1]

Those are cute end notes....oh wait the RX 270 doesn't draw nearly 180 watts on those tests and the rx 470 draws more than 110 watts. Simple math shows AMD's footnotes are either lies or are incredibly misleading and still untrue.

So you've re-run those tests and proved them wrong? They list out the exact games and test conditions, please show us your source proving them wrong. Different games use different amounts of power, I'm sure you know that. But since this is completely off topic, why don't you create a new thread with your proof showing them wrong.

https://www.computerbase.de/2016-07...mm-leistungsaufnahme-des-gesamtsystems-xcom-2

390x for example goes from 393watt -> 460watt system depending on the game.

 

[AtenRa]

Those are cute end notes....oh wait the RX 270 doesn't draw nearly 180 watts on those tests and the rx 470 draws more than 110 watts. Simple math shows AMD's footnotes are either lies or are incredibly misleading and still untrue.

Just because you dont like the way they measure perf/watt (board power) doesnt make them lies or misleading.

 

[zinfamous]

Up 16% today. Not too shabby...

 

[MrTeal]

Just because you dont like the way they measure perf/watt (board power) doesnt make them lies or misleading.

He's not complaining about that, it's the second comparison where they don't use system power or actual measured GPU power, they just appear to use the TDP for the card.

 

[AtenRa]

He's not complaining about that, it's the second comparison where they don't use system power or actual measured GPU power, they just appear to use the TDP for the card.

Thats what i said, just because they used Board Power TDP doesnt make it a lie or misleading its just a different methodology. Yes it does produce a higher perf/watt than other methods (measuring total system power or actual Graphics Card energy consumption etc) but they said "up to" and with their method they got up to 2.8x. Im not saying this is the method i would use, but since they provided exactly the conditions of their methodology its valid.

 

[MrTeal]

Thats what i said, just because they used Board Power TDP doesnt make it a lie or misleading its just a different methodology. Yes it does produce a higher perf/watt than other methods (measuring total system power or actual Graphics Card energy consumption etc) but they said "up to" and with their method they got up to 2.8x. Im not saying this is the method i would use, but since they provided exactly the conditions of their methodology its valid.

Ah, my apologies then. I thought you were referring to the P11 demo as they specifically call that board power in the footnotes while not mentioning it in the P10 notes.

I'd disagree about your second point though, I would definitely say it's misleading as they chose really the only pair of GPUs and metrics that could produce those kinds of numbers. Most of that is due to the overstating of the board power of the amazingly efficient Pitcairn card, but they also use a board power for the 470 that it will regularly exceed even in gaming loads. By any objective standard Polaris 10 is nowhere near 2.8x the efficiency of Pitcairn or any comparable die from their 28nm line, and claiming that it has up to 2.8x perf/watt when you need to clarify in the footnotes that you really mean perf/rated board power is disingenuous especially given the arbitrary nature of board power ratings.

 

[Qwertilot]

Of course it was misleading! It was marketing for pities sake