AMD Mantle API [updating]

[mc6809e]

I am still slightly confused. The original AMD "45% faster with Mantle" slide specifically stated the hardware was Kaveri 7850k and 290x. Does this imply the more powerful the GPU the more load Mantle can offload from the CPU to the GPU and thus improve overall performance?

This is all about reducing the amount of work the CPU must do between the time a render function is called by a program and the time when the GPU takes over to do its job. There's plenty of code must be executed by the CPU before the GPU is ready to work. Mantle reduces this work.

What AMD has figured out is that users would rather not have to upgrade their entire system to take advantage of ever faster GPUs. A user that spends less money on a CPU (especially a high end Intel processor) has more to spend on an AMD GPU.

 

[Dave Glue]

What AMD has figured out is that users would rather not have to upgrade their entire system to take advantage of ever faster GPUs. A user that spends less money on a CPU (especially a high end Intel processor) has more to spend on an AMD GPU.

...not to mention the significant hassle of upgrading your CPU/motherboard, or purchasing a new system. New video card is a 10 minute install vs. several hours for a new motherboard.

What I'm really curious now though is to see if this would enable more disparate crossfire setups to work effectively. I'm talking within the same GPU generation of course, and I wouldn't expect a DDR3 $50 board to work well with a 290x, but say someone with a R9 270 upgrades to a 290X...but they they want to keep the R270 and have it assist in rendering where it can. Would Mantle help to enable this, or at least effectively enough to make it worthwhile in Mantle-supported titles?

 

[liolio]

I wish somebody will test a Kabini based part, it is too bad it is not on sale yet in a socketable format.
I want to know what Mantle or something akin to it could bring to the mobile realm.

 

[loekf]

Total fail by AMD for not releasing Catalyst 14.1 yesterday AT THE SAME time as the BF4 patch came out.

What are they thinking ?

 

[Psycho]

Darn... http://www.techpowerup.com/197425/amd-catalyst-mantle-driver-delayed.html

AMD now tells us that their teams have been working overnight to fix the driver, and should have their next status update for us by mid-afternoon EST (New York time), later today. That update doesn't necessarily mean a driver release at that point in time, and so AMD might be forced to label it Catalyst 14.2 beta, keeping up with its calendar-based driver version naming. Catalyst 14.1 beta was expected to ship the first public distribution of Mantle, AMD's ambitious 3D graphics API to rival Direct3D and OpenGL.

AMD also mentioned a 24 hour exclusive period for press to evaluate the driver before public release, so it looks like your download will be at least 36 hours away.

 

[SimBy]

Total fail by AMD for not releasing Catalyst 14.1 yesterday AT THE SAME time as the BF4 patch came out.

What are they thinking ?

They were thinking, we need to fix it so that it works as intended. Some fail to grasp what AMD and devs achieved here in a very short time. How long did MS take to get DX to a level it is today? And Mantle wants to be above that.

 

[Picao84]

They were thinking, we need to fix it so that it works as intended. Some fail to grasp what AMD and devs achieved here in a very short time. How long did MS take to get DX to a level it is today? And Mantle wants to be above that.

We are not failing to grasp anything. It is not about achievement. It is about timing and presentation.

Furthermore, its not like they started Mantle at the same level DX was started. There is a difference between technical implementation and concepts. Many of the things that were added to DX over the years were not even conceptualized at the time DX development started. So Mantle got at least the benefit of those concepts existing already. They did not start from tabula rasa.

 

[Osamar]

I want to see when possible a test with A10-7700K and R7 260 or something similar.

 

[shiznit]

The results from the blog are skewed by the use of 4xAA, disable that and Mantle will likely offer a bigger boost even in the high end Intel CPU scenario. