FutureMark Showcases 3DMark's DirectX 12 'Time Spy' Benchmark

[HKZeroFive]

From Guru3D:

 

With DirectX 12 available on Windows 10 it's just a matter of time before everybody jumps onto it. Not just the hardware and games, the test slash benchmark softwares will adapt as well. Futuremark has been hard at work to up-date their 3DMark benchmark software.

During a presentation Futuremark at GOC in Asia they have showed the first footage of the new version that is DX 12 enabled. The presentation is an early beta from the looks of it. Pasi Virtanen from Futuremark mentioned that the new version of 3DMark will get two GPU and one CPU test, one of the names would be Time Spy.

From WCCFtech:

 

While game developers ready their new titles, Futuremark’s Product Manager, Pasi Virtanen, announced at the GALAX GOC 2015 event that they are developing their latest 3DMark benchmark which will use the DirectX 12 API. The new 3DMark will focus around the tech and features which DirectX 12 is going to offer in future next-generation AAA titles. The benchmark will include several game tests, one of which known as “Time Spy” was showed at the event that revolves around a museum filled with highly detailed objects. The whole tests makes use of 3D rendering using DirectX 12 API to test performance of graphics cards. The benchmark test runs at 2K or 1440P resolution and is 5 times (5x) more heavier than 3DMark Fire Strike which is available in the 3DMark 2013 benchmark suite. That means that the graphics test will require some serious horsepower to render and only the best performance cards will be able to score decent numbers in the test.

 

From Tom's Hardware:

 

FutureMark says it has been following the news break of the benchmark and has noticed that there appears to be concerns of bias based on the Nvidia logo seen above the video presentation. The company was clear with us that the logos were there for the Galax event and not its own presentation, so the Nvidia part was purely coincidental.

 

FutureMark said that the Time Spy 3DMark benchmark is 'being developed with input and feedback from all our Benchmark Development Program partners' and added that it has been using this approach to develop all of its benchmark tools since 1997. FutureMark said said all BDP members have the opportunity to review the source code and make suggestions towards and improving the benchmark and that BDP partners that had a hand in developing Time Spy include AMD, Nvidia, Intel and Microsoft, among others.

 

Video:

Another DX12 benchmark means exciting news! I'm assuming this is basically 3DMark with DX12, and it looks promising. You gotta wonder though what involvement NVIDIA has in the benchmark program since they've been falling behind in DX12 benchmarks lately in Ashes of the Singularity and Fable Legends. Hope nothing uncompetitive or dodgy is going on because starting another controversy isn't favourable. Otherwise looks good!

Sauces:

http://www.guru3d.com/news-story/futuremark-shows-first-footage-3dmark-directx-12.html

http://wccftech.com/futuremark-3dmark-time-spy-directx-12-benchmark/

 

http://www.tomshardware.com/news/3dmark-time-spy-coming-2016,30758.html

[Ashaira]

seriously? you make a benchmark that is supposed to test graphics without preference  and you put the logo of a card manufacturer on it? rly?

[That Norwegian Guy]

You gotta wonder though what involvement NVIDIA has in the benchmark program since they've been falling behind in DX12 benchmarks lately such as in Ashes of the Singularity and Fable.

 

I also worry for Nvidia's upcoming architecture (Pascal) - due to how they responded to AotS vs. current cards, it seems they haven't even designed the architecture to get a decent async compute functionality in Pascal either. I think the first 1 or 2 generations of cards in the DX12 and Vulkan era is about to be extremely profitable for AMD.

[Swndlr]

seriously? you make a benchmark that is supposed to test graphics without preference  and you put the logo of a card manufacturer on it? rly?

You ever stop to think that maybe Nvidia sponsored the event and is giving the event holder graphics cards to run the benchmark?

[Mr_Troll]

dat NV logo on top.... rip amd performance.... Nvidia is always better in synthetic benchmarks... so not suprised

[AlwaysFSX]

Now hopefully games will finally start looking good with decent textures, AO, bloom, reflections, global illumination, etc because of this showing what DX12 can do on hardware in the very near future. Even if it's not at that bar, just under is still a large step up from where we're at right now.

 

You ever stop to think that maybe Nvidia sponsored the event and is giving the event holder graphics cards to run the benchmark?

Nah, remember you have to hate Nvidia right now, that's the meta after all.

[Swndlr]

Now hopefully games will finally start looking good with decent textures, AO, bloom, reflections, global illumination, etc because of this showing what DX12 can do on hardware in the very near future. Even if it's not at that bar, just under is still a large step up from where we're at right now.

 

Nah, remember you have to hate Nvidia right now, that's the meta after all.

Says who

[Ashaira]

You ever stop to think that maybe Nvidia sponsored the event and is giving the event holder graphics cards to run the benchmark?

you ever stop to think that is a conflict of interest?

