• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

[Rumor (Various)] AMD R7/9 3xx / Fiji / Fury

Page 86 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Can someone play with the Ultra settings on FC4 and do some benchmarks to give us an estimate of the effect of tweaking the "Ultra" settings. We already know that on "asterisk free Ultra" settings we get these results:

74776.png


35% faster than 980ti.
 
I wonder if Fury Nano is that misterious Captain Jack sample from Chiphell leaks, it had ~16% higher performance than the GTX 980, with an average of 197W of power consumption in games.
 
Jaydip said:
So a question for you guys, I have never installed any wc cards before so I just buy and install it like regular cards? (maybe sounds like a weird question though 🙂 )
Click to expand...

Plug it in, find a free 120mm fan area (or replace a current 120mm fan exhaust) for the radiator, works best as exhaust, so it acts both as heat exhaust for the GPU but it will be a air exhaust for your case. Win win scenario.

Btw, that * just says: Refer to page X for system config (Rig used to run bench). Ultra would imply the standard Ultra option in FC4.
 
Tapoer said:
I wonder if Fury Nano is that misterious Captain Jack sample from Chiphell leaks, it had ~16% higher performance than the GTX 980, with an average of 197W of power consumption in games.
Click to expand...

Most likely it is.

Faster than 980 with similar power usage. Which led to me claiming it would be ridiculous if true, because that would require a MASSIVE efficiency jump from R290X, even bigger than the Kepler > Maxwell efficiency jump.

Turns out its true, 2X eff!

ChipHell also had Fiji "cut" (air version), so it seems they got their hands onto all 3 Fiji ES SKUs along with Titan X/980Ti. These chips were all produced in Q3 2014. Definitely the leaker would be an insider from an AIB that sells both AMD/NV products.
 
Silverforce11 said:
Most likely it is.

Faster than 980 with similar power usage. Which led to me claiming it would be ridiculous if true, because that would require a MASSIVE efficiency jump from R290X, even bigger than the Kepler > Maxwell efficiency jump.

Turns out its true, 2X eff!
Click to expand...

They say up to 2x. And we all know what that means. Kepler->Maxwell was the same.

Also remember Nano is a 600mm2 die. I am pretty sure you can also make a GTX980Ti perform like a GTX980 at 125W or so. Intel did the exact same thing with underclocked GT3 in some SKUs that performed like GT2 SKUs. Just with much lower power consumption. Just throw more lower clocked transistors at the problem. Works great in GPUs in performance/watt metric.
 
Tapoer said:
I wonder if Fury Nano is that misterious Captain Jack sample from Chiphell leaks, it had ~16% higher performance than the GTX 980, with an average of 197W of power consumption in games.
Click to expand...

So its about the same in efficiency as the 980? If you overclocked a 980 about 16% (or got a factory OC version), you would have similar TDP.

I am still unsure why everyone is so excited about the card? Its fine, but its nothing we haven't seen before in either performance, efficiency or TDP. If this was 125w (as other posters seemed to suggest previously) then this would be pretty special. ~200w for 980 performance? Yeah, been there...
 
exar333 said:
So its about the same in efficiency as the 980? If you overclocked a 980 about 16% (or got a factory OC version), you would have similar TDP.

I am still unsure why everyone is so excited about the card? Its fine, but its nothing we haven't seen before in either performance, efficiency or TDP. If this was 125w (as other posters seemed to suggest previously) then this would be pretty special. ~200w for 980 performance? Yeah, been there...
Click to expand...

Well given that as soon still as last week, Perf/Watt was apparently the most important metric. But of course as of this week now its meaningless
 
exar333 said:
So its about the same in efficiency as the 980? If you overclocked a 980 about 16% (or got a factory OC version), you would have similar TDP.

I am still unsure why everyone is so excited about the card? Its fine, but its nothing we haven't seen before in either performance, efficiency or TDP. If this was 125w (as other posters seemed to suggest previously) then this would be pretty special. ~200w for 980 performance? Yeah, been there...
Click to expand...
Except a gtx 980 is a larger card. Far larger. We'll have to see reviews though.

I do agree I don't get what's so exciting a out fiji nano but that's because I don't see why people are putting such a massive premium on power consumption and pcb size along with air cooling.

Gtx 970 mini itx is probably not too far behind fiji nano anyway. It's all about fiji wce I really don't see why you'd spend this much money on a gpu and not step up to the wc edition since it just adds so much more for the incremental cost. We'll have to see reviews to confirm that though.
 
exar333 said:
So its about the same in efficiency as the 980? If you overclocked a 980 about 16% (or got a factory OC version), you would have similar TDP.

I am still unsure why everyone is so excited about the card? Its fine, but its nothing we haven't seen before in either performance, efficiency or TDP. If this was 125w (as other posters seemed to suggest previously) then this would be pretty special. ~200w for 980 performance? Yeah, been there...
Click to expand...

Just remember, AMD's TDP != NV's TDP.

R290X is 300W, uses ~230-250W in games.
 
