• 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.

[H] Tests Crossfire V SLI with Mantle

Will Robinson

Golden Member
Some good results for the Mantle code path here...better FPS and frame pacing results.
http://www.hardocp.com/article/2014/03/04/bf4_amd_mantle_video_card_performance_review_part_1/3

1393620031nTfVKdLjSj_3_1.png

And frame pacing...
1393620031nTfVKdLjSj_5_2_l.png
 
The increase in performance is nice, but the elimination in frame latency spikes is the most impressive improvement, for me. Looks like a single GPU instead of typical AFR inconsistencies seen in the DX frame/time graphs. "Smoother and Faster" No more giving up absolute performance to have the best latency times. 😉
 
That surprises me. I thought Mantle would offer much more benefits in CF. I kept hearing it would offer much more boosts.

It appears the main benefits of Mantle, as of now, is CPU bottleneck removal, and more consistent frame times.
 
+12% speed is nice, but making it work more like a single card is very nice.

If enough demanding games get Mantle support this could make CF the preferred choice for multi-card setups.

I'm still happy with my single GTX 680 though 🙂
 
That surprises me. I thought Mantle would offer much more benefits in CF. I kept hearing it would offer much more boosts.

It appears the main benefits of Mantle, as of now, is CPU bottleneck removal, and more consistent frame times.

Yes, as of now that seems to be what it's offering. It's not like that's not a worthwhile thing though. It's only the first game to use it and get it out in mass on peoples computers. Early early days right now.
 
Yes, as of now that seems to be what it's offering. It's not like that's not a worthwhile thing though. It's only the first game to use it and get it out in mass on peoples computers. Early early days right now.

Its good to see they finally moved to Win 8.1 for testing BF4 and the results are a good sign for future Frostbite titles.

We have to go back to why Mantle exists, and its because devs wanted something better than what Microsoft was willing to offer with their DirectX. But with DX12 with its focus on less CPU overhead, not sure if Mantle has longevity since Microsoft appears to be stepping up.

In the meantime, its great for BF4 players!
 
That surprises me. I thought Mantle would offer much more benefits in CF. I kept hearing it would offer much more boosts.

It appears the main benefits of Mantle, as of now, is CPU bottleneck removal, and more consistent frame times.

I'm sure removing MSAA would see the performance jump exponentially. Mantle takes a serious hit with MSAA right now.
 
12% at 1600p with 4xMSAA is a decent result, I really like the frame times too. Would like to see 1080p results for 120/144hz users, I expect Mantle will have a better showing at that res too.
 
It no surprise that the gains are rather small. They did say they haven't done any GPU optimisations yet.
 
If one reads the final conclusion, using the settings tested, which were gpu limited, they state that there is really no final difference in gameplay on high end single cards, but a significant improvement in frametimes with crossfire. Also still not working on lower end cards.
 
If they haven't done any GPU optimizations yet, that means only CPU bottleneck is addressed. Right? Then where are the SLI mantle scores? 😉
 
CF XDMA + Mantle: big alternative to Nvidia frame pacing in SLI.

If they haven't done any GPU optimizations yet, that means only CPU bottleneck is addressed. Right? Then where are the SLI mantle scores? 😉

Is optimized, but still a little buggy.



And a reminder: 2560*1600 4xMSAA is heavily GPU bound. 1920*1080p no AA is much less.
 
It seems reasonable to suppose that it maybe a fair learning curve for the devs now that they have greater access to the GPU processing units rather than a library of D3D effects to work with.
I expect Mantle GPU optimizations will come in fits and starts as they figure out how to exploit the new,lower level access.
 
If one reads the final conclusion, using the settings tested, which were gpu limited, they state that there is really no final difference in gameplay on high end single cards, but a significant improvement in frametimes with crossfire. Also still not working on lower end cards.

What doesn't work on lower end cards?

Amazing how many people are now interested in 1080p/0xAA benchmarks...

They should be interested because mantle adds around 50 fps higher minimums for 120/144hz gamers.
 
The fact that AMD was able to nail down frame pacing, which has plagued multi-GPU setups since the beginning, I think is the most impressive part of this. There's no point in getting high FPS if it still feels like it's lower. Very promising.
 
I'm sure removing MSAA would see the performance jump exponentially. Mantle takes a serious hit with MSAA right now.

The reasons I was surprised as the small boost (the boost I predicted prior to Mantle's release), was because 1) I heard a few things that suggested CF was supposed to scale much better under Mantle and 2) CF tends increases your FPS, so your CPU becomes more of a bottleneck.

I do realize that if you turn down settings for higher FPS, it should show to be a bigger improvement. While most people do not do this, it is something I'd do.
 
There are really two types of PC gamers: those that go for all out IQ, and those that go for high FPS.

Mantle is ideal for those after the highest FPS possible.
 
The fact that AMD was able to nail down frame pacing, which has plagued multi-GPU setups since the beginning, I think is the most impressive part of this. There's no point in getting high FPS if it still feels like it's lower. Very promising.

Yup, I'm getting more and more happy with my 7990 investment this fall. I got it from new for $750 when the shop cleared their inventory (before price were $1300).
 
Back
Top