• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Rise of the Tomb Raider DX12 Performance Benchmarks

Page 5 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Mahigan said:
I'm not sure if NV is running the Async on feature in Hitman or if the default to Async off.
Click to expand...

No matter if its async or not, why does Tonga, aka half Fiji regress in performance in DX12? Why doesn't GCN 1.0 get any benefits? Why is it that pretty much just GCN 1.1 gets all the fun? And for the rest, its simply about CPU overhead.
 
ShintaiDK said:
No matter if its async or not, why does Tonga, aka half Fiji regress in performance in DX12? Why doesn't GCN 1.0 get any benefits? Why is it that pretty much just GCN 1.1 gets all the fun? And for the rest, its simply about CPU overhead.
Click to expand...
Probably because the DX12 implementation in the game is in it's early stage and there is a lot more work to be done.
 
Noctifer616 said:
Probably because the DX12 implementation in the game is in it's early stage and there is a lot more work to be done.
Click to expand...

Even then, if you are not CPU limited in any way. Then DX12 isn't going to be better as some kind of defacto standard. (Besides in performance/watt on the CPU side maybe.)

Its quite clear that with DX12 that developers now have to do everything for, is a far reach from DX11 with years of optimization that has been done by AMD/NVidia. So for a foreseeable future this is the result I expect. Haswell i3 gains 17% for example, while a 6700K loses. And for console ports, the first task is to help the anemic console CPU.

Its no different from the move from DX9 to DX10. Even with 20% lower CPU overhead, DX10 ended up slower.
 
Last edited:
I'm going to assume that the developers already gathered enough information about performance regresses in live systems that they didn't see in their testing equipment to improve DX12 performance in the next patch.

That's by far the most logical and simple explanation that I can think of, since they were very clear that most people should see performance improvements, not regresses.
 
ShintaiDK said:
Most people already do. All those that dont own a 5960X/6700K etc.
Click to expand...

While we expect this patch to be an improvement for everyone, if you do have trouble with this patch and prefer to stay on the old version we made a Beta available on Steam, Build 629.1, that can be used to switch back to the previous version.

We will keep monitoring for feedback and will release further patches as it seems required. We always welcome your feedback!
Click to expand...
They are being very specific.
 
Update:
Just played it for another two hours, half an hour on DX11 and the remaining time on DX12. Couldn't tell a difference, certainly nothing like the 39 fps vs. 32 fps that the benchmark would lead me to believe. By this point I don't think there is a performance problem specific to Tonga outside of the benchmark.
 
Piroko said:
While we expect this patch to be an improvement for everyone, if you do have trouble with this patch and prefer to stay on the old version we made a Beta available on Steam, Build 629.1, that can be used to switch back to the previous version.

We will keep monitoring for feedback and will release further patches as it seems required. We always welcome your feedback!
Click to expand...

They are being very specific.
Click to expand...

Seems like a very indirect way of acknowledging there is an issue. Hopefully they don't consider that fixing it.
 
ShintaiDK said:
No matter if its async or not, why does Tonga, aka half Fiji regress in performance in DX12? Why doesn't GCN 1.0 get any benefits? Why is it that pretty much just GCN 1.1 gets all the fun? And for the rest, its simply about CPU overhead.
Click to expand...

Tahiti doesn't get much benefit in Ashes when testing Async on/off as well. 2 ACEs may not provide big performance gains when factoring in context switch/fencing overhead and perhaps non-optimal implementation. This is the reason Sony worked with AMD from 2009 to iterate and improve GCN to have 8 ACEs.

Their current implementation results in only weak FX and i3s seeing some gains (when paired with top GPUs, a config that we can agree, few would ever run), and only in some areas where its a CPU bottleneck. However, the MTR should be a benefit. When folks in the know present about DX12/Vulkan, the few key points has always been a massive reduction in CPU overhead, getting more performance out the CPU. It should not regress in performance, period. That implies some poor coding.

They are in the learning phase. So one should wish them luck and look forward to more patches, rather than champion what is a broken implementation.
 
Last edited:
Actually is it all that far off the sort of CPU/GPU config you might see in gaming laptops?

Mobile GPUs are about to roughly double in speed after all. Something like a ULV dual core + 1060M configuration? Can quite easily see that existing.

Or I guess if we're allowed to dream about external GPUs, but those probably are a rather long shot.

Do think broken is a bit strong - you need to leave adjectives for stuff like GoW! A little sub optimal, but maybe expected for a first/initial try.
 
Sigh. I thought I had a dialed in solid OC on my video card after hours of gameplay in other games without issue. This I get a nvidia graphics driver crash running dx12 with this game...Would you consider this a game issue or an unstable OC?

BTW Very High Preset @ 1440p
DX12: 82 FPS
DX11: 86 FPS
 
MBrown said:
Sigh. I thought I had a dialed in solid OC on my video card after hours of gameplay in other games without issue. This I get a nvidia graphics driver crash running dx12 with this game...Would you consider this a game issue or an unstable OC?

BTW Very High Preset @ 1440p
DX12: 82 FPS
DX11: 86 FPS
Click to expand...

I'd just lower the overclock and see what happens if it was me.
 
prtskg said:
What a game! Regression of performance going from dx11 to 12. Why would anyone bother with dx12 in this game?
Click to expand...
I get ~25% higher minimum framerate with DX12 on my already ancient 975Mhz 270 2GB (Haswell 4C/8T @ 3.7 Ghz, 32gb RAM). Game runs and feels way smoother. Definitely, no complains here. As a bonus, my CPU utilization is now properly spread across the cores/threads. A win-win.

P.S. DX11 feels like a bad dream to me, now. Windows Vista/7/8/8.1 are now officially dead for gamers. Hail to the new King, baby!

steam.png
 
Last edited:
You must log in or register to reply here.

TRENDING THREADS

Back
Top