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

Page 79 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

tkrushing

Junior Member
Jan 10, 2008
14
0
0
We need more actual performance benchmarks though. Definitely promising...

3D Mark can be affected my memory clocks and so forth in odd ways. HBM could either be over or under-represented in the real-world benches.

Can't wait to see some real reviews soon. Any benchmarks should be read with caution until then IMO.

100% agree. Maybe it gives us a ballpark idea though.
 

AnandThenMan

Diamond Member
Nov 11, 2004
3,991
626
126
Efficiency does not mean less power draw. It means better performance for a given amount of power usage. If you take a card that draws 100W and gets 50fps, and another card that draws 100W but gets 100fps, the second card has 2x the efficiency.
You could have a GPU that draws 100 milliamps and gets 0.4 fps. But it would still have the best perf/watt out there, of course you would not want to game on it.
 

crashtech

Lifer
Jan 4, 2013
10,692
2,289
146
You could have a GPU that draws 100 milliamps and gets 0.4 fps. But it would still have the best perf/watt out there, of course you would not want to game on it.
This has been phrased and rephrased, perf/watt is only relevant when comparing similarly performing parts, this should be obvious to all, no?
 

exar333

Diamond Member
Feb 7, 2004
8,518
8
91
Though perhaps lacking above that, or at high AA settings.

CF under DX12 will be interesting though, if the sharing of memory works well.

You can see 4K is the limit for the card on the graph. 5K drops a good amount in comparison to TX/980Ti and 8K tanks completely and the 390X is way ahead due to the 8GB.

Add different games, textures and AA and the picture gets more and more complicated. I think this will be a focus in the reviews as a good deal of people will want to be confident the 4GB isn't a limiting factor (myself included).
 

AnandThenMan

Diamond Member
Nov 11, 2004
3,991
626
126
This has been phrased and rephrased, perf/watt is only relevant when comparing similarly performing parts, this should be obvious to all, no?
Of course, but for some reason the discussion about perf/watt is sometimes seen in a vacuum, it is one metric that has to be weighed against actual performance. I mean I can overclock my CPU to silly levels and I get the performance, but power usage goes up exponentially.
 

Enigmoid

Platinum Member
Sep 27, 2012
2,907
31
91
Efficiency does not mean less power draw. It means better performance for a given amount of power usage. If you take a card that draws 100W and gets 50fps, and another card that draws 100W but gets 100fps, the second card has 2x the efficiency.

No duh. but the 980 is not twice as efficient as the 770 no matter how you cut it.


Though perhaps lacking above that, or at high AA settings.

CF under DX12 will be interesting though, if the sharing of memory works well.

Need some magic sauce to get any large amount of data over the PCIe bus however. Its 16 GB/s for 3.0 x16.

You can see 4K is the limit for the card on the graph. 5K drops a good amount in comparison to TX/980Ti and 8K tanks completely and the 390X is way ahead due to the 8GB.

Add different games, textures and AA and the picture gets more and more complicated. I think this will be a focus in the reviews as a good deal of people will want to be confident the 4GB isn't a limiting factor (myself included).

Reviews will show that now 4GB isn't too bad. However, it looks like the release of shoddy console ports could easily push future games over 4 GB, especially at 4 K.

Gotta love AMD marketing. 4GB is plenty for Fiji, meanwhile look at our 4K ready 8 GB Hawaii chips (which add ~$30-40 to the BOM). That inconsistency. D:
 

zagitta

Member
Sep 11, 2012
27
0
0

gUaIaG2.png


Interestingly enough it looks like project quantum uses a custom ~ITX sized motherboard, take a look at the oddly placed power connector next to the network interface, and it also looks like an Intel CPU to me :D
 

.vodka

Golden Member
Dec 5, 2014
1,203
1,537
136
At least they're honest and realize their CPUs have nothing to do in a scenario like that, driving such powerful GPUs. Nice way of thinking. Should change once Zen products replace the faildozer family. (Carrizo is something else, but not suited for high power scenarios)
 

happy medium

Lifer
Jun 8, 2003
14,387
480
126
So I skimmed over the last few post, so basically we really know nothing more about actual performance of the Fury than we did yesterday? Just more guesses and speculation?
 

Despoiler

Golden Member
Nov 10, 2007
1,968
773
136
No duh. but the 980 is not twice as efficient as the 770 no matter how you cut it.




Need some magic sauce to get any large amount of data over the PCIe bus however. Its 16 GB/s for 3.0 x16.



Reviews will show that now 4GB isn't too bad. However, it looks like the release of shoddy console ports could easily push future games over 4 GB, especially at 4 K.

Gotta love AMD marketing. 4GB is plenty for Fiji, meanwhile look at our 4K ready 8 GB Hawaii chips (which add ~$30-40 to the BOM). That inconsistency. D:

What is inconsistent when they use two different types of memory? False equivalence is a logical fallacy.
 

LTC8K6

Lifer
Mar 10, 2004
28,520
1,575
126
So I skimmed over the last few post, so basically we really know nothing more about actual performance of the Fury than we did yesterday? Just more guesses and speculation?

I would say that's a fair comment.

300 series seems to be largely an unexciting rebrand.

Fury is still mostly a mystery as far as performance. It's been described in terms that could be taken more than one way.

It's mostly described in terms of the 290X.

We have some synthetic benches, that probably don't reflect gaming.

So, I would say it was a very interesting reveal, but so far, not of Cat's Ass caliber.

Subject to change with more data, of course.
 
Last edited:

MrTeal

Diamond Member
Dec 7, 2003
3,917
2,704
136
Nice!
FuryX faster than TITANX and fury pro at 980TI level for 550usd.

Fury Pro will be an interesting card, hopefully that triple fan cooler is good. The Fury X costs $100, but that's really only 18% more. For your extra $100 you get the CLC and 14% more shaders if the 4096/3584 split holds up. Depending on how they test, this might be one of the few times where the top end card is actually the better value and not just the better performer.
 

ShintaiDK

Lifer
Apr 22, 2012
20,378
145
106
What is inconsistent when they use two different types of memory? False equivalence is a logical fallacy.

And? FP, EDO, SDR, DDR1-4, GDDR1-5 had zero difference. But somehow HBM changes all that? 4GB is 4GB nomatter what type of memory. I can guarantee you it was the original plan to only ship it with 4GB. But thats because Hynix couldnt deliver.
 
Last edited:

therealnickdanger

Senior member
Oct 26, 2005
987
2
0
It's my somewhat educated opinion that the 8GB VRAM on 390/390X is there to combat the 970/980. We all know people that buy 4GB GT 730 graphics cards because it has "m0ar vrams". It's a one-up against the 970's 4GB* and 980's 4GB... perhaps even the 980Ti's 6GB. These things are already on sale at Best Buy and people that know nothing about GPUs are buying them (as evidenced by the videos posted so far), so it's clearly working.
 

.vodka

Golden Member
Dec 5, 2014
1,203
1,537
136
I can only wonder what the comments would be if nV was the first to HBM1, and AMD would've had decided to wait until HBM2 next year.


Ahh, reviews, where are you?



EDIT: That Fiji X2 board looks amazing. Hmm, the VRM area seems similar between it and the 295x2, I count a similar number of chokes on both:

Radeon-R9-295X2-PCB-Front-TPU.jpg


This time at least the VRM should be watercooled too, considering how compact everything has become.
 
Last edited: