• 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.

Ashes of the Singularity Coming, with DX12 Benchmark in thread.

Mauller said:
How about the 770ti getting a 180% performance boost? is that a negative?
Click to expand...

Why would that be a negative? And I thought Nvidia abandon their previous generation stuff. Or, at least the AMD fans tell me.
But surely it's a negative thing that their latest stuff is getting a performance hit?
Surely you can't believe that it's accurate?

Also, I just looked at OP's post, no 780Ti there.
 
Last edited:
Martini1991 said:
Why would that be a negative? And I thought Nvidia abandon their previous generation stuff. Or, at least the AMD fans tell me.
But surely it's a negative thing that their latest stuff is getting a performance hit?
Surely you can't believe that it's accurate?
Click to expand...

the 770 result is on wccftech here, And the funny thing is, by the look of the test, if they used the 5960x instead of the 3570k it could have had an even bigger boost in performance.


I still think it is the way that DX11 will sometimes submit older frames buffing the fps slightly at the expense of latency.

where as DX12 will most likely be rendering all new and up to date frames.

It might not be the case, it could even be that the shaders in the nvidia pathway need some tweaking still. but the driver overhead impact is only shown when you look at the individual tests and see that DX12 has an advantage over DX11 when the High draw call test is being used.

Things are not as clear cut with the averaged results most sites are showing. Although in general AMD show large improvements, nvidia only sees them on their current high end cards when the High draw call test is performed.
 
Last edited:
Can you hear that??? if you listen carefully, its the scream of a million tormented souls, crying out in anguish that AMD are actually better!!!!!!

Seriously though, its an alpha demo right? to be fair i wont put stock in any dx12 "benchmark" until its part of an actual released game with proper drivers for the cards, then we can talk, until then its all just speculation
 
Mauller said:
I still think it is the way that DX11 will sometimes submit older frames buffing the fps slightly at the expense of latency.

where as DX12 will most likely be rendering all new and up to date frames.

It might not be the case, it could even be that the shaders in the nvidia pathway need some tweaking still. but the driver overhead impact is only shown when you look at the individual tests and see that DX12 has an advantage over DX11 when the High draw call test is being used.

Things are not as clear cut with the averaged results most sites are showing. Although in general AMD show large improvements, nvidia only sees them on their current high end cards when the High draw call test is performed.
Click to expand...


Like I say, nothing can be taken away from AMD, but I'll wait till we see games start to come DX12 as mainstream.

RTS has never really been a good genre for GPU benchmarking. CPU's are great though.
 
Martini1991 said:
Like I say, nothing can be taken away from AMD, but I'll wait till we see games start to come DX12 as mainstream.

RTS has never really been a good genre for GPU benchmarking. CPU's are great though.
Click to expand...

Oh yeah, in general if you are looking for pure GPU performance then no game is a good benchmark. Although this benchmark does show a good comparison between DX11 and DX12. On the tests results it does show the TX and 980TI at 100% load and the benchmark being GPU bound when paired with a 5960x. IF you look at the ones that show the in game bench result and not just accumulated results.

Here is the 770 result. Link

Oh wait, i just noticed something, you can ignore the 770 test since WCCFtech ****ed it up. He chose the wrong Benchmark and tested theoretical CPU performance. It says at the top of the benchmark window in the picture 'CPU' instead of 'Full system test'
 
Last edited:
Mauller said:
How about the 770ti getting a 180% performance boost? is that a negative?
Click to expand...

That just proves that there is some serious flaws in that benchmark. A low end GPS should see the least gains in DX12, a high end GPU should see the most. The 770ti results show that Nvidia have excellent DX12 drivers and can expect big improvements, there should be even bigger improvements on faster hardware as the API becomes less of a limiting factor. The fact that the 980TI gets slower DX12 scores than DX11 is proof positive that the benchmark is absolutely flawed. This may be a Nvidia driver issue, but Nvidia have already indicated it is a game engine issue. Nvidia have a lot of experience with DX12 on different game engines and internal benchmarks, they know exactly what kind of performance benefits to expect, they have warned the ;public that this benchmark is not representitive, and that is blatantly obvious sicne the results are all over the place.
 
Last edited:
D.P. said:
That just proves that there is some serious flaws in that benchmark. A low end GPS should see the least gains in DX12, a high end GPU should see the most. The 770ti results show that Nvidia have excellent DX12 drivers and can expect big improvements, there should be even bigger improvements on faster hardware as the API becomes less of a limiting factor. The fact that the 980TI gets slower DX12 scores than DX11 is proof positive that the benchmark is absolutely flawed. This may be a Nvidia driver issue, but Nvidia have already indicated it is a game engine issue. Nvidia have a lot of experience with DX12 on different game engines and internal benchmarks, they know exactly what kind of performance benefits to expect, they have warned the ;public that this benchmark is not representitive, and that is blatantly obvious sicne the results are all over the place.
Click to expand...

I actually discovered the problem, WCCFTech chose the wrong benchmark option in that test. They chose the theoretical GPU performance option which just tests how many frames the cpu can pump out. instead of the 'full system test' option.

D.P. said:
Because Nvidia can enjoy the exact same API benefits unless you want to believe some kind of grand conspiracy theory.
Click to expand...

But nvidia do see gains in the High draw call versions of the test. you just need to look at the benchmark screenshots. it could just be that the normal and medium version of the test don't push the draw call limit that hard.
 
Last edited:
D.P. said:
Because Nvidia can enjoy the exact same API benefits unless you want to believe some kind of grand conspiracy theory.
Click to expand...

I don't think he said that as in nVidia shouldn't or deserve DX12 benefits or gains but rather why should they get same sorta gains as AMD are demonstrating or rather this bench is. Like many have said anyways, It's in alpha and it doesn't look quite right! especially when other nVidia cards are showing improvements and the 980Ti isn't quite getting anything as noticeable.


All we can take away is that AMD show some very good gains and that DX12 is great for performance and great for gamers. Many have said they didn't expect any performance uplift but things are starting to show overwise!
 
Yeah, not expecting nvidia to see double the frame rates as their DX11 performance is already very good, but to show negative scaling going to DX12... Well, something isn't right
 
D.P. said:
Because Nvidia can enjoy the exact same API benefits unless you want to believe some kind of grand conspiracy theory.
Click to expand...

They have different architectures, different drivers, there is entirely nothing anywhere to suggest we should see the same gains on an AMD card as an Nvidia card. With Nvidia guys constantly banging the drum about higher draw call performance in DX11 than AMD, the only logical conclusion is they WON'T make the same gains AMD will.

But the other factor is AMD have been working on low level API performance for longer than Nvidia, they are involved in two consoles for multiple years that run on low level APIs that they help with and their architecture seems to have been Mantle/DX12/low level API oriented feature wise since GCN 1.0, significantly longer than Nvidia. None of these things would make me presume that Nvidia would automatically make the same gains as AMD from DX12.
 
Funny how nvidia releases ashes of singularity driver few days ago, and suddenly something is broken in that game for nvidia :D I mean is they did not release the dedicated driver for this game, I would understand, but now, come on :D
 
drunkenmaster said:
They have different architectures, different drivers, there is entirely nothing anywhere to suggest we should see the same gains on an AMD card as an Nvidia card. With Nvidia guys constantly banging the drum about higher draw call performance in DX11 than AMD, the only logical conclusion is they WON'T make the same gains AMD will.

But the other factor is AMD have been working on low level API performance for longer than Nvidia, they are involved in two consoles for multiple years that run on low level APIs that they help with and their architecture seems to have been Mantle/DX12/low level API oriented feature wise since GCN 1.0, significantly longer than Nvidia. None of these things would make me presume that Nvidia would automatically make the same gains as AMD from DX12.
Click to expand...

something else people are ignoring, is that the normal test mode does not necessarily use over the 2million odd draw calls that the Nvidia DX11 driver can manage. Just becasue DX12 can do 12 - 20 million draw calls, does not mean they have done that in this test.

But when the High draw call test is performed, then nvidia cards see a performance improvement in DX12 over DX11.

And you can see the above on the PC Per benchmarks.
 
Last edited:
What i have seen in the most test, AMD promised DX 12 improvements are nowhere, i could say, hi i am here make some noise then gone. AMD FX series cpu wise , completely dead in this benchmark. Personally i don't see that hyped super performance at all. Especially amd said , they don't do dx 11 profile at all for this game/benchmark in DX 11. So the "massive" gain is with lot of salt. I bet AMD rep didn't said anything about this.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom