videocardzFirst NVIDIA GeForce GTX 1080 3DMark Benchmarks

[csbin]

http://videocardz.com/59558/nvidia-geforce-gtx-1080-3dmark-benchmarks

The second benchmark we have is from FireStrike Extreme, which as you probably already know is rendered at 2560×1440 resolution. This is actually the first benchmark where you will see 1.860 GHz clock. In this scenario GTX 1080 is faster than typical overclocked GTX 980 Ti (~8700 points). However if we compare it GTX 980 Ti running at almost the same frequency (1.8GHz with LN2 cooling), GTX 1080 is actually much slower clock to clock (link).

 

[Head1985]

1080 is 26% faster than reference GTX980TI.Stock reference 980TI have 8000points in firestrike extreme GPU score.
http://www.guru3d.com/articles_pages/nvidia_geforce_gtx_980_ti_review,24.html

And 980TI at 1500/8100 have 9900points so its even slightly faster than 980TI 1500/8000.

 

[airfathaaaaa]

what "no FM approved" means?

 

[MangoX]

Means Futuremark hasn't tested the driver yet.

 

[Silverforce11]

Looks about spot on, 25-30% above 980Ti.

1.86Ghz boost is crazy high though, well done to NV & TSMC.

Have to see whether it can really OC much higher than that, since a 980Ti at 1.5ghz pretty much close that gap.

 

[Glo.]

More importantly, what will be the TDP of that GPU with that core clocks.

 

[xpea]

These preliminary 1080 results are right where most of us have been expected (if not fake of course), event a bit higher than the 10~20% above 980Ti consensus.
What looks a bit over the top is the incredible 1.8GHz boost clock of this reference card
and how much headroom is left for AIB overclocked versions ?

 

[railven]

WHen is the bigger pascal expected? Early 2017 or mid 2017? (ANd also, the AMD competitors?)

As it is now, I got a blank check ready to go to someone. ~30% over my 980 Ti stock is enticing, hopefully OCs better and NV fixes their DX12 issues.

 

[thilanliyan]

Don't 980Ti cards already hit 1500mhz easily? Or is that what golden cards can do?

 

[railven]

Don't 980Ti cards already hit 1500mhz easily? Or is that what golden cards can do?

I've read here most cards can hit 1400mhz+, mine was definitely a gold card because it could hit 1504mhz with stock cooler @ 100%.

Using a EVGA Hybrid kit, my card stays at 1514mhz boost clocks while 42C. Stock BIOS is keeping me from going higher.

 

[xpea]

Don't 980Ti cards already hit 1500mhz easily? Or is that what golden cards can do?

you must compare apples and apples. reference card vs reference card. custom vs custom. oc vs oc. Don't mix up everything please ()
in other world, don't compare "not overclocked reference" vs "custom factory overclocked" boards.

For now we only have the result of a not overclocked reference. Thus:
Reference GTX980Ti : 1000 MHz base / 1075 MHz boost
Reference GTX980: 1126 MHz base / 1216MHz boost
Reference GTX1080: around 1600MHz base / 1860MHz boost

Big big boost in clocks !

 

[FatherMurphy]

Do we think the 1080 is going to have a similar TPD as the 980 (165W) at those clocks? That's an impressive jump.

 

[airfathaaaaa]

you must compare apples and apples. reference card vs reference card. custom vs custom. oc vs oc. Don't mix up everything please ()
in other world, don't compare "not overclocked reference" vs "custom factory overclocked" boards.

For now we only have the result of a not overclocked reference. Thus:
Reference GTX980Ti : 1000 MHz base / 1075 MHz boost
Reference GTX1080: around 1600MHz base / 1860MHz boost

Big big boost in clocks !

and how do you know that this is not a "burn" test to see how the card performs ?

 

[Glo.]

Do we think the 1080 is going to have a similar TPD as the 980 (165W) at those clocks? That's an impressive jump.

I already said, I think the TDP will be around 200W.

I know it is hype and everything, but:
http://www.3dmark.com/3dm11/11061015
GTX 1070 might have 7.5 GB of RAM. It might be the same history as is with GTX 970, and GPC arrangement.

 

[digitaldurandal]

If this is accurate I won't have any trouble waiting for big die GPU releases after all.

 

[Face2Face]

So I guess if this isn't fake, then it's safe to assume the GTX 1070 is about as fast as a GTX 980Ti, while using much less power.

 

[FatherMurphy]

I already said, I think the TDP will be around 200W.

I know it is hype and everything, but:
http://www.3dmark.com/3dm11/11061015
GTX 1070 might have 7.5 GB of RAM. It might be the same history as is with GTX 970, and GPC arrangement.

I know TDP is not necessarily power draw, but if 1080 has a TDP of 200W and only barely outperforms an overclocked 980ti, that would be a surprisingly low jump in perf/w, no?

 

[el etro]

For start is the all-time biggest node jump for GPUs. Is not crazyness to expect this huge performance jump, especially on AMD case(AMD will bring their real new architecture in FOUR years). I expect all this performance jump is done at GTX 980 TDP levels.


Now the guys understand why they tell that Polaris 10 can achieve 980Ti performance at 232mm^2 and low TDP levels.

 

[FatherMurphy]

For start is the all-time biggest node jump for GPUs. Is not crazyness to expect this huge performance jump, especially on AMD case(AMD will bring their real new architecture in FOUR years). I expect all this performance jump is done at GTX 980 TDP levels.


Now the guys understand why they tell that Polaris 10 can achieve 980Ti performance at 232mm^2 and low TDP levels.

The big question (I think) from AMD's perspective is whether their architecture can clock substantially higher than GCN chips have for the past 4 years. Clearly, Nvidia has had the clock speed lead for some time now and it appears (based not only on this leak but on the impressive clock speeds of GP100) that Nvidia is able to take advantage of 16nm's clock speed boost.

I wonder if the 1800 Mhz in this leak is overclocked or regular boost.

 

[MangoX]

That is a low perf/w jump, considering AMD is claiming up to a 2.5 perf/w performance increase. Nvidia making their architecture more GCN-like has increased power usage.

 

[railven]

Considering what happened with 28nm, I sort of expect both AMD and NV to milk this node as long as they can.

I've pretty much given up on historical trends, welcome to the 20-30% increase per generation era. They's got to make their money, after all.

 

[airfathaaaaa]

Considering what happened with 28nm, I sort of expect both AMD and NV to milk this node as long as they can.

I've pretty much given up on historical trends, welcome to the 20-30% increase per generation era. They's got to make their money, after all.

wait you actually expected MORE % from the first gen of a new node? D:
this never happened
even if you follow the mobile soc that snapdragon just die shrink the maximun they saw was 36% node to node

 

[railven]

wait you actually expected MORE % from the first gen of a new node? D:
this never happened
even if you follow the mobile soc that snapdragon just die shrink the maximun they saw was 36% node to node

No, I always predict lower so I'm not left disappointed. Being a prior AMD fan conditioned me to this.

 

[Det0x]

You know its bad when csbin start posting nvidia news

 

[DarkKnightDude]

Isn't that worst, clock for clock? Or am I looking at this wrong?