Amd mantle and ms dx12 support for assetto corsa

Firstly I'd like to say Hi to everyone on the forums and enjoying Kunos sim, and \/ to Kunos Dev Team.



I have been thinking how to approach Kunos about this suggestion and future development encouragement from the AC Community as a whole.

Let me first say, Assetto Corsa is a very beautiful sim, and I (and i hope most people would) appreciate the work and detail that have gone into this game. Both graphically and them masterful physics. Currently the graphics engine as I'd assume both Kunos and the AC players is not satisfied with its current efficiency, performance, consistency. DX11 clearly showing its limitations in 2014 from a lot of titles, not only Assetto Corsa but AAA titles as well with Ubisoft taking a huge hit with the the release of Watch Dogs & AC:Unity porting from Console (low level graphics API) to DX11.x (High overhead API).

I am humbly asking Kunos to take this matter seriously as it will have a ripple effect on the rest of the game development industry and to show everyone how its done. It is a lot of work (more so time consuming) to recreate the graphics engine in Mantle API. (Indie Dev Oxide Games managed in less than 3 months with 6 development members without any proper documentation as to how to do things in Mantle in Beta).

Currently Systems with intel and AMD CPUs with a AMD GPU is struggling heavily with a full grid on the track, there is also a heavy performance hit from 6 Cars or more in Multiplayer for both nVidia Setups and AMD Setups, with AMD GPUs struggling more so lagging 20% behind in performance in some cases.

Low Level Graphics API will be beneficial for VR users. We all know the importance of high FPS and VR screen low persistence is for a good VR Experience, as well as lowering motion sickness due to lag and stutter. It would be hugely beneficial if someone with a mid range PC, to get 100fps in medium graphics settings in either Mantle/DX12. VR Devices will most definitely have high ppi screens (lowest would be 1440p) and in Assetto Corsa we are struggling to get +100fps unless I am the only Car on the track I get 120fps with a R9 290. As a consumer I'm looking at this and wondering if developers are taking Low Level Graphics API seriously. The benefits is clear with Consoles. Low Level API you can harness emmense amount of performance by using it instead of DX11.x.

DX12 is not far from Mantle, as well as OpenGLNext(dare I say cousins) using a lot of code elements from mantle to implement within their own API and software ecosystem. (it is highly likely that Valve will be using OpenGLNext and Mantle for Source 2 for cross platform compatibility between Windows, Apple OS and Linux. This will be key pillars for the success of Source 2 and SteamOS. They see a clear future for these platforms. (hint hint Half Life 3 will be released in 2018).

Intel have to not only show interest but create a driver for DX12/Mantle as it wouldn't be any real difference between the two. This will bring Intel's onboard GPUs closer to current mid range GPU performance and that says a hell of a lot about the benefits. AMD Onboard GPUs clearly benefitting greatly by using Mantle. This brings the cost of the required gaming setups down, which means us the consumers benefit not only from the performance, but our wallets as well. This allows us to buy more games and or peripherals like Oculus Rift, better wheel, shifter, etc. Developers that have used mantle and working on DX12, outing Mantle to be more feature rich than DX12 upon release. DX12 will not be coming to Windows 7, but Mantle is available for Windows 7, 8, 8.1 and 10.

It would be beneficial for you to do Mantle in order to possibly prepare for DX12 and looking at the possibility of bringing AC it to consoles as porting between platforms is much easier from Mantle to PS4/XBone.


Please give us the options to select graphics APIs between Mantle and DX.

I have created a poll where us the players/gamers/sim racers/enthusiasts/etc and you Kunos can vote about this matter. You are welcome to post the Poll to your facebook page to allow some input from your fans.

http://strawpoll.me/3005590


We can expect AMD Mantle to be released within the next few weeks, (before December 31 2014).
Kunos New 2015 Year's Resolution (from the community to you).

• Familiarize yourself with Mantle ASAP
• Announce within the first 4 months of 2015 if you will add mantle support
• If, Yes Add as free "Mantle(low level API) DLC" for Assetto Corsa to list of DLC planned
• If successful, look at 2016 for DX12 possibility

PS* Kunos you have a bigger impact on game development globally than you would assume.

Kind Regards
Your Fan/Customer


Note* I know other users have posted in the forums about this topic. I'm not completely sure if a valid argument for Mantle for AC was made. As I am use you are well aware of AMDs announcement.

http://bit.ly/1xvfC14

 

I'd better go for DX12 since mantle is limited only to AMD users

 

Wrong. Mantle goes 'Open' if you will later this year, If I'm not mistaken.

 

Mantle is Open to All including nVidia, Intel for free, no licencing fee required to use. Leveling the playing field for game developers.
Mantle can be used crossplatform, as well as windows 7, 8.x, 10
Mantle has More features than DX12
Easy to Port from Mantle to Consoles or vise-cersa

vs

DX12 Comptibile with all GPU venders, does not work on linux, does not work on AppleOS
DX12 High licencing fee
DX12 Windows 8 and windows 10 Only, MS will not be releasing it for Windows 7 (announced last week/2 weeks ago) thus force OS Upgrade.

Don't forget Kunos is an Indie Dev, how money is spent matters.
If nVidia does not support mantle, well thats on nVidia, the door is wide open.

 

In that poll, where do I click to vote?

 

Interesting post...but here's the thing: 120 FPS is by no means bad. I think there's a little bit of an FPS war going on that eclipses the point. When you think about the baseline gaming FPS rate, i.e., 60FPS @1080p, your getting 200% more with a buffer of 60 FPS. Then you have the simple fix to all of this, which is to lower the graphics yourself as the user. Consoles impose this on you, PC gaming allows you to select the options to optimize your own system.

Understand that the reason FPS gets lower on multi-car races is not the graphical data, it is the physics the CPU processes. That accounts for "drops" on every system high and low. Console-esque features to PC in this regard seem redundant. And when you look at similar sim titles like pC, whose optimization is far less than AC, the point is even clearer, as AC has done about as much optimizing as you'd want. The point of AC is not to simply "perform" in the purest sense of the word, but to provide rich graphical landscapes and cars that simulate the real things. Console racing does not allow this and is the reason 67,000 members have chosen to play AC and contribute to the forums, m8.

Plus, I don't think you have considered Kunos' role in all of this, i.e., what their goals are for their simulator. I don't know how much you know about Kunos Simulazioni, but they are not a gigantic Turn10 Studios or Criterion firm that can crank out a game for millions of people. Kunos is a privatized group of enthusiasts whose aim is to bring a custom simulator to a group of fans who feel the same way about performance driving as they do.

My point? Kunos is not interested in reaching untold millions of users. Why? Because of their company ethos.

While we can all agree that massive numbers of fans are impressive, stop and think about the quality of those fans for a moment in games like Forza, GT, etc: out of their combined numbers, how many know the first thing about performance cars or how to properly drive them? Better yet, how many of those fans race on more than the most casual of gamer levels?

There is a reason the sims games have fewer members than the arcade/simcade games. There is also a reason why the sim crowd is a smaller, more dedicated group than the hordes who follow GT6, etc. It has to do with the reason why we game.

The sim crowd is a close-nit group of members who want to replicate real driving in every possible way from the physics to the graphics. If you go lowering the standard of the graphics department to gain new members, you would be doing so without considering what kind of new members you would be acquiring.

Kunos' ethos is to allow racing enthusiasts an affordable digital chance to experience the kinds of experiences real drivers feel. Because the group of people who want to "game like reality" is limited, a public outreach to millions by lowering aspects of the game is not really an option.

 

Id say in a few years time if Kunos make an AC2 they will likely look at this. They arent using their own gfx engine using yebis 2 I think it is. So unless that was to get updated to dx12 or mantle then its extremely unlikely for kunos to re write a new game engine at this point so close to release.

They may upgrade to a newer yebis engine if allowed in their license and if it doesnt take too much work with a significant benefit in doing so.

 

In the day`s of old, DX used to be a free update, not tied to whatever operating system.

MS saw this as a money making options, to force users to update the O/S.

Nvidia wrote code to hamper AMD cards in regards to early development of tessellation dx10.

I think M$/Nvidia/AMD need to be put in a boxing ring, and beaten with a wet fish until they see sense.
It`s the end user that suffers in the end.

mantle could be used on Nvidia cards, but you know NVidia to proud and all that, they see DX12 as there option to off load the cpu work load to the gpu.
Bit like what mantle does.

I would rather like to see ac use a work map for the cores, example of codemasters ego engine.

Code:

<?xml version="1.0" encoding="utf-8"?>
<workerMap>

  <workerThread name="core_0" priority="normal" core="0" stacksize="131072" />
  <workerThread name="core_1" priority="normal" core="1" stacksize="131072" />
  <workerThread name="core_2" priority="normal" core="2" stacksize="131072" />
  <workerThread name="core_3" priority="normal" core="3" stacksize="131072" />
  <workerThread name="core_4" priority="normal" core="4" stacksize="131072" />
  <workerThread name="core_5" priority="normal" core="5" stacksize="131072" />
  <workerThread name="core_6" priority="normal" core="6" stacksize="131072" />
  <workerThread name="core_7" priority="normal" core="7" stacksize="131072" />
  <workerThread name="gc_01" priority="abovenormal" core="0" stacksize="32768" />
  <workerThread name="gc_02" priority="abovenormal" core="1" stacksize="32768" />
  <workerThread name="gc_03" priority="abovenormal" core="2" stacksize="32768" />
  <workerThread name="gc_04" priority="abovenormal" core="3" stacksize="32768" />
  <workerThread name="gc_05" priority="abovenormal" core="4" stacksize="32768" />
  <workerThread name="gc_06" priority="abovenormal" core="5" stacksize="32768" />
  <workerThread name="gc_07" priority="abovenormal" core="6" stacksize="32768" />
  <workerThread name="gc_08" priority="abovenormal" core="7" stacksize="32768" />
  <workerThread name="debug_server_processor_0" priority="belownormal" core="0" stacksize="32768" />
  <workerThread name="debug_server_streamer_0" priority="belownormal" core="7" stacksize="32768" />
  <workerThread name="egonet_processor_0" priority="belownormal" core="0" stacksize="32768" />
  <workerThread name="egonet_processor_1" priority="belownormal" core="0" stacksize="32768" />
  <workerThread name="egonet_streamer_0" priority="belownormal" core="7" stacksize="32768" />
  <workerThread name="egonet_streamer_1" priority="belownormal" core="7" stacksize="32768" />
  <workerThread name="tessellation_0" priority="belownormal" core="0" stacksize="32768" />
  <workerThread name="tessellation_1" priority="belownormal" core="1" stacksize="32768" />
  <workerThread name="tessellation_2" priority="belownormal" core="2" stacksize="32768" />
  <workerThread name="tessellation_3" priority="belownormal" core="3" stacksize="32768" />
  <workerThread name="tessellation_4" priority="belownormal" core="4" stacksize="32768" />
  <workerThread name="tessellation_5" priority="belownormal" core="5" stacksize="32768" />
  <workerThread name="tessellation_6" priority="belownormal" core="6" stacksize="32768" />
  <workerThread name="tessellation_7" priority="belownormal" core="7" stacksize="32768" />
   
  <workerSet name="tessellation" jobpoolsize="1024" >
    <worker name="tessellation_0" />
    <worker name="tessellation_1" />
    <worker name="tessellation_2" />
    <worker name="tessellation_3" />
    <worker name="tessellation_4" />
    <worker name="tessellation_5" />
    <worker name="tessellation_6" />
    <worker name="tessellation_7" />
  </workerSet> 
 
  <workerSet name="generic_workers" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="collision" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <!-- solver cannot run on same threads as softbody -->
  <workerSet name="solver" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
  </workerSet>

  <workerSet name="softbody" jobpoolsize="1024" >
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="plugin_finalise" jobpoolsize="1024" >
    <worker name="core_1" />
  </workerSet>

