someone leaked couple of GTX1080 benchmarks of Ashes, DX12

[zMeul]

source: http://www.ashesofthesingularity.com/metaverse#/personas/4a36165f-4e15-4c4f-af35-7340095285f5/match-details/eb04c772-0095-4429-b1de-199d6e00c4c6

 

he has a bunch of runs, broken down: (video card / score / avg framerate) all of them are at 1440p with crazy quality preset 

• GTX 1080    4600    47.6
• GTX 1080    4900    50.0
• "R9 Fury"    4100    42.6
• "R9 Fury"    4100    42.4
• GTX 1080    4600    47.6
• "R9 Fury"    4100    41.8
• "R9 Fury"    4000    41.1
• "R9 Fury"    3900    40.7
• "R9 Fury"    4000    40.8
• GTX 1080    4900    49.6
• GTX 1080    4800    49.3
• GTX 1080    4600    47.4
• GTX 1080    4700    47.6
• GTX 1080    4600    47.4
• GTX 1080    4600    47.4
• GTX 1080    4600    47.3
• GTX 1080    4900    49.7
• "R9 Fury"    4100    42.0
• "R9 Fury"    4300    44.4

 

some system details:

• CPU: i7-5930K CPU @ 3.50GHz
• 32GB RAM

 

I'd wait on some other benchmarks

the differences against the Fury X (I presume) is quite negligible; and Ashes is not the most trustworthy benchmark as multiple runs can produce different results 

 

---

 

May 13th - Videocardz has some 3DMark results: http://videocardz.com/59871/nvidia-geforce-gtx-1080-3dmark-firestrike-and-3dmark11-performance

Spoiler

 

[eLucid]

Seems legit.

[Philcat101]

meh

some troll might have exploited system properties to rename the GPU

er somthin like that

/xD

[zMeul]

7 minutes ago, Philcat101 said:

meh

some troll might have exploited system properties to rename the GPU

er somthin like that

/xD

there's also a nVidia employee, Sean Pelletier, with an Ashes test: http://www.ashesofthesingularity.com/metaverse#/personas/81b0e74e-f164-4349-8c62-fc1ddfaa9df9/match-details/4b660527-0716-4c31-9c30-27a3334e688f

1080p, high

[Misanthrope]

It looks like what I'm thinking: compute performance still sucks but they're just brute forcing their way with extremely high clock speeds it seems. Not a bad way to transition unless you really like the like 2 or 3 DX12 games that will really benefit from better compute performance.

[handymanshandle]

Hmmm interesting. I won't cast it off as false yet. But remain skeptical.

 

[Tech_Dreamer]

when they said 2x performance over 980..

[zMeul]

Just now, Tech_Dreamer said:

when they said 2x performance over 980..

they clearly said VR

[Tech_Dreamer]

Just now, zMeul said:

they clearly said VR

shit! so t'was just another amped promo

[PerfectTemplar]

15 minutes ago, Tech_Dreamer said:

shit! so t'was just another amped promo

They also said that performance per watt is still performance. So they added the frame rate increase + the lower power consumption to come up with 2x performance. It's pretty damn good though. It was 3x performance for VR.

[FakezZ]

46 minutes ago, zMeul said:

they clearly said VR

I didn't see them specify it was for VR only... Where did you get that from?

[Dietrichw]

8 minutes ago, FakezZ said:

I didn't see them specify it was for VR only... Where did you get that from?

they had another slide which had it broken up by VR and a few Gameworks games. The VR performance was through the roof (due to the software they introduced) and the games were more modest and what was expected

 

-edit here it is 

 

The games are assumed to be best case scenario but are more reasonable when compared to a 980. So about the same leap from a 780ti to 980 now with the 980ti to 1080. It just looks better to compare it to the 980

[FakezZ]

1 minute ago, Dietrichw said:

they had another slide which had it broken up by VR and a few Gameworks games. The VR performance was through the roof (due to the software they introduced) and the games were more modest and what was expected

LOL another pitiful marketing trick

[Fetzie]

So that's about 5% higher than a fury x, for about the same price as a fury x?

 

Nvidia really pushed the boat out this gen if this is a representative benchmark /s.

[zMeul]

33 minutes ago, Fetzie said:

So that's about 5% higher than a fury x, for about the same price as a fury x

it's more like ~10%, and doesn't need liquid cooling to function

[zMeul]

36 minutes ago, FakezZ said:

I didn't see them specify it was for VR only... Where did you get that from?

[Fetzie]

3 minutes ago, zMeul said:

and doesn't need liquid cooling to function

Eh, if they had needed an aio to cool a 16nm card I'd be worried.

 

Still, how come we haven't got a post slating Nvidia for their focus on VR from you? That's pretty much all they talked about on Friday.

[zMeul]

4 minutes ago, Fetzie said:

Eh, if they had needed an aio to cool a 16nm card I'd be worried.

it still has a substantial power draw, ~180W if I recall?!
Fury X on average uses ~220W on average: http://www.tomshardware.com/reviews/amd-radeon-r9-fury-x,4196-7.html

[Link4]

Those scores are from a regular Fury, Fury X FPS in Crazy preset with 4x AA is about 47.  http://www.hardocp.com/article/2016/04/01/ashes_singularity_day_1_benchmark_preview/3#.VzAt3FgrKUl

[Kinda Bottlenecked]

10 minutes ago, Link4 said:

Those scores are from a regular Fury, Fury X FPS in Crazy preset with 4x AA is about 47.  http://www.hardocp.com/article/2016/04/01/ashes_singularity_day_1_benchmark_preview/3#.VzAt3FgrKUl

 

So the difference isn't that big huh... I hope that gtx 1080 wasn't on 2.1ghz

[MrDynamicMan]

If this is both legit and accurate... Well, shit. Looks like nvidia dun fucked, if their 1080 is almost on par with a FURY.

[unsubscribed]

A lot of worries based on presumably somewhat accurate leak. I'd say wait until reviews are in (only 8 days left), keep an eye on bundled games (fingers crossed we'll get something worth it) and get one.

 

Every time new GPU is in from either red or green team benchmarks flood in, everyone panics, some even jump the ship etc. In my opinion the following does not matter:

 overclocking potential (I am not going to push my card to breaking point to get from 60 to 65fps, I rather lower  settings in game)

 1-5% performance difference between AMD and NVidia alternatives

 one game with 10-15% gain comparing similarly priced red and green GPUs

What does matter to me is:

 performance sustainability and thermals (both related and both important)

 feature set (weather it is VR optimization, game recording, streaming or even simply user friendly drivers)

 aesthetics (I would happily sacrifice 10% performance for GPU that I want to get for the looks and acoustics)

 resale value

 

If you like the way GTX 1080 performs and looks, if you like the driver feature set and you don't mind the price - get it. Unless performance is just barely reaching GTX 1080Ti levels (highly doubt) you will have one beast of GPU in your rig and you will enjoy the hell out of it. I certainly will 

[Guuhan]

Still want the performance on their roadmap with the "rough estimates" ! :"( 

[zMeul]

1 hour ago, MrDynamicMan said:

If this is both legit and accurate... Well, shit. Looks like nvidia dun fucked, if their 1080 is almost on par with a FURY.

how'd you figure that out?!

the GTX1080 is the new "king of the hill" in single GPU performance and it (will) offer it at aprox the same cost as a Fury X and at a lower power consumption; and this is just one game

it offers above Titan X perf at ~1/2 the cost - how exactly did they fckd up?!

 

not to mention, theoretically there will be two more cards above the GTX1080, the Ti and the Titan replacement

 

---

 

I'd like to see Polaris 10 even getting the same level of perf as Fury X - it won't!

[Sintezza]

Since the GTX1080 isnt for sale in the wild yet.

I have my doubts about these numbers.

Nvidia might just start to send out review samples by now.