• 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.

Another Developer Rebellion Throws Weight Behind AMD & Mantle

Nickolp1974 said:
I really hope it works and Nvidia don't adopt it, as they will probably just produce a faster card to compensate and that could possibly mean a faster card in ALL games not just ones supported by mantle. It also may put an end to them holding back and just release the true high end card initially.
Click to expand...

Nvidia already push the boundaries on how fast there cards can be. They can't just suddenly magic faster cards. Nvidia have the same manufacturing process to work with that amd do, this round Nvidia never held anything back they just couldn't manufacture there top end chips in enough numbers (due to poor yields) to release them at a decent price. Nvidia are not in the game of holding back especially when they don't currently hold the fastest gpu ie 7970/gtx680. Nvidia will most likely find it even harder on 20nm to get there huge 780ti like chip out and may well use a similar strategy to what they did this round.
 
Last edited:
MjFrosty said:
I don't think the developers would agree with you there. A lot of developers have been victim to the sums that EA throw at them, only to later on realise the pressure they're put under. You'd be extremely naive to believe that EA don't have a very large input into their games. I could probably find a list of developers that have gone under with EA involved.

When you're paying sometimes up to as much as 860 million dollars, you expect results. Also it would have most definitely been EA that pushed BF4 through quality control, not DICE.



http://www.dsogaming.com/editorial/...image-and-the-resignation-of-its-co-founders/

http://beforeitsnews.com/entertainm...-of-bioware-warns-of-pressure-ea-2449168.html
Click to expand...

Gregster said:
If anyone thinks that Dice have as much time as they need to develop BF4, they are seriously deluded. EA will be cracking the whip and deadlines must be met.

I do not sympathise with EA at all.
Click to expand...

Fair point.
 
For anyone interested. This is a recording of the back room Mantle developers summit, its the boring technical stuff that wasn't streamed or recorded, except for some guy in the audience.

 
Last edited:
LtMatt said:
We do but most of them get ruined with fan boy smack talk so sometimes its nice to start a fresh one.
Click to expand...

I feel the biggest problem facing Mantle is the talk talk talk and nothing backing it up as yet. Not one performance graph (we all love a graph) and speculations of 50% faster in Mantle games are getting the industry buzzing but still nothing from the ones developing it.

When Mantle is up and running, it would be hard to smack talk if it produces.... G-Sync threads were the same and if I recall Matt, you did your best in most of them to dismiss the tech. You won't allow anything bad said against Mantle but G-Sync is rubbish? You can see how this makes you look :)

Not digging by the way and just saying how I see it :)
 
humbug said:
For anyone interested. This is a recording of the back room Mantle developers summit, its the boring technical stuff that wasn't streamed or recorded, except for some guy in the audience.

Click to expand...
Interesting to hear how it works. I am not a rendering expert but it seems to me that the biggest advantages appear to be:

- Command buffers and resource descriptors to allow more pre-calculation for batches of commands that can be used across frames. This could dramatically reduce the number of calls being made to the driver per frame.

- Command buffers can be constructed in different threads and are only synchronised when being queued for render (not sure how this would work in practice, how would you guarantee rendering order?) allowing multiple cores to be used for rendering.

- Exposing individual GPUs allows the developer to control what gets processing on each. This is particularly interesting for a-symmetrical cases (like an APU and a PCI-E GPU) where sharing rendering evenly would produce poor results. The developer could decide to use one half as a co-processor for physics or something.

- Direct memory management of VRAM. Should allow developers to make much more efficient use of space and could lead some innovative uses for content streaming and stuff.

Sounds like a fair bit of work to make to best of it but the potential sounds good.
 
Last edited:
You have to remember that on consoles most of that is pretty standard anyway, controlling memory and having direct control of the GPU. That means most dev's have been doing most things Mantle offers for PC for several generations of console to one degree or another and will most certainly be doing almost all of it for GCN on consoles already. This is precisely why AMD/DICE have said along the lines of, extremely similar to coding for PS4, and taking optimisations from PC to a console version of vice versa, becomes very little work.

Ultimately we're talking about so much similarity that most of what you do is applying the same knowledge and control to another platform. The side effect of giving far more control, is far less overhead, which will increase performance on gpu and cpu.

That the consoles are so similar to PC games and in particular AMD architecture that as support for games in 360/ps3 drops off and the dev's learn to code particularly for that hardware, most of that improvement will be directly applicable to Mantle.

With Kaveri and HSA you have an even more similar platform to the consoles with unified memory which should help cut down on wasted copys and pci-e traffic, every little bit here and there ends up meaning the gpu is waiting for data that 1ms less long. At 120fps you're talking about 8ms a frame, every single microsecond you can save doing something that isn't needed is time something needed gets done and it all does it earlier.
 
Why do peeps want Mantle to Fail, if AMD could get this of the ground it means more $$$$$ coming in to keep them moving forward.


Plus it could bring forward Steamroller as well.:)
 
OLDPHART said:
Why do peeps want Mantle to Fail
Click to expand...

Scepticism =/= wanting It to fail..

If you have been a pc gamer long enough you will be familiar with the amount of rubbish we have been told about that was supposed to bring huge performance gains...but never really did.
 
Gregster said:
When Mantle is up and running, it would be hard to smack talk if it produces.... G-Sync threads were the same and if I recall Matt, you did your best in most of them to dismiss the tech. You won't allow anything bad said against Mantle but G-Sync is rubbish? You can see how this makes you look :)
Click to expand...

See below Greg. I mentioned that once, at the time when it was announced and kept it to one thread.

Jono8 said:
Scepticism =/= wanting It to fail..
Click to expand...
 
A brand new API that reduces abstraction is going to be faster, it's how these things work. One that can run rendering threads and scale to any number of cores is going to make a big difference, it can't not.
 
You must log in or register to reply here.
Back
Top Bottom