  <workerSet name="cloth" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="handling" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="ai" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="vehicleSingleThreaded" jobpoolsize="1024" >
    <!-- IMPORTANT - maximise thread usage for this workerSet as it is used only during the single threaded update, hence all threads are waiting for it -->
    <worker name="core_0" />
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="pfx" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="visibility" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="ground_cover" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="ground_cover_workers" jobpoolsize="1024" >
    <worker name="gc_01" />
    <worker name="gc_02" />
    <worker name="gc_03" />
    <worker name="gc_04" />
    <worker name="gc_05" />
    <worker name="gc_06" />
    <worker name="gc_07" />
    <worker name="gc_08" />
  </workerSet>

  <workerSet name="anim" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="build_crowd" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="in_car_characters" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="storage" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="pssg_storage" jobpoolsize="1024" >
    <worker name="core_2" />
  </workerSet>

  <workerSet name="audio" jobpoolsize="1024" >
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>
 
  <workerSet name="damageSkinning" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <workerSet name="integrator" jobpoolsize="1024" >
    <worker name="core_1" />
    <worker name="core_2" />
    <worker name="core_3" />
    <worker name="core_4" />
    <worker name="core_5" />
    <worker name="core_6" />
    <worker name="core_7" />
  </workerSet>

  <!-- Debug server uses blocking IO and requires two independent worker sets to function -->
  <workerSet name="debug_server_processor" jobpoolsize="256">
    <worker name="debug_server_processor_0" />
  </workerSet>

  <workerSet name="debug_server_streamer" jobpoolsize="256">
    <worker name="debug_server_streamer_0" />
  </workerSet>
 
  <!-- Egonet uses blocking IO and requires two independent worker sets to function -->
  <workerSet name="egonet_processor" jobpoolsize="256">
    <worker name="egonet_processor_0" />
    <worker name="egonet_processor_1" />
  </workerSet>

  <workerSet name="egonet_streamer" jobpoolsize="256">
    <worker name="egonet_streamer_0" />
    <worker name="egonet_streamer_1" />
  </workerSet>

</workerMap>

Example 1.0 of Dirt3 8 core workmap.

so users could remap the work load, and in time when new cores/thead`s are added,
we could simply add/spread the loads to the core workmaps when new processor are released onto the market.

anyhoo just venting

 

I think your understanding of my post is hit and a miss.
1. I'm get 120fps alone on the track... Only my car, I'm not going to hotlap for the next 4-8 years alone in AC, I want to race with AI and with people, getting 40fps or lower with 6 or more cars on the track is really bad.
2. Mantle (low Level APIs) enables low CPU overhead which gives higher fps. AC has plentiful physics processes for the CPU. Thus AC will benefit 100% from Mantle. Which means any 4Core CPU with a mantle comptible GPU will get +60fps with a full grid of vehicles on the track.
3. I'm pretty clued up about the size of Kunos is less than 15 devs, I don't expect them to make a game to satisfy millions. If the software (the game) runs terrible on most hardware it was designed for then it should be redesigned. Creating a sim for PC is just that, the possibility of millions of different combinations, they only have to optimize for the most used, not for all.
4. m8 the things I higilighted have nothing to do with the topic at hand.
5. I honestly don't care about console players the untold millions you speak of. I said the option is there, i did not say they must create it for console.

"Your last line confused me" : Kunos' ethos is to allow racing enthusiasts an affordable digital chance to experience the kinds of experiences real drivers feel. Because the group of people who want to "game like reality" is limited, a public outreach to millions by lowering aspects of the game is not really an option.
6. Last Line what do you mean by lowering the aspects of the game? Using mantle will raise it. You can't game like reality when the sim is stuttering, you can't use Oculus Rift DK2 with an FPS lower than (minimum)75fps soon 90fps(minimum) with the new CV1 Dev Kit @ 1440p.

Most of the things you mentioned above have nothing to do with what members are concerned with. I share the passion you have for AC and am part of a local sim racing community as well. But it seems you aren't informed on the techincal aspects of creating a digital experience as you express and what is reality is far from eachother. Thus leveraging the technology at hand to bring the experience closer to what you talk about off topic. I also want to point out that my "performance" suggestions have a big impact on the experience and feel of the sim. a Big impact positively. The game will be exactly the same, the only difference is it won't require superman on a treadmill to run it.

I mean no offence to you in any way. I hope you are not offended by a bit of sarcasm hidden in there. Simply put your lack of programming experience shows in your post. Nothing is being lowered, no standard is being lowered graphically, Mantle is not inferior to DX11.x in anyway, it is superior. Also it would benefit you as a driver who want an even better experience with it, it will make it possible to run triple screen setups with higher FPS meaning smoother and more responsive racing and a better experience.

PS* It is currently cheaper to go to a track day for 1 year than to build a PC to run this decently. Thus You will get a better experience + trainer for cheaper. The affordable digital experience currently does not exist of which you speak.

Also look at my signature.. look at my rig.. The Sim is excellent the experience @40fps and lower is terrible for a rig this powerful.

 

my apologies i posted the link to the results, I edited the link, check the first post.

@KillZoneGB It would require the engine to be multithreaded for them to assign tasks to different cores, I'm not sure how well Kunos did on that part of the engine. I doubt it would be heavily multithreaded as there is no benefit reported between 4Cores and 8 Cores, or even 4 Cores + Hyper threading.

This would mean that the engine is not well optimized for multi-core multithread processing. It is also 32Bit which is even bigger handicap for current gen systems. This is also the reason as to why Intel CPUs is performing much better than AMD in this sim (this graphics +physics engine ) They can use OPenCl for physics, again there is many options..

Intel has better clock for clock 32bit performance than AMD's current FX Line. Intel Also has better single threaded performance than AMD. This is why there is such a huge performance CAP between using an Intel CPU and an AMD CPU. It is also one of the reasons Project cars can look so good and run so smooth, it is better optimized for multithreaded CPUs and has a 64bit variation of the engine which is more efficient and can manage memory better. In Assetto Corsa There is no benefit to having more than 4Gb RAM or a multicore CPU.

This is cause for concern for the rest of the sim racers and it shows there is still a lot of work to be done. They will have to rewrite + recompile the engine completely anyways in order for it to run natively 64bit on 64Bit OS and to take full advantage of the current Hardware available.

 

Also.. please don't be childish to 1 Star this forum thread as it is very important and beneficial to all of us. To whomever voted 1 Star.

 

I would like to see opengl and or mantle used instead of directx.. But i would like to see that for every game..
dont think it will happen though. Not unless sony decide not to use dx on consoles and Nvidia do chose to use mantle.
Probably not gonna happen though. and i doubt we will see anything like that happen for a nother 10 years or so atleast..
SO basically if your targeting gamers your looking at coding for direct x.. Anything on top of that is extra work and cost "time is money"
if your game is dx 11 right now it is REALLY unlikley that you will be changing to dx 12. The coding is Very different. And you have to be more carefull about how you code and then release the hooks and so on. "current dx and older ones would do checks for that sort of stuff. but this also makes it a bit slower"

and then you have mantle. I guess you could try and impliment it but unless nvidia decide to back it too then your looking to do that extra work for prehaps 40%-50% of the users (i am being generous there)

right now i can play online with a steady 60+fps at 1080p just had to mess with some smoke settings and so on. even if i we had mantle i wouldnt be able to max it out, Having said that it would be nice to be able to max out the settings and play online at 1080p.. But to do that i would need a couple of upgrades.
new cpu and new Gpu.. and i would need that either way so i dont see how it really affects us that much..

You cant just assume mantle will guarantee double your FPS. if the game is coded really well for the API it is using, and it has resources to spare mantle isnt going to magically double your fps.. AC is not THAT graphically intensive. its not really that cpu heavy either. Most of the issues with frame dorps and stuff happen when cpu and gpu useage is below 70% which means they both had room to work.
Mantle cant really help with that..

any way yeah id like to see mantle added to this game. I dont see dx12 being added at all. Not unless they make AC2 and start from scratch.
But i do not see it making the impact you think it would either.

Prehaps we just need slightly better coding so the game distributes loads better accross cores and utilizes more of them when needed..
"i have a sneeking feeling they tried to limit the cpu usage to stop those cpu warning messages"


P.S

Please dont double and tripple post.. you can use the edit button.

 

It may be more important to some then others, I have never even noticed stars or used them. But being a public forum everyone can choose how they rate a certain topic if its an option.

In a dictatorship you would just rate your own thread but it doesnt really show public opinion that way. Of course if you start a thread you think it has value but let the public decide for themselves if thats the case.

By making that post it may mean more people take notice of the starts and so will vote. You may not like the result but thats just bad luck.

 

No, you did not say specifically they need to create a console, you just asked for features that consoles have

Again, you're missing the point: AC is a game in development, the team is very aware of the areas which need improvement. The game's AI, MP, and damage physics will be some of the last things to finalize and will probably have future improvements...the main thing is these issues are already on their to-do list and will receive future optimizations. We must sit back and wait.

I think your post hits on the issue of optimization, but your solution is far too complicated. Just look at games like iRacing with 40,000+ members: they are still using DX9.0, and no one really cares (well, people do care, but you get my point ). The sim racing crowd is less interested with cutting edge DX technology, more interested in the driving experience. Kunos understands this and updates accordingly.

When you think about the fact that the game just came out a year ago, and that it has gone from 0 AI to a 24+ grid that is decent to race, there is really no need to complain or make requests. The team knows what needs to be done and will get it done in due time. I looked into this Mantle business, and I am just not that impressed. It certainly stands as an option for allowing massive optimizations, but from what the team has indicated so far, their planned optimizations are along a different rout. Believe me, they have similar hardware to yours, they know the issues and will make sure the latest cards and CPUs won't disappoint.

 

think that these stars "that i had not noticed before either" can help show what the community think of the idea as well.
if it ends up with 400 votes all 1 star then that speaks volumes..
if it has 400 votes only 40 1 stars then that also speaks volumes..
would i pay attention to a thread with 5 starts more than one with 1 or 0??.. nope makes no difference to me lol.. But possibly the devs care??
i dunno.

 

Ok ty, this is the poll: http://strawpoll.me/3005590 Had to take the /r out, but in your original thread post the link still goes to /results.

I also support Mantle, heck, even Directx12. Anything better than dx11 should be developed for the game.

 

well i cant vote on that poll or if i do I have to just select "no" It does'nt seem to have any other realistic options like.
"mantle would be a nice addittion if they have time to add it eventually"

p.s

just to illustrate what i was saying in my 1st post here..

depending on the game what mantle brings to the table is not always that great..



only a few fps here. Less than 10.. and its not like it takes the game from unplayable to now playable either. it really is not that much of a thing with The above game..

Obviously on other games like thief "which were really made to showcase mantle" there is a huge difference. But thief is a horrible game coded very strangely. when it released i used it for some benches "before it had mantle support" and i did say i think it performed like that in the bench tests just so it could do better when mantle was added to it..
A q6600 4 core at 3.7 scored the same "slightly better" than an 8120 at stock. and an i5-2500k at stock was 50% better. and watching the cpu usage you could see it was not using the systems to the full potential.

To me i believe i was proven right when mantle support was added and these results seem to confirm that too...


i wont say mantle does not allow games to utilize the hardware better. What i will say is if they take the time to utilize the hardware the best they can with dx 11 then using mantle instead will not massivly improve the frames per second..
Thiefs cpu usage is horrible And there is no reason for it to be like that Other than it was there to show off mantle.
Battelfield was made to work as best as it could and as a result mantle bearly makes a difference at all.
So like i said.. I really dont think it would make the impact you think it would.

 

I think you will be holding your breath for a long time waiting for AC to be 'updated'.

There really is no need.

 

you will have to excuse me, ive had a very rough night.

the thing is ac is not using 4 cores @ best dual core.
so you still got 2 cores on a fag break.

which could be put to better use, I/O lets say!

"many hands create less work"

 

Mantle only benefits people with slow CPU and fast GPU by reducing CPU bottlenecks. Dx12 seems a lot better as far as multi threading goes as it will spread the load equally. My experience with mantle has been nothing but a horror show..still get horrible lag spikes in bf4 using mantle and I get better fps using dx11.2.