Headfoot said:
Well given that as soon still as last week, Perf/Watt was apparently the most important metric. But of course as of this week now its meaningless
Click to expand...

Efficiency is great, but this isn't showing us anything new. Its competitive, and if priced lower or the same as the 980, is a great alternative. Not disputing that. Just saying its not anything special, but is a lot better vs. 290/290x in efficiency.

The Fury and Fury X are exiting. This is ok, but just the same as the 980/970 in a smaller package. Cool, but not game-changing or anything.

Honestly, instead of offering an alternate to the 390/390x, it would have been a lot more interesting if THIS was the 390 and 390x. Seems a bit odd to have another SKU in the same performance area, but with totally different tech and TDP. We see this in CPUs, but usually they are lower-freq parts with lower voltage rather than a very different arch.

The Nano should replace all the 3xx parts faster than the 380 IMHO. The power consumption and efficiency brings more to the table vs the 8GB upgrade that adds cost to the 390/390x.
 
Last edited:
Not sure why you think 16% (leak from last year) more performance at similar power use (the actual leak chart had them very close in power usage in games) is the same by comparing it to an OC 980. Are you implying Nano can't OC?

AMD is late to the party but if they can show up with Nano and its ~15-20% faster than 980 with similar power use, that's a great achievement. In fact, something like that was deemed impossible by many.

People were seriously stating that AMD cannot improve efficiency by much due to lack of R&D funds etc, that Fiji is going to be a 375W monster (or worse, some people felt R290X + 50% = 400W+).

Now if they show up with 2x efficiency gain on the same node, that's nothing short of amazing.
 
Silverforce11 said:
Not sure why you think 16% (leak from last year) more performance at similar power use (the actual leak chart had them very close in power usage in games) is the same by comparing it to an OC 980. Are you implying Nano can't OC?

AMD is late to the party but if they can show up with Nano and its ~15-20% faster than 980 with similar power use, that's a great achievement. In fact, something like that was deemed impossible by many.

People were seriously stating that AMD cannot improve efficiency by much due to lack of R&D funds etc, that Fiji is going to be a 375W monster (or worse, some people felt R290X + 50% = 400W+).

Now if they show up with 2x efficiency gain on the same node, that's nothing short of amazing.
Click to expand...

I don't know if we have an 'official TDP' but the slide posted on the thread states 175w 'typical power use' (whatever that means...😛) The 980 is 165w and a factory-OC model would typically be about 10-15% faster and would use 10-15w more power. So that's what we get with the 980 already, plus some additional room to OC.

We don't know how the nano performs, but if you compared a stock nano to a standard 980 with some freq bumps, you have simiar power and performance. That's all I am saying.

Am I missing something amazing from this card?

Edit: If the TDP is even lower than 175w, then you have a class-leading TDP part with mid-high range performance. If it performs similarly as a 980, with similar power, its competitive. That's awesome! But not something to freak-out about. You could have had all that 9 months ago!
 
Last edited:
Silverforce11 said:
Not sure why you think 16% (leak from last year) more performance at similar power use (the actual leak chart had them very close in power usage in games) is the same by comparing it to an OC 980. Are you implying Nano can't OC?

AMD is late to the party but if they can show up with Nano and its ~15-20% faster than 980 with similar power use, that's a great achievement. In fact, something like that was deemed impossible by many.

People were seriously stating that AMD cannot improve efficiency by much due to lack of R&D funds etc, that Fiji is going to be a 375W monster (or worse, some people felt R290X + 50% = 400W+).

Now if they show up with 2x efficiency gain on the same node, that's nothing short of amazing.
Click to expand...

Totally agree. 175w is not less than the 980 though. Again, we don't have official specs yet or reviews.

If it can perform the same as the 980 with 20% less power-consumption, its a win. If it can perform 20% better with the same power-consumption, its a win. If its slightly more power for slightly less power, or about the same for both, its competitive but not amazing. That's all I am saying...
 
exar333 said:
Totally agree. 175w is not less than the 980 though. Again, we don't have official specs yet or reviews.

If it can perform the same as the 980 with 20% less power-consumption, its a win. If it can perform 20% better with the same power-consumption, its a win. If its slightly more power for slightly less power, or about the same for both, its competitive but not amazing. That's all I am saying...
Click to expand...

A 980 uses ~185 W during gaming loads.
 
The take home message for me was - Fury X is 1.5x more power efficient than 290x and Lisa mentioned 400 amps!

I'd like to see gaming results for this setup.
 
jackstar7 said:
So can we officially add perf/watt to FCAT results to the list of things that no longer matter?
Click to expand...


Perf/Watt is hugely important, and the efficiency gains AMD appears to have made are more important IMO than HBM for Fury's success.

At this point, we're just waiting for independent confirmation of their claims.
 
exar333 said:
True. Some use closer to 160w though.... depends on the review.
Click to expand...

I think you mean it depends on the game because the card isn't just going to use that much less for one person versus another. We'll find out when nano releases, but if it actually uses 175 W during heavy gaming, it will use less than a 980.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top