Ah come on, that's rubbish. A good programmer is a good programmer and will make full use of the tools available. A bad/lazy programmer won't become a good hard working programmer by removing the drawcall limitation.
-
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 DEVELOPER VIDEO 44 mins
- Thread starter Farooq
- Start date
Ah come on, that's rubbish. A good programmer is a good programmer and will make full use of the tools available. A bad/lazy programmer won't become a good hard working programmer by removing the drawcall limitation.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
It's NOT a 10% increase in cost, the quote is talking about the specific small situation of porting the game. IE $50mil game dev cost, $10mil to port to two platforms at $5mil a piece, that means Mantle add's $500k to the cost... or, of $60mil, a tiny amount.
Porting already costs, that's how it works, it's benefit vs cost. HD3D, Eyefinity is used by a fraction of those Mantle would effect, and both add cost and time to ports. THat is life, the quote specifically indicates the 10% is for the porting cost ONLY and adds that it is a drop in a bucket compared to the full game development costs.
So cost isn't an issue, games cost money to develop, always have done, porting has always cost money, always has done, adding new features has always added more costs, always has done. This is nothing new, Mantle, at 10% increase in porting costs, which can benefit every single AMD user that has bought a GCN gpu based product in the past 3 years and into the future is a massive user base, 10% is nothing.
THen we get to the "we want numbers" people, what numbers. In general optimisation is the LAST step of releasing a game. Meaning to compare Mantle numbers you HAVE to have the shipping product, the final version, anything else is retarded, and it's ridiculous to suggest you want some.
If Game X runs at 30fps unoptimised 3 months from launch, and runs at 60fps at launch because they finished the game then spent 3 months optimising it. What possible use could releasing the 30fps number be? There is zero benefit to anyone. This isn't a chip, an Intel chip that gets demoed 12 months before release, is already the final hardware, even then we often get demoed a half speed chip..... In this situation the shipping code will be significantly different from 3 months before shipping, to the shipping date.
Why would anyone compare finished and optimised(for performance, not stability
) BF4 DX11 vs unoptimised unfinished Mantle?
Lastly, the function of increased performance is to USE that performance.
Most gamers if they have 120fps, do not want 180fps, they want increased visual quality. An increase in IQ that without the "extra" performance would drop the previous 120fps down to 80fps. But due to the new found performance these effects get added, and you still get 120fps.
This is what happened with DX10, and 11 for the large part. DX10 games didn't instantly run faster, because in reality DX10, and 11, and every new version usually adds support in hardware for things people wanted to do but it was too costly in the previous API. So DX10 came about, performance didn't jump 20%(in situations where the API could fully enable higher performance), they would add a new effect, and often times they added an effect that made it slower overall.
IE DX9 at 60fps, + lighting feature that needed huge hacks to make work, huge overhead, and killed performance to 20fps... so dev's don't enable it in DX10. under DX10 this overhead was fixed, so they added this lighting feature, it still dropped performance from 60fps, but only to 50fps.
Porting already costs, that's how it works, it's benefit vs cost. HD3D, Eyefinity is used by a fraction of those Mantle would effect, and both add cost and time to ports. THat is life, the quote specifically indicates the 10% is for the porting cost ONLY and adds that it is a drop in a bucket compared to the full game development costs.
So cost isn't an issue, games cost money to develop, always have done, porting has always cost money, always has done, adding new features has always added more costs, always has done. This is nothing new, Mantle, at 10% increase in porting costs, which can benefit every single AMD user that has bought a GCN gpu based product in the past 3 years and into the future is a massive user base, 10% is nothing.
THen we get to the "we want numbers" people, what numbers. In general optimisation is the LAST step of releasing a game. Meaning to compare Mantle numbers you HAVE to have the shipping product, the final version, anything else is retarded, and it's ridiculous to suggest you want some.
If Game X runs at 30fps unoptimised 3 months from launch, and runs at 60fps at launch because they finished the game then spent 3 months optimising it. What possible use could releasing the 30fps number be? There is zero benefit to anyone. This isn't a chip, an Intel chip that gets demoed 12 months before release, is already the final hardware, even then we often get demoed a half speed chip..... In this situation the shipping code will be significantly different from 3 months before shipping, to the shipping date.
Why would anyone compare finished and optimised(for performance, not stability
) BF4 DX11 vs unoptimised unfinished Mantle?Lastly, the function of increased performance is to USE that performance.
Most gamers if they have 120fps, do not want 180fps, they want increased visual quality. An increase in IQ that without the "extra" performance would drop the previous 120fps down to 80fps. But due to the new found performance these effects get added, and you still get 120fps.
This is what happened with DX10, and 11 for the large part. DX10 games didn't instantly run faster, because in reality DX10, and 11, and every new version usually adds support in hardware for things people wanted to do but it was too costly in the previous API. So DX10 came about, performance didn't jump 20%(in situations where the API could fully enable higher performance), they would add a new effect, and often times they added an effect that made it slower overall.
IE DX9 at 60fps, + lighting feature that needed huge hacks to make work, huge overhead, and killed performance to 20fps... so dev's don't enable it in DX10. under DX10 this overhead was fixed, so they added this lighting feature, it still dropped performance from 60fps, but only to 50fps.
Thats not the point I'm making. Not sure how clearly its coming across, I dislike enabling people who do sloppy work in the first place.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
As with everything, a good programmer will find workarounds. Currently the work goes on working around draw calls, remove the draw call limit........ you WILL find another bottleneck, so the programmers will find something else to work around, some other way to improve performance. This is how coding works, removing the biggest bottle neck that takes the most work to fix, means, they can find something else big to work on.
It's also worth pointing out that basically everyone working on it has said the thin layer and basically all the responsibility being on the dev means they can try entirely new techniques and play around, finding new things to do. Often with the current DX and driver situation, they can have a great idea, code it, but find that it doesn't work... this could be something, a new lighting technique, who knows, whatever, but it might not work purely because DX or the driver is doing something daft.
Dev's left to their own devices with nothing really stopping them, WILL find cool new things to do, and one of the key features of Mantle, is it gets out of their way.
Caporegime
- Joined
- 18 Oct 2002
- Posts
- 33,188
NO, it wasn't, the first mention of it was december, with no date at all given. Dice have since said late december. Either way, if DICE are late then DICE are late. It's got nothing to do with AMD or Mantle. When, i actually forget who it was, Ubisoft delayed Watchdogs by several months, is that the fault of MS because DX sucks, or because, you can't always judge how long it will take to finish a project?
Dice and Oxide and a couple of others got the first(I assume stable) build of Mantle 2.5 months ago or so, how long guys take to implement it, is how long they take and zero reflection of Mantle or AMD.
No I understand what you are trying to say. But, I don't and can't agree with it. A sloppy programmer will always be sloppy and it will eventually catch up with him/her.
The drawcall issue is holding back performance on a PC and hamstringing developers. They currently are wasting time getting around the drawcall issues when they could be working on the harder problems that you mentioned earlier.
And I don't get your enabling comment either. Read Frenchtart's comment earlier in the thread, he says it very well.
Soldato
- Joined
- 28 May 2007
- Posts
- 10,222
I think mantle itself could be the incentive. The ability to showcase what you are capable of bringing to the table could be enough of an incentive as dx is holding everyone back from fully utilizing the power of the pc platform. The original crysis for instance was a pc showcase mainly in the graphical department. A lot of people don't rate the game but it still gets talked about these days due to its visuals. This game is what really made crytek and maybe others will want to follow. Mantle could make things easier for any developer wanting to showcase a pc only next gen visual master piece. The dx version of any such game would still be available for nvidia but would just run slower on there hardware.
Last edited:
Funny that since I own two 7990s
Check your facts before you start accusing someone of being a 'fan boi'
I just can't see why they can state mantle improves performance by 5%, 10%, 20% or whatever if the whole thing isn't ready yet. I'm not saying I want a playable demo but a simple monitor by monitor video of one setup using a 290 card and the other running Mantle on a 290 card to show the increase in performance.
If they can state it increases performance then I want to see for myself now.
Last edited:
Yes but imagine what a good programmer/game dev can do when they don't have to worry about draw call limitations as much anymore. They can move on to big and better things making more use of their time. Still those "bad/lazy" programmers will still be bad/lazy programmers but it will also make things easier for them too. End of the day I can see it benefitting gamers either way.
Soldato
- Joined
- 28 May 2007
- Posts
- 10,222
Funny that since I own two 7990s
Check your facts before you start accusing someone of being a 'fan boi'
I just can't see why they can state mantle improves performance by 5%, 10%, 20% or whatever if the whole thing isn't ready yet. I'm not saying I want a playable demo but a simple monitor by monitor video of one setup using a 290 card and the other running Mantle on a 290 card to show the increase in performance.
If they can state it increases performance then I want to see for myself now.
I believe he was talking about no 1 dave with regards to the gtx780. I believe people in this thread have stated why amd can't and won't put a % on mantle's benefits. The main reason being that mantle will never give a fixed % performance boost.
Tell that to Rroff, that's part of the point I have been trying to get across.
ahh yea sorry, just read through again. But yea I agree
Funny that since I own two 7990s
Check your facts before you start accusing someone of being a 'fan boi'
I just can't see why they can state mantle improves performance by 5%, 10%, 20% or whatever if the whole thing isn't ready yet. I'm not saying I want a playable demo but a simple monitor by monitor video of one setup using a 290 card and the other running Mantle on a 290 card to show the increase in performance.
If they can state it increases performance then I want to see for myself now.
What are you talking about, you just picking posts to try and look cool with your 2 x 7990's
Last edited:
Hi lads nice to see the video got us talking so here is another enjoy its Mantle from Johann Anderson from DICE
http://www.youtube.com/watch?v=N_6CAneoW-0
so here is another enjoy its Mantle from Johann Anderson from DICEhttp://www.youtube.com/watch?v=N_6CAneoW-0
Last edited: