Soldato
- Joined
- 28 May 2007
- Posts
- 10,146
Tbh i had never seen this so must have missed it.
-
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.
AMD Mantle Vs. Direct X Battlefield 4 Multiplayer Benchmarks
- Thread starter LtMatt
- Start date
Tbh i had never seen this so must have missed it.
Soldato
It's anecdotal. Not seen any tests which show this 'smoothness' factor - those that say it are more often than not running multi GPU so they are likely seeing more smoothness from an increased and more consistent FPS along with higher corresponding minimums. None of which you're going to see on a single card if you're only getting an 8% bump.
Not to say something isn't there but I'm not convinced yet from what I've read. I put anecdotal evidence (i.e. "it feels smoother lads") pretty low down on the weighting. Again, not doubting this is being perceived but the reasons behind it to me seem correlative to the FPS being received.
Last edited:
Caporegime
8% is a substantial increase!
The difference between a 290 and 290x is often less than that, but people pay the extra £100+ eagerly, Mantle gives you the same for free.
Soldato
The 1st, 2nd and 4th graphs in the OP don't show a massive increase in 4K framerates, which is a shame because you have to think that is the way we'll be going (maybe not soon, but...).
Also a pity there was no eyefinity comparison, not that I use it, but I know some do.
Also a pity there was no eyefinity comparison, not that I use it, but I know some do.
In 1 game so far...
8% isn't substantial, especially on cards like 290's or 290x's where BF4 already runs very well anyway. Above 60fps at all times.
Soldato
It's not really when you look how much FPS it equates to. Would you be able to tell the difference between PC A and B where one is 45 FPS and one is 48.6 FPS? Probably not.
Anything is of course better than nothing and I do feel that BF4 isn't exactly showcasing Mantle's benefits in the proper way...
Soldato
- Joined
- 28 May 2007
- Posts
- 10,146
In the videos i have seen with performance overlays mantle graphs look smoother than dx. I don't know how this translates as i have never used it but it's not only crossfire. There's a good few on here saying the same with testing single card. It should be more apparent when mantle or bf4 becomes more stable.
Soldato
A graph looking smoother isn't the same as the actual game feeling smoother
.I just think you (figurative you) need to let the hype die down and then analyse properly what benefits are being provided. On the face of it 8% bump on max settings at 1600p ain't going to really make a difference.
For me on my 780 because I run it overclocked and drop MSAA to 2x I run with an average of 100-110 FPS at 1080p. It's smooth as it can get so if you translate the equivalent experience to Mantle capable cards I just don't think they're going to gain a phantom increase in smoothness.
Soldato
- Joined
- 28 May 2007
- Posts
- 10,146
People using 4k probably won't be running single cards though. It does see a good benefit in crossfire. 4 x aa does seem to lessen the benefits though but i doubt those frames on the 4xaa graph would be acceptable for a serious fps gamer. Lowering the settings would probably widen the gap again and adding another card in would probably do the same.
Caporegime
Look at the graphics card industry and tell me 8% is nothing. Look at the factory overclocked cards that provide a few percent benefit over the stock version. Look at the flagship cards from either manufacturer often there is a few percent between them, yet they'll push for every last percent to have the fastest card. Look at all the driver cheating that went on for a few percent increase.
Yes it's in one game at the moment but even so I can't see how ATI knocking out an increase in FPS, for free, is anything but insignificant. The fact it even works is pretty impressive to be honest.
Last edited:
Soldato
- Joined
- 28 May 2007
- Posts
- 10,146
A graph looking smoother isn't the same as the actual game feeling smoother
I just think you (figurative you) need to let the hype die down and then analyse properly what benefits are being provided. On the face of it 8% bump on max settings at 1600p ain't going to really make a difference.
For me on my 780 because I run it overclocked and drop MSAA to 2x I run with an average of 100-110 FPS at 1080p. It's smooth as it can get so if you translate the equivalent experience to Mantle capable cards I just don't think they're going to gain a phantom increase in smoothness.
"Where AMD Mantle actually did improve things though is in the frame times. While we weren't able to increase the gameplay settings, we did notice a smoother overall experience while gaming. This in itself is a big plus. The graphs backed up our feelings by showing AMD Mantle to be smoother, more consistent, and also produce lower time between frames, which is very important. This is probably AMD Mantle's biggest claim to fame for us right now, is the improvement in frame times which smooths out your gameplay experience. If this carries through to multi-GPU CrossFire, this could be a huge benefit."
http://hardocp.com/article/2014/02/03/amd_mantle_performance_preview_in_battlefield_4/5#.UwE2M4VV_Gc
Single card testing.
As i say plenty have said this is the case. Yea there is hype and it really would be nice to have a crystal ball to see what the case is when it's in a decent amount of games.
Soldato
You're coming at it from a different angle, one which isn't relevant to the point I'm making. As I said look at the FPS it translates to: it's neither here nor there for single cards. Again, anything is of course better than nothing but I wouldn't be shouting from the rooftops about a 3 FPS increase at 1440p on max settings is all.
Oh I thought AMD had already fixed frame times. My frame times are fine in DX so I don't necessarily believe that it is a flaw of DX and as a result, a benefit of Mantle.
Soldato
- Joined
- 28 May 2007
- Posts
- 10,146
Amd did fix there frame times to where the 7970 was only bettered by the Titan at the time. I have not looked into the 290x but i would imagine it's better than the 7970 and on par with Nvidia. Mantle is just a step above in this department from the graphs i have seen.
Caporegime
Different angle maybe, but you writing an almost 10% gain minimum off as it being neither here nor there is total nonsense I'm afraid.
We all overclock our kit, you could argue that just gets us a few FPS more. Why overclock your 780 if you wouldn't notice the difference compared to stock? You still do it though for the free performance. This is the same.
Last edited:
Soldato
The other issue with these graphs is that BF4 doesn't have an inbuilt benchmark, so every playthrough would have been done manually (wouldn't it?). This means there is obviously a chance for some variation to creep in.
Maybe the smaller gaps would actually be a little larger and the larger gaps a bit smaller if every playthrough was the same.
It's like the BF4 benchmark threads here, unless everyone that posted a result also posted a video of the playthrough you can't tell how closely they matched someone else's run. perhaps, just because of their playstyle they looked at the ground or sky more than someone else. Even one person playing through multiple times will likely get differing results.
It's a pity these games don't have inbuilt benchmarks and that the Star Swarm demo didn't have a benchmark option/mode.
Maybe the smaller gaps would actually be a little larger and the larger gaps a bit smaller if every playthrough was the same.
It's like the BF4 benchmark threads here, unless everyone that posted a result also posted a video of the playthrough you can't tell how closely they matched someone else's run. perhaps, just because of their playstyle they looked at the ground or sky more than someone else. Even one person playing through multiple times will likely get differing results.
It's a pity these games don't have inbuilt benchmarks and that the Star Swarm demo didn't have a benchmark option/mode.
Caporegime
Read the very first quote in the OP, it's a standardised benchmark.
Soldato
Make it clear in the OP that the percentage you're stating is for CF then.
8% on single card is a bit meh IMO. Need to see what it's like when it's been implemented from the off as opposed to patched in.
8% is 8%, for a 'patched in' attempt. I seem to recall the green team were blowing trumpets over driver improvements around the same margin. As TRD says below:
You won't notice 8%,
, what is it with people? One minute when slapping graphs everywhere it's everything e.g. my i5 is 8% better than your FX in crysis 1 <insert graph> yet in a thread comparing a new API working it suddenly a insignificant improvement.
Caporegime
It's because it's ATI only sadly.
Last edited:
Soldato
It's not total nonsense at all. How is a 3 FPS increase at 1440p really that significant? You would be guessing trying to pick a 45 FPS average out from a 48 FPS average line up.
Whereas I over clock my 780 because it provides a hell of a lot more than 8%
.You won't notice 8%,
No - you won't notice 8%.
Stating it's 8% faster (or whatever) is fine but is that really a game changer? Not really on its own. You seemed to have missed the part where I said twice that any improvement is not to be sniffed at. And again, when I said the true benefits will hopefully be shown when it's built in at an earlier stage as opposed to patched in. Of course these are AMDs first drivers too.
Last edited: