• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

3DMark brand new DX12 benchmark "Time Spy"

Associate
Joined
8 May 2014
Posts
2,288
Location
france
as titles say new DX12 Bench, i wonder how Nvidia and AMD will compare in this, here is the new trailer, and confirmation from futuremark about async.

WoPBjvk.jpg

To verify if Time Spy incorporates Async Compute technology we reached out to James Gallagher, a Marketing Manager at Futuremark:

Time Spy uses asynchronous compute to overlap rendering passes. GPUs supporting async compute can process work from multiple queues in parallel. There are three types of queue: graphics, compute, and copy. A graphics queue executes rendering commands and can also handle other work types. A compute queue only accepts compute work, and a copy queue only accepts copy operations. The queues all race for the same resources so the overall benefit depends very much on the workload. We’ll be talking more about the technical side of Time Spy in the coming weeks.

I also asked if 3DMark Time Spy is designed for high-end systems and if Futuremark is planning to include Time Spy into current 3DMark benchmark stack or replace any of the existing modules:

Time Spy will be added to 3DMark as a new test alongside the others. It won’t replace Fire Strike. Both are aimed at high-end gaming PCs but because DirectX 12 is a low-level, low-overhead API, Time Spy is able to render significantly more detail and complexity to the screen than Fire Strike.

So there you have it.

Futuremark confirmed that they will be sharing more details about TIme Spy after Summer Steam Sales. We should expect info on pricing, edition features and release date.

Here’s the press release that was released today:

Since 1998 gamers have turned to 3DMark benchmarks to see the latest advances in real-time graphics made possible with each new version of DirectX. That’s why today I am excited to share the first official trailer and screenshots from 3DMark Time Spy, our new DirectX 12 benchmark test.

Time Spy is a new DirectX 12 benchmark test, coming soon to all Windows editions of 3DMark. With its pure DirectX 12 engine, built from the ground up to support new features like asynchronous compute, explicit multi-adapter, and multi-threading, Time Spy is the ideal benchmark for testing the DirectX 12 performance of the latest graphics cards.

Developed with input from AMD, Intel, Microsoft, NVIDIA, and the other members of the FuturemarkBenchmark Development Program, Time Spy shows the exciting potential of low-level, low-overhead APIs like DirectX 12.

In the trailer, you’ll get a glimpse of gaming’s future while revisiting classic 3DMark scenes from the past. For a bit of fun, see how many can you spot.


source: Videocardz
 
Last edited:
Soldato
Joined
17 Oct 2012
Posts
5,273
Location
Leeds
Let's hope this benchmark is actually hard to run. I don't think there is a single benchmark that stresses the newer cards at all even heaven 4 in 4k and firestrike ultra aren't really pushing cards.
 
Soldato
Joined
17 Oct 2012
Posts
5,273
Location
Leeds
I would hardly call the very low fps a driver problem when the score is at the top of the leaderboard.

It looks like it's getting over 20fps which is too much. A decent benchmark should be getting 20fps with 4 of the latest cards. Not 20fps with a single card which isn't even top of the line. When big Pascal comes out you are going to be getting nearly 60fps with 2 of them which is no good.
 
Soldato
Joined
6 Jan 2013
Posts
21,855
Location
Rollergirl
It looks like it's getting over 20fps which is too much. A decent benchmark should be getting 20fps with 4 of the latest cards. Not 20fps with a single card which isn't even top of the line. When big Pascal comes out you are going to be getting nearly 60fps with 2 of them which is no good.

What logic is this? The goal is to push as many frames as possible, not the least.
 
Soldato
Joined
17 Oct 2012
Posts
5,273
Location
Leeds
What's the point in a benchmark where you get 60fp that's not pushing cards and it's not a game so it doesn't matter if it's smooth. On ln2 with lod you can't really tell much difference between loading screens and the benchmark anyway.
 
Soldato
Joined
28 May 2007
Posts
10,071
In short he wants the Crysis of Benchmarks. We want a benchmark that will last and stay relevant until the next one. Firestrike with extreme and ultra has done a decent job tbh.
 
Soldato
Joined
6 Jan 2013
Posts
21,855
Location
Rollergirl
No its to get less, the more frames you pushing the more cpu power you need at which point it becomes a cpu benchmark and might as well be 2d.

Nonsense. My setup can push Heaven over 200fps with the GPU's at 99% and the CPU at barely 20%.

Explain that one, Einstein.

It's a benchmark because a 980ti will get a better score than a 980 but a lesser score than a 1080. The CPU is practically irrelevant, as I'd probably get the same score with my 6600k than I do with my 5930k.
 
Man of Honour
Joined
21 May 2012
Posts
31,940
Location
Dalek flagship
It looks like it's getting over 20fps which is too much. A decent benchmark should be getting 20fps with 4 of the latest cards. Not 20fps with a single card which isn't even top of the line. When big Pascal comes out you are going to be getting nearly 60fps with 2 of them which is no good.

I have got 4 of the fastest cards around and it still looks bad @4k.

You do realise that you can run Firestrike @8k if you want to providing you have a stack of cards to drive it in quad SLI and it uses a lot of memory.

Perhaps you should try Firestrike at 8k and let us know how you get on lol.
 
Soldato
Joined
6 Jan 2013
Posts
21,855
Location
Rollergirl
Yes pretty much :)

You do realise people don't just use these facilities to benchmark Quad 1080s? If you got this card crushing benchmark then you'd find that 99% of the hardware couldn't run it, and of the 1% that could actually push those 4 frames... what would be the margin of division? Every card that could run it would get a score of "4". :rolleyes:
 
Soldato
Joined
17 Oct 2012
Posts
5,273
Location
Leeds
You do realise people don't just use these facilities to benchmark Quad 1080s? If you got this card crushing benchmark then you'd find that 99% of the hardware couldn't run it, and of the 1% that could actually push those 4 frames... what would be the margin of division? Every card that could run it would get a score of "4". :rolleyes:
Yes of course. I was benchmarking fs extreme with gtx670 just last night. My fastest single gpu is only a fury x.
 
Man of Honour
Joined
21 May 2012
Posts
31,940
Location
Dalek flagship
You do realise people don't just use these facilities to benchmark Quad 1080s? If you got this card crushing benchmark then you'd find that 99% of the hardware couldn't run it, and of the 1% that could actually push those 4 frames... what would be the margin of division? Every card that could run it would get a score of "4". :rolleyes:

The 1080s can not run Firestrike in 4 way SLI.:eek:

The fastest setup you can run is 4 way TitanXs and on Firestrike Ultra you don't even need to overclock the CPU as the fps are quite lol.
 
Back
Top Bottom