AMD Radeon R9 285 Review: Tonga and GCN Update 3.0

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Could you make a graph to show other cards power consumption? I find it hard to relate to without them.
Click to expand...

Yep! That would be nice. The power usage graph is quite useful as it says how this card use electricity, but it is not very useful, for comparison.
So old fashionable power usage charts would be nice to have!
 
THG, I love you. I come here daily, but you consistently leave out important metrics when reviewing GPUs. Idle power usage in multi monitor scenarios and while playing a video off of YouTube for example. Surely I'm not the only one that cannot stand AMD's latest GPUs because of this. They stay in low 3d clocks, using ridiculous amount of power and heats the GPU to unreasonable idle temperatures. What's worse, simply connecting the second monitor to an AMD GPU prevents Zero Core from working. That makes red cards use 4x more than the green ones. Since my computer spends most time either idling or viewing videos, AMD cards are a big no no for me.

Because of this I returned my 7950 and got 660ti instead. It's not even funny how bad AMD is here and not pointing this weakness is doing nothing to push AMD to fix it. I have to go to techpowerup now to check if AMD addressed either of these deal breakers with its latest card.
 
So old fashionable power usage charts would be nice to have!
Click to expand...
Where is the problem? Take this:

Idle:
http://www.tomshardware.co.uk/charts/2014-vga-charts/22-Power-Consumption-Idle,3614.html

Gaming:
http://www.tomshardware.co.uk/charts/2014-vga-charts/23-Power-Consumption-Gaming,3615.html

Maximum:
http://www.tomshardware.co.uk/charts/2014-vga-charts/24-Power-Consumption-Maximum,3616.html

Efficiency:
http://www.tomshardware.co.uk/charts/2014-vga-charts/25-Watts-per-Index-Percent,3617.html

Idle power usage in multi monitor scenarios and while playing a video off of YouTube for example.
Click to expand...
I've compared a setup with 1, 2 or 3 monitors - no differene. If you try different monitors the chance is 50:50. That means it depends simply at your periphery.

I think, if we have more time to write such reviews (please don't forget the X99 last week), we can addd such info too. But who is watching youtube in HD 24/7? If you find the numbers from TPU more exact, it is your own decision. 😉
 
 


That is not an assumption.....20nm will be GCN 2.0

 


I went back and tried to isolate any non single monitor scenarios in the original review. Aside from diagrams being borderline unreadable, I found no multimonitor scenarios.

As for viewing a video 24/7, I'd normally start many videos be it on YT or off my local HDD. Pause and forget them as I do many things, like reply to an email/sms... I may forget about it, but AMD GPU does not. It will not enter idle state until you shut down every single video. To make things worse (as if they weren't bad enough) flashplayer likes to stay active even after closing internet videos, at least inside the latest firefox. This means that AMD GPU won't go into idle even on a single monitor system unless you shutdown your browser every time after you watch a video. Forget about Zero Core- a simple 2d idle is not going to happen.

Also, I did go to TPU. They showed improvement in single monitor power usage and very limited improvement in multi monitor scenario, but no improvement when watching what they call a blu ray video. Any video of any kind in any GPU accelerated player will cause these clocks/power usage.

I know all of that to be true from personally testing my own 7950 before returning it for precisely these reasons. I'll quote TPU on 285 power usage:

Idle power consumption with a single monitor is greatly improved. With 9 watts, it is now down to levels appropriate for a modern GPU. Multi-monitor and Blu-ray power consumption is still bad - while AMD has made small improvements here, the gap to NVIDIA cards is still gigantic.

In games, compared to the R9 280(X), we roughly see a 10% improvement in performance per watt, which still makes the R9 28x Series the least power-efficient product series on the market. NVIDIA's new Maxwell architecture is roughly twice as power efficient as Tonga.
Click to expand...

For example, with two monitors r9 285 uses 37w vs 10w of gtx 770. Or 51w for viewing a video on r9 285 vs 17w on gtx 770. Their test does not show it, but from personal experience, I know that that watt readouts translate into skyrocketing of "idle" temperatures and fan noise that goes with it. It's flat out unacceptable for AMD to be this bad essentially from the introduction of 6xx series by nvidia, some 3 years ago.
 
With 9 watts, it is now down to levels appropriate for a modern GPU
Click to expand...

9 watts are simply impossible. I've tried 3 different cards and I have the best equipment - no chance to get such numbers. We've tested the same cards in another lab with the same result. - 15 watts and more in idle.

At the end it is not important to have the largest amount of numbers but the right 😀
 
I noticed in the small print that in the test setup it says (underclocked GPU to reference 918 MHz specification for benchmarks). So you used a Asus Strix at the reference for the gaming benchmarks. Although it seems that power numbers are from the Gigabyte R9 285 WindForce 2X with an OC that fairly high 973 MHz (6%), is this what we are looking at in the data? You say, "The Gigabyte R9 285 WindForce OC and its 176 W for gaming comes in almost 40 W lower than a moderately overclocked AMD Radeon R9 280 reference graphics card." Is there the same published data on that 280? While if there’s 40W between a "moderately" OC 280 and one of the highest OC (~6%) I'm really more interested in power for the reference Asus Strix, or B-M score from the Gigabyte. It would be nice to get the power number Sapphire's Dual-X R9 280 OC from just last August 25, 2014 in this same format.

The review says "looks like it’s (the Gigabyte) comparable to Nvidia’s older Kepler-based GeForce GTX 760 (at reference). If winning the efficiency war was the goal, then it most certainly hasn’t been reached". I'm not sure how you can conclude that? I found in the 280X review from Oct 7, 2013 that place the GTX760 at 150W. Do you have these same sophisticated numbers for a GTX760 at reference or say 1137Mhz / Boost 1202 Mhz (6%)?
I’d say we need the data, either...
A) The Gigabyte B-M’s and those average in at the conclusion.
B) Power findings on the Strix at reference
C) Power from a 760 and even better pulled from one with 6% OC.
 
i have a AMD R9 280x 3gb vram (Factory overclocked) by gigabyte and it is wonderful! It runs every game on ultra and is one of the best cards in my opinion!
 


thanks for the head up.
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom