Futuremark 'Time Spy' DX12 Benchmark Coming Early 2016

News
By Kevin Carbotte
published

Futuremark, makers of well-known 3D benchmark tool 3DMark, has a new benchmark tool coming in the early part of next year. The company revealed the first information about the new tool at a recent overclocking competition in China hosted by Galax.  

The GALAX GOC overclocking competition in Wuhan, China took place on December 11 and 12, and 14 professional overclockers competed for a chance to win a grand prize of 30,000 chinese yen (approximately $4,650 USD) by pushing Galax hardware to the limits. As part of the presentation for the event, Galax offered Futuremark some stage time to reveal the first glimpse of its upcoming Direct X 12 benchmark tool. A video of the presentation surfaced this weekend on Chinese video service YouKu and has since been reposted to YouTube by Guru3d.

"Time Spy" Benchmark Coming

We reached out for comment from Futuremark about the new benchmark and we were told that after a short speech, Pasi Virtanen, Futuremark’s 3DMark product manager, presented a short trailer of the upcoming Time Spy benchmark tool, which will be added to the 3DMark benchmark in early part of 2016.

Time Spy will consist of three tests -- two graphics tests and a single physics test -- which will test GPU and CPU performance. The test will also include a demo reel and soundtrack, as we've seen with previous 3DMark benchmark tests.

Futuremark said the Time Spy benchmark is being developed using a Direct X 12 engine that it created in-house to showcase the API's new features. The company pointed out that one of DX12’s greatest benefits is its ability to have far more draw calls than the previous version of the API. We were told that “Time Spy has roughly five times as much content on screen as Fire Strike.”

On The Subject Of Bias

Futuremark said it has been following the news break about 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 of 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 all BDP members have the opportunity to review source code and make suggestions towards improving the benchmark and that BDP partners that had a hand in developing Time Spy include AMD, Nvidia, Intel and Microsoft, among others.

Futuremark wasn’t willing to be more specific than to say that Time Spy would be made available in early 2016 and that more information would be revealed in the new year.

______________________________________________________________________

Kevin Carbotte joined Tom’s Hardware in early 2015. He writes GPU and VR hardware reviews and contributes to the news channel in the areas of computer graphics, water cooling, VR and other immersive technology. Kevin’s personal interests include technology advancements, fast cars and collecting video games that he doesn't have time to play.

Follow Kevin Carbotte @pumcypuhoy. Follow us on Facebook, Google+, RSS, Twitter and YouTube.

Kevin Carbotte

 Kevin Carbotte is a contributing writer for Tom's Hardware who primarily covers VR and AR hardware. He has been writing for us for more than four years. 

See more GPUs News
13 Comments Comment from the forums
  • dragonsqrrl
    "Futuremark’s 3DMark product manager, presented a short trailer of the upcoming Time Spy benchmark tool, which will be added to the 3DMark benchmark in early part of 2016."

    So will this be an update to the existing version of 3DMark? That would be pretty sick.
    Reply
  • fuzzion
    Why are all the top performing futuremark benchmark cards from nvidia?

    Reply
  • rmpumper
    17129350 said:
    Why are all the top performing futuremark benchmark cards from nvidia?

    Because all top performing GPUs are from nvidia.
    Reply
  • Piki__
    I remember when benchamrsk were pretty accurate. 2 or 3 years ago. I remember how some AMD vs NVIDIA cards were standing. Some times ago I have noticed a big jump in NVIDIA scores that did not reflect that well with benchmarks in games. I have a suspicion that this scores have been rigged. Mybe NVIDIA logo is there for a reason.
    Reply
  • jasonelmore
    I remember when benchamrsk were pretty accurate. 2 or 3 years ago. I remember how some AMD vs NVIDIA cards were standing. Some times ago I have noticed a big jump in NVIDIA scores that did not reflect that well with benchmarks in games. I have a suspicion that this scores have been rigged. Mybe NVIDIA logo is there for a reason.

    Old Futuremark benchmarks increased the score if the card could do Physx, Since AMD cards did not do Physx back in those days, they got lower scores.

    Later on, Physx was taken out the scoring completely after much uproar from the overclocking community
    Reply
  • ohim
    People make the mistake of still making Futuremark or w/e Mark a benchmark guideline.

    If real life gaming scenarios taught us something is that no game will perform the same with each GPU. I don`t even get it why would you buy a GPU based on Futuremark scores ... are you going to play Futuremark with the GPU or actual games? Also benchmarks scores can be driver optimized to get better scores while gaming is the same ...This is proven by lots of gaming benchmarks on YT.

    Just think about it, do you care about 3D Mark scores of over 9000!!! or do you care about 10-20- 50 FPS more in the game that you actually play?
    Reply
  • jaber2
    People make the mistake of still making Futuremark or w/e Mark a benchmark guideline.

    If real life gaming scenarios taught us something is that no game will perform the same with each GPU. I don`t even get it why would you buy a GPU based on Futuremark scores ... are you going to play Futuremark with the GPU or actual games? Also benchmarks scores can be driver optimized to get better scores while gaming is the same ...This is proven by lots of gaming benchmarks on YT.

    Just think about it, do you care about 3D Mark scores of over 9000!!! or do you care about 10-20- 50 FPS more in the game that you actually play?

    I would use it when I upgrade my video card, test it before and after, just so I can say to myself I spent all that money and got X amount of increased FPS
    Reply
  • ComputerSecurityGuy
    We will see... I still wish more DX12 games were out. We could see if AMD has a solid lead in all DX12 gaming or just the ones currently out.
    Reply
  • Mousemonkey
    17129812 said:
    I remember when benchamrsk were pretty accurate. 2 or 3 years ago. I remember how some AMD vs NVIDIA cards were standing. Some times ago I have noticed a big jump in NVIDIA scores that did not reflect that well with benchmarks in games. I have a suspicion that this scores have been rigged. Mybe NVIDIA logo is there for a reason.

    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.

    You should actually try READING the articles in future.
    Reply
  • clonazepam
    For me, Futuremark's benchmarks are in the "hey that looks neat" category. When I was into it, I was far more impressed by doing huge gpgpu calculations with cuda or opencl. Make sure that puppy can actually last to the end with accuracy.

    I don't know what is a good, readily available benchmark anymore that comes close to representing actual gameplay. Developers these days always seem to go over their performance budget. I imagine them huddled around the screen, egging each other on to squeeze in one more asset or effect to make it look perfect or something.

    Anyway, my preferred benchmark analysis stresses the importance of the minimum fps score. Time and time again, I see the particular hardware combinations leading here while maybe losing the avg and maximum fps. That's fine by me.

    Futuremark needs to get more sponsors from all the major players in the field, and get a full-featured benchmark out there at zero cost. The exception would be business licenses only. On principle, I just refuse to pay for this due to its inaccuracies and being more of a spectacle than anything truly important.
    Reply