The AMD Mantle Thread

[krumme]

You posted the COMPUTE benchmarks, not the actual game benchmarks.

I posted the latest Civilization 5 game benchmark on Anandtech above.. The GTX 680 and 770 are both more GPU limited at those settings, than the GTX 780..

Who cares. Its yesterdays engine that took forewer to build and cost a fortune. Tomorrow its oxide SWARM. No lead thread. Using mantle. Its 300% difference on the application level. Rts just need a new api. Always had. Period. And i bet and hope nv is going to do the same for swarm too. As indicated by oxide themselves.

 

[itsmydamnation]

another set of benchmarks not matching pclab

http://www.purepc.pl/karty_graficzn...nych_battlefield_4_mantle_vs_directx?page=0,2

again the mantle numbers from pclab.pl are low in comparison


Carfax83 i'll reply to your post tomorrow.... its late and im tired.

 

[DarkKnightDude]

Basically all the benchmarks are useless in MP to be honest, since nothing is ever the same? Personally my experience of Mantle has increased my FPS and smoothness like pretty nicely, they just need to fix the random framerate dips.

I guess at this point its dependent on your setups.

 

[itsmydamnation]

Basically all the benchmarks are useless in MP to be honest, since nothing is ever the same? Personally my experience of Mantle has increased my FPS and smoothness like pretty nicely, they just need to fix the random framerate dips.

I guess at this point its dependent on your setups.

i disagree i have been logging all my games via the console and when i do a bell curve distribution the mean and avg are within 1 fps on the same map on pretty much every run i have done.

I have a 6970 and i run BF on medium with no AA AO and a 3770k@4.3

 

[sontin]

another set of benchmarks not matching pclab

http://www.purepc.pl/karty_graficzn...nych_battlefield_4_mantle_vs_directx?page=0,2

again the mantle numbers from pclab.pl are low in comparison


Carfax83 i'll reply to your post tomorrow.... its late and im tired.

AFAIK they used the SP. And so far there is no difference to the pclab.pl numbers from the SP.

 

[Arkadrel]

another set of benchmarks not matching pclab

http://www.purepc.pl/karty_graficzn...nych_battlefield_4_mantle_vs_directx?page=0,2

again the mantle numbers from pclab.pl are low in comparison


Carfax83 i'll reply to your post tomorrow.... its late and im tired.

Those where all under windows 7 64bit though.

Still it seems fishy if no one else is giveing results near what pclab posted.
And its not unthinkable that a review site would fake results to show amd in bad light/nvidia in better light, if their bias went that way.

 

[McLovin42]

Basically all the benchmarks are useless in MP to be honest, since nothing is ever the same? Personally my experience of Mantle has increased my FPS and smoothness like pretty nicely, they just need to fix the random framerate dips.

I guess at this point its dependent on your setups.

Always something that has to be "fixed" with AMD.........round and round on the AMD merry go round we go! :whiste:


We don't tolerate thread crapping here and you should know this.

-Rvenger

 

[McLovin42]

The fact that mantle is beta, the driver is beta and this is the first release of mantle on BF4.


i dont get what people dont want to analyze this properly, they would rather "WIN"..........

And this is the problem......everything with AMD always seems to be a "beta" product.............by the time the bugs are fully worked out, the ap is properly streamlined etc, the entire hype will be yesterdays news and not really matter performancewise...........AMD releases too much early cr@p to try and keep pace with Nvidia's own solutions.

WHat AMD needs to concentrate on is less marketing fluff and hype trying to compete with Nvidia, and put more money into hardware/parts improvements, design improvements, QC, and software/driver development.

Mantle overall is IMO a waste of time and effort and is no better than a good quality driver optimization in most scenarios. AMD just used their marketing team to create the aura to all the AMD lemmings that it was going to be the next latest greatest thing for their built in bacon n egg breakfast cookers inside their case. In reality its on the same level as a quality driver release. Who cares how a low cost budget cpu/gpu combo gets a big boost though regardless? LOL!


You can have a day off for that and when you get back, stay out of this thread.

-Rvenger

 

[GaiaHunter]

Mantle overall is IMO a waste of time and effort and is no better than a good quality driver optimization in most scenarios. AMD just used their marketing team to create the aura to all the AMD lemmings that it was going to be the next latest greatest thing for their built in bacon n egg breakfast cookers inside their case. In reality its on the same level as a quality driver release. Who cares how a low cost budget cpu/gpu combo gets a big boost though regardless? LOL!

If it is cheaper to create a good quality driver optimization that gives the same performance boost then it makes no sense to spend time and money on Mantle.

Of course a good quality driver optimization apparently does nothing for low cost CPU while Mantle does as you point out.
So maybe some people care more abut the low budget PCs than the high budget PCs.

 

[McLovin42]

If it is cheaper to create a good quality driver optimization that gives the same performance boost then it makes no sense to spend time and money on Mantle.

Of course a good quality driver optimization apparently does nothing for low cost CPU while Mantle does as you point out.
So maybe some people care more abut the low budget PCs than the high budget PCs.

Yes and no.

I have seen driver releases do wonderful things for budget cpu/gpu setups in the past. But not to the extent that Mantle has shown with a big asterisk since its only with very limited and few specific combos on the lower tier end.

Problem is, AMD misled people on the marketing hype..........specifically to those with upper tier products and setups..........I don't remember hearing them state to people in that crowd that the benefits would be minimal at best..........did you? Or anyone else?

Again, its what AMD/ATI has done year after year.......lead people on, or rather mislead people on raw fps #'s even though overall the game may look and play like crap with bugs galore and lots of annoying artifacting...............................but hey, that's what sells products. ^_^

 

[krumme]

Basically all the benchmarks are useless in MP to be honest, since nothing is ever the same? Personally my experience of Mantle has increased my FPS and smoothness like pretty nicely, they just need to fix the random framerate dips.

I guess at this point its dependent on your setups.

I share your experience but remember all the benchmarks of MP show a good/significant improvement of Mantle. And ofcourse it varries wildly dependent on how the test is made and the setup. But thats also how gaming in reality is - it is varried. What map, what route, what is happening. All that needs also to be explained in the test.

There is one exception to this picture but to make up for it, its quoted like 50 times in this thread.

 

[krumme]

i disagree i have been logging all my games via the console and when i do a bell curve distribution the mean and avg are within 1 fps on the same map on pretty much every run i have done.

I have a 6970 and i run BF on medium with no AA AO and a 3770k@4.3

I noticed and saw that. Its a good point and well argumented.

 

[krumme]

AFAIK they used the SP. And so far there is no difference to the pclab.pl numbers from the SP.

Can you explain why SP numbers is interesting?

 

[Shivansps]

Who cares. Its yesterdays engine that took forewer to build and cost a fortune. Tomorrow its oxide SWARM. No lead thread. Using mantle. Its 300% difference on the application level. Rts just need a new api. Always had. Period. And i bet and hope nv is going to do the same for swarm too. As indicated by oxide themselves.

no individual ai and dumb group "follow the leader AI" is incluided in that? Star Swarm does not have 5000 real units on screen its about 100, its like counting the units inside a unit on CIV5.
Rendering things that are not needed? ugly effects and graphics? no real combat?

If that the future, i rather go back to 90s thanks.

I can accept that BF4 MP shows some results, but Star Swarm is just the worst tech demo i ever see. In fact in the way they are counting the units, ive seem larger battles on sins of a solar empire rebellion, 4 Advent will bring intro battle a total of about 8000 fighters + full capitals.

 

[GaiaHunter]

Problem is, AMD misled people on the marketing hype..........specifically to those with upper tier products and setups..........I don't remember hearing them state to people in that crowd that the benefits would be minimal at best..........did you? Or anyone else?

I'm sure everyone that bought an AMD card did so because it had Mantle and not because it had a better price/performance than the competition.

Also it is fun to notice that somehow a BETA of Mantle is the definitive improvement you will ever see from Mantle.

Which is interesting because DX that has been going for years can only get better but Mantle that is in a BETA release will never improve.

It is also interesting to notice that Crossfire somehow isn't AMD upper tier product and setup.

 

[96Firebird]

It is also interesting to notice that Crossfire somehow isn't AMD upper tier product and setup.

I may have missed them, can you link to reviews that use CrossFire and Mantle?

 

[GaiaHunter]

I may have missed them, can you link to reviews that use CrossFire and Mantle?

Not that I'm aware.

What does exist is the Dice CF numbers with crossfire but considering that the other numbers released by Dice have been pretty much spot on, I see no reason to doubt them at the moment.

EDIT: Actually there maybe some but I don't know what kind of reliability/reputation this site has.

http://www.hardwareluxx.com/index.p...tx-update-multiplayer-benchmarks.html?start=1

 

[sushiwarrior]

Problem is, AMD misled people on the marketing hype..........specifically to those with upper tier products and setups..........I don't remember hearing them state to people in that crowd that the benefits would be minimal at best..........did you? Or anyone else?

Again, its what AMD/ATI has done year after year.......lead people on, or rather mislead people on raw fps #'s even though overall the game may look and play like crap with bugs galore and lots of annoying artifacting...............................but hey, that's what sells products. ^_^

I don't seem to recall AMD stating anything other than "Mantle will remove API overhead" and "It can improve performance by 45% on an A10-7850k". Where did you get the idea that removing overhead would help systems that can already handle the overhead easily?

I don't think AMD needs to spend anything on "better quality parts/components", whatever that means I agree that Mantle is probably too cost/effort intensive to provide a good return on investment, but if all/most Frostbite titles use it then there's definitely more to it than just BF4.

I'm curious to see how much new Mantle driver updates can improve performance, I bet there's a ton of horsepower still on the table for GCN1.0 cards that just needs an optimized driver to fix.

And just a slight tip for all those experiencing chugs every 10-15 seconds, based on my experience with the dev build, this is due to objects not loading correctly in BF4. Stuff like meshes and models not being fetched properly and throwing errors, or even things like explosions which throw ray-casts too far (not sure how that happens exactly, but that's what the error console says). So as far as I know, this can be fixed with a BF4 patch, and is not necessarily Mantle's fault...

 

[OCGuy]

at just needs an optimized driver to fix.

And just a slight tip for all those experiencing chugs every 10-15 seconds, based on my experience with the dev build, this is due to objects not loading correctly in BF4. Stuff like meshes and models not being fetched properly and throwing errors, or even things like explosions which throw ray-casts too far (not sure how that happens exactly, but that's what the error console says). So as far as I know, this can be fixed with a BF4 patch, and is not necessarily Mantle's fault...

If it works fine on Sony's API and D3d, but the game needs to be patched so Mantle works, does that go against one of the selling points of Mantle?

 

[sushiwarrior]

If it works fine on Sony's API and D3d, but the game needs to be patched so Mantle works, does that go against one of the selling points of Mantle?

I don't know, I'm just sharing an observation based on the dev build, since AMD is slow with driver releases then at least there's the hope that Dice can patch the issue faster than AMD.

There's likely to be issues when you overhaul any build, one of the selling points of Mantle isn't "it's 100% error free always and forever no matter what". They obviously have changed some code with regards to fetching assets and rendering explosions, which needs a patch from either AMD or Dice to correct. I'm not saying it's Dice's fault, I'm not trying to save face for Mantle, because Mantle is having issues either way and that's bad for a new product (and especially for a product with so many delays). Just giving a tip which may suggest the issue can be fixed quickly.

 

[bystander36]

If it is cheaper to create a good quality driver optimization that gives the same performance boost then it makes no sense to spend time and money on Mantle.

Of course a good quality driver optimization apparently does nothing for low cost CPU while Mantle does as you point out.
So maybe some people care more abut the low budget PCs than the high budget PCs.

This helps more than those with budget CPU's, although BF4 may not need Mantle as much as some types of games.

Mantle will also help those with a desire to get a constant 100+ FPS. Those with Lightboost monitors will like this a lot.

 

[Carfax83]

Carfax, your argument for Civ5 makes no sense to me. Besides that, I found a test with an x290 and a 780 in Civ5.

http://www.overclockers.com/amd-r9-290-graphics-card-review

In an even more cpu stressful situation the Nvidia cards still don't show that much of a lead. They are leading the AMD cards, sure, but not by anything you'd find too unusual outside of game by game variance. I think you have a point with the Nvidia drivers being more efficienct than the AMD ones in general, but not anywhere near the level you are trying to make it out to be (or the effect of dx11 multithreading).

One of the reasons why I, and many others come to Anandtech is because Anandtech uses thorough, detailed and consistent testing procedures.

That Overclockers review doesn't even list the drivers they used.. But more to the point, as someone else pointed out, that's not the in game benchmark.

That's obviously something else that is less reliable, as the GTX 780 is leading the Titan, something which shouldn't happen.. Anandtech uses the late game benchmark, which simulates a late game workload, and is way more consistent.

 

[Carfax83]

So what you are saying is that for comparable graphic cards from AMD and NVIDIA there is no difference.

I'm not sure how that is proving your point.

I don't know how you get that from what I said

My point was that DX11 multithreading DOES work, otherwise the GTX 780 wouldn't have such a large lead.

People keep forgetting that DX11 multithreading only works in CPU limited circumstances. That's why I used the GTX 780, because it's the most CPU limited card in that test.

At 1080p, the GTX 770 has a 15% lead over the 7970 GE, and the GTX 780 has maxed out the benchmark FPS. If Anandtech benched at a lower resolution, the GTX 770's lead would widen over the 7970 GE, because the benchmark would become more CPU limited.

 

[sushiwarrior]

I don't know how you get that from what I said

My point was that DX11 multithreading DOES work, otherwise the GTX 780 wouldn't have such a large lead.

People keep forgetting that DX11 multithreading only works in CPU limited circumstances. That's why I used the GTX 780, because it's the most CPU limited card in that test.

At 1080p, the GTX 770 has a 15% lead over the 7970 GE, and the GTX 780 has maxed out the benchmark FPS. If Anandtech benched at a lower resolution, the GTX 770's lead would widen over the 7970 GE, because the benchmark would become more CPU limited.

All you're showing is a GPU limited benchmark, where a faster GPU performs better. That has nothing to do with CPU performance or CPU driver performance. Note how the 7970GE is faster than the 7970, if they were tied, then maybe you'd have somewhat of a point.... but you're not proving any point with that graph.

 

[Spjut]

All you're showing is a GPU limited benchmark, where a faster GPU performs better. That has nothing to do with CPU performance or CPU driver performance. Note how the 7970GE is faster than the 7970, if they were tied, then maybe you'd have somewhat of a point.... but you're not proving any point with that graph.

Ye, and notice the 670 and 680 don't have a strong lead over the 7970 and 7970GE