[Zah]

I also worry for Nvidia's upcoming architecture (Pascal) - due to how they responded to AotS vs. current cards, it seems they haven't even designed the architecture to get a decent async compute functionality in Pascal either. I think the first 1 or 2 generations of cards in the DX12 and Vulkan era is about to be extremely profitable for AMD.

Yet nvidia's shown to have also performed better with newer drivers on AotS, and no one knows what conusmer pascal chips will have.

[AlwaysFSX]

Says who

Minimum half this forum. 

 

you ever stop to think that is a conflict of interest?

Every benchmark is biased in that case.

[Zah]

you ever stop to think that is a conflict of interest?

A conflict of interest would be nvidia giving money to make the benchmark.  We do not have legitimate evidence of that happening.

[Ashaira]

A conflict of interest would be nvidia giving money to make the benchmark.  We do not have legitimate evidence of that happening.

oh sure. i don't give money to armies to sustain the war. i just give them guns to test their soldiers performance on.

 

notice the similarities?

[Zah]

oh sure. i don't give money to armies to sustain the war. i just give them guns to test their soldiers performance on.

 

notice the similarities?

No I don't because that was one of the dumbest analogies I've heard.

[Swndlr]

you ever stop to think that is a conflict of interest?

How so, I see nothing wrong with them using Nvidia cards for a benchmark

[HKZeroFive]

Yet nvidia's shown to have also performed better with newer drivers on AotS, and no one knows what conusmer pascal chips will have.

Haven't seen any benchmarks in which NVIDIA beats AMD in AotS. Unless you're referring to this:

Even then, it's in 1080p, in which the 980Ti only beats the Fury X by 0.9 FPS so it's safe to say that the Fury X will beat the 980Ti at higher resolutions. All other NVIDIA cards are being beaten by their AMD counterparts. Feel free to provide another benchmark source, this is the only one I could find using the 'latest drivers'.

[That Norwegian Guy]

no one knows what conusmer pascal chips will have.

 

If Nvidia designed async compute into Pascal, they'd be better served to make their statement to AotS after the initial benchmark to be more positive, rather than getting contrarian, combative and huffy about it. 

 

They could have said: "We are aware of our current generation lacking async compute capabilities, but are on track to remedy that situation with our upcoming Pascal architecture"

 

Their actual statement: (paraphrased) "ASYNC COMPUTE IS NOT IMPORTANT AND YOU CHEATED! YOU'RE IN BED WITH AMD! WE'RE TAKING THE BALL AND WE'RE GOING HOME!"

[Stadin6]

why not Vulkan? 

[Bouzoo]

While I'm interested in this I wonder, what are the chances of 3D Mark for Vulkan? 

[Bouzoo]

why not Vulkan? 

 

Afaik most of the devs are focusing on DX12. Can't remember a game based on Vulkan was announced. Many people still have no idea what Vulkan is while DX12 is the one that's been talked about. 

[Zah]

why not Vulkan? 

Not out yet.

 

 

Haven't seen any benchmarks in which NVIDIA beats AMD in AotS. Unless you're referring to this:

-snip-

Even then, it's in 1080p, in which the 980Ti only beats the Fury X by 0.9 FPS so it's safe to say that the Fury X will beat the 980Ti at higher resolutions. All other NVIDIA cards are being beaten by their AMD counterparts. Feel free to provide another benchmark source, this is the only one I could find using the 'latest drivers'.

 

I was in fact referring to that, yes, although I was unaware of how close it was.  But my original point stands in that drivers still matter, and the game is not yet out, and this benchmark is not yet out, and people are already upset about things.

[Alurit]

why not Vulkan? 

it's much earlier in development than dx12, it's promising but it will take even more time for it to be out

[Exotath]

They've had vendor brand logo's in their benchmarks for quite some time now. MSI, Gigabyte, and NV have been corporate partners of theirs for a few years now.

[-BirdiE-]

Haven't seen any benchmarks in which NVIDIA beats AMD in AotS. Unless you're referring to this:

Even then, it's in 1080p, in which the 980Ti only beats the Fury X by 0.9 FPS so it's safe to say that the Fury X will beat the 980Ti at higher resolutions. All other NVIDIA cards are being beaten by their AMD counterparts. Feel free to provide another benchmark source, this is the only one I could find using the 'latest drivers'.

 

Umm.. It's really not "safe to say" if the benchmarks haven't been run. The FuryX, while it has faster ram, has not shown to be any better at 4K than the 980ti.

 

 

[Doobeedoo]

Looks really solid. Since it's heavy, as it says, it will be funny to see next top cards scores.

[spartaman64]

oh sure. i don't give money to armies to sustain the war. i just give them guns to test their soldiers performance on.

 

notice the similarities?

and dont you want the army to help you not the other side so your analogy proves the opposite point