Possible GTX 880 benched in 3DMark Firestrike Extreme!

[Cloudfire777]

This was posted at the forum at Notebookreview.com today under the Maxwell thread
http://forum.notebookreview.com/gam...ew-maxwell-cards-incoming-19.html#post9726698

The GPU name was censored but one can see "Nvidia Geforce GTX 880" if you look closer.

The GTX 880 scored 6110 in total in 3DMark Extreme and it scored 6578 in GPU score.

In comparison, GTX 780 Ti score 5153 (+28%) in GPU score while R9 290X 5063 (+30%) in GPU score on the same bench.

I tried checking all entries that scored 6110 on 3DMark.com but nobody matches the GPU score.

The leaker mention "Gamescom" which happen between August 13th -17th, and there was a previous rumor that 2nd generation Maxwell will finally be presented around that date.

Fingers crossed the bench isnt a fake one. The card should be a beast if legit

 

[Grooveriding]

Better hope it's fake because the lowest ranked 780ti with 4770K on 3DMark is faster than that. Not sure where you found those abysmal 3D Mark benches...

http://www.3dmark.com/fs/1884843

I could post some of my FirestrikeX runs, but that wouldn't be fair with the overclock I have on them, I get over 7000 GPU score on a single 780ti and my overall score is much higher, but that is due to a 6 core chip rather than 4 as seen in this bench.

I'm thinking has to be fake because the performance is terrible for a new part.

 

[Cloudfire777]

Dude, stop with the overclocked benches. 1195MHz on the core. 1995MHz on the VRAM. Please dont ruin this thread with posts like that.
The results I posted are mirrored by several other respectable reviewer sites.

Legitreviews:
5173 in GPU score

Bj0rn3d.net
Around 5000 if you add the extra specs for 780 Ti

Hardwarezone.com
Around 5000

Overclockersclub.com
Around 5000

The GTX 880 seems to score around 30% more than a GTX 780 Ti stock

 

[Grooveriding]

Yeah the more I look at these they must be fake. Your average 780ti boosts to about 1100-1150 on the core out of the box.

Certainly a fake, there is no way the GTX 880 will perform this poorly. They're not going to release a new flagship that needs to be overclocked to be faster than the previous one.

 

[toyota]

"Not sure where you found those abysmal 3D Mark benches"


Grooveriding, come on now. that list of scores he listed is accurate and line right up with my 770 and 780 scores at various speeds.

 

[Cloudfire777]

Try reading my post again instead of trolling Grooveriding

 

[Grooveriding]

Grooveriding, come on now. those score he listed are right and line right up with my 770 and 780 scores at various speeds.

Sure, but a 780ti is about 20% faster than a 780. Even a conservatively overclocked 780ti would be faster than this card. I'm not expecting much from the 880, but certainly not performance this poor.

 

[Cloudfire777]

I speculated on +30-35% over GTX 780 Ti. Looks like I might be right. Wonder what the TDP will be

 

[Alatar]

Better hope it's fake because the lowest ranked 780ti with 4770K on 3DMark is faster than that. Not sure where you found those abysmal 3D Mark benches...

That 780Ti is 1200MHz + boost meaning that it's running at around ~1350MHz.

You need a 1400MHz+ 780Ti to break 7K GPU score. And even then it usually requires LOD tweaks etc. Can also be done on Titans (I know because I've done around 7.1K GPU score on a Titan, but it requires close to 1500MHz)

On topic : the screen seems a bit too convenient.

 

[Cloudfire777]

That 780Ti is 1200MHz + boost meaning that it's running at around ~1350MHz.

You need a 1400MHz+ 780Ti to break 7K GPU score. And even then it usually requires LOD tweaks etc. Can also be done on Titans (I know because I've done around 7.1K GPU score on a Titan, but it requires close to 1500MHz)

On topic : the screen seems a bit too convenient.

Here are all the entries on 3DMark.com running Firestrike Extreme with i7 4770k and scoring 6110. None match the GPU score. I know some doesnt upload the results, but promising that no other previous results there match

http://www.3dmark.com/search?_ga=1.66069161.626027082.1405978530#/?mode=advanced&url=/proxycon/ajax/search/cpu/fs/X/1605/6110?minScore=6110&cpuName=Intel%20Core%20i7-4770K

 

[ams23]

Dude, stop with the overclocked benches. 1195MHz on the core. 1995MHz on the VRAM. Please dont ruin this thread with posts like that.
The results I posted are mirrored by several other respectable reviewer sites.

Legitreviews:
5173 in GPU score

Bj0rn3d.net
Around 5000 if you add the extra specs for 780 Ti

Hardwarezone.com
Around 5000

Overclockersclub.com
Around 5000

The GTX 880 seems to score around 30% more than a GTX 780 Ti stock

You are correct, and there are several other websites that show a reference GTX 780 Ti with a Fire Strike Extreme graphics score of just slightly above 5000:

PCPer: http://www.pcper.com/reviews/Graphi...-Card-Review/3DMark-Fire-Strike-and-Fire-Stri

Tweaktown: http://www.tweaktown.com/reviews/58...i-3gb-reference-video-card-review/index5.html

etc.

 

[Grooveriding]

That 780Ti is 1200MHz + boost meaning that it's running at around ~1350MHz.

You need a 1400MHz+ 780Ti to break 7K GPU score. And even then it usually requires LOD tweaks etc. Can also be done on Titans (I know because I've done around 7.1K GPU score on a Titan, but it requires close to 1500MHz)

On topic : the screen seems a bit too convenient.

I break 7000 on a 1300Mhz 780ti with a 5ghz 3930K, no tweaks beyond having texture filtering set to best performance, but none of the lucid virtu cheats or any of the outright cheats that you can use. Granted it is a 3D mark ranked score I offered, but it was the lowest one I could find using the same CPU.

Anyway I agree, everything about this image stinks to me. The performance is too low and it's likely a photoshop troll. Releasing a new flagship that amounts to the performance of an overclocked 780ti is just not reasonable.

 

[Cloudfire777]

Uhm, I found this entry for Nvidia at Gamescom which the leaker mentioned. Notice the picture.
http://www.nvidia.de/object/gamescom2014-de.html

Then look at the teaser from Nvidia Germany and UK

I break 7000 on a 1300Mhz 780ti with a 5ghz 3930K, no tweaks beyond having texture filtering set to best performance, but none of the lucid virtu cheats or any of the outright cheats that you can use. Granted it is a 3D mark ranked score I offered, but it was the lowest one I could find using the same CPU.

Anyway I agree, everything about this image stinks to me. The performance is too low and it's likely a photoshop troll. Releasing a new flagship that amounts to the performance of an overclocked 780ti is just not reasonable.

Yes buhuuu, a much lower TDP GPU beating a 250W monster GPU from Kepler by 30%.
Seriously, what else would you expect?

 

[SinOfLiberty]

On topic : the screen seems a bit too convenient.

Too convenient?

 

[ams23]

Yes buhuuu, a much lower TDP GPU beating a 250W monster GPU from Kepler by 30%.
Seriously, what else would you expect?

This is actually a reasonably good result for GTX 880 (if true), especially if power consumption is lower in comparison. GTX 680 was "only" ahead of GTX 580 by ~ 20-40% with most games (and obviously with lower power consumption too).

 

[Cloudfire777]

This is actually a reasonably good result for GTX 880 (if true), especially if power consumption is lower in comparison. GTX 680 was "only" ahead of GTX 580 by ~ 20-40% with most games (and obviously with lower power consumption too).

Totally agree, just in line with how GTX 680 vs 580 (295mm2 - 195W vs 520mm2 - 250W) like you say. Very curious about the TDP here since they are still on 28nm

Another good thing is that this baby will be going inside notebooks, which will make it superfast compared to GTX 880M. Obviously not as fast as GTX 880, but maybe GTX 780 Ti.

 

[blackened23]

Really? Comparing stock scores to overclocked scores, ESPECIALLY 3DMARK, is pretty silly. 3dmark is and also is heavily influenced by overall CPU speed, memory and GPU speed, and as you can guess, the leaderboards for all of the 3dmark scores are heavily overclocking the CPU, RAM, and GPU. Yet, when you compare stock CPU with a stock GPU setting the scores will (OBVIOUSLY) be much lower. And that is how websites do their testing. Everything stock.

Conversely, when you go to the 3dmark leaderboards with overclocked CPUs, RAM, and GPUs, that's where everyone disables LOD and uses high performance mode in NV control panel, disables tesselation in the AMD CP, and all of the other cheats that leaderboard level scorers use. IF you want to be silly, you can easily find tons of prior generation GPUs that score higher than next gen stock ones. Does that mean anything? No, that simply means the leaderboards are full of people running their systems at non sustainable overclocks/control panel tweaks/cheats just to get on 3dmark leaderboard. That said, I really don't know if that score is accurate, pretty sure no one does. It's safe to say that it will be a good deal faster than the 780ti, NV will not release a new flagship that isn't quite a bit faster than the prior one.

 

[Attic]

Price it at $500, win.

Scores look pretty good, i'm sure the main thing for it vs 780ti will be 6gb VRAM standard. 30% bump is decent, but OC'ing potential will matter as well. The 780/780ti really shine when unleashed.

 

[3DVagabond]

Price it at $500, win.

Scores look pretty good, i'm sure the main thing for it vs 780ti will be 6gb VRAM standard. 30% bump is decent, but OC'ing potential will matter as well. The 780/780ti really shine when unleashed.

780 ti +30% for $500? Considering nVidia's typical pricing, that will be a real breath of fresh air. I'll be really shocked though as it will be a major reduction in nVidia's ASP. Add to that there will be a lot of GK110 cards that will turn into unsalable stock if they do that.

 

[Attic]

Yea, unfortunately I think 6gb and $799 if this performance holds true.

 

[ocre]

I just dont know.

How much faster was the stock 290x over the 7970ghz?

Is it reasonable to think that Nvidia can get +30% over a 780ti while using less power?

This is actually a reasonably good result for GTX 880 (if true), especially if power consumption is lower in comparison. GTX 680 was "only" ahead of GTX 580 by ~ 20-40% with most games (and obviously with lower power consumption too).

That was with a node shrink. 28nm shrink played a huge role. It was significant yet people seem to forget this lately.

I just dont know. I know that >30% over a 780ti on the same node could happen for sure. But all while reducing power at the same time? I just dont know about that.

You have to think about it. While the gk104 accidentally became the gtx 680. It was origanlly aimed at being the 770ti but they found the chip to have turned out much more capable. It came back from the fab better than expected. It was originally designed for much more conservative clocks but they quickly found it was much more capable. After AMD launched the HD7970, nvidia pushed the gk104 up to notch out a win over it. But it was purely luck. Nvidia engineers were blessed with the gk104 on the 28nm node.

All of this is important because with the gm204 Nvidia has had much time to prepare. They know the node very very well and not only that, they have a really good indication of how maxwell will perform on 28nm because of the 750ti. Where the 750ti had a performance gap already slotted for it fill, the next maxwell doesnt.

Nvidia has AMD to worry about. With the 290x, AMD has shown that they are ready to take the fight all the way up to the boundaries of reasonable TDP. So for these reason, I am not sure Nvidia will be so conservative on their power consumption target. They know the node and how maxwell performs on the node. I am pretty sure they can project with great accuracy and target the TDP exactly where they want. I just dont see them leaving a lot left unless they really arent concerned at all what AMD will do.

So if the gm204 is on 28nm and it is a large chip, do we really expect it to targeted at 200watts? And is it reasonible to think a chip on the same node could be +30% faster while using a lot less power? This is a pretty big chip.

I believe Nvidia will aim at the higher end of the TDP scale to try to be certain they retain the GPU crown. I dont think they will leave much......
That it might even take 240watts to be 30% faster than a 780ti.

Think its gonna be that much faster while using a lot less power sounds like a lot to expect from a chip on the same node.

 

[toyota]

oops looking at wrong thing

 

[alcoholbob]

28% faster means...2x GTX 880s can hit 39fps in Crysis 3 at 4K

(That said, if the Titan II really is 600mm², a pair of them might hit 60fps at 4K).

 

[Mondozei]

Framerate differences tend to be lower in actual games than in synthetic benchmarks.

If we get 20% then that is good by most comparisons, but not great considering what we've accustomed ourselves to in the GPU space.

When AMD/NVIDIA both go out with 14-16 finfet nm with stacked memory then I'll upgrade from my 290 Tri-X(to whichever company has the best card under 500 dollars). So no more than 2 years.

If Torvalds is right, I might not buy all too many dGPUs after that, especially since we're seeing 2-3x improvements in some iGPUs over each generation.

 

[boxleitnerb]

My bad, GPU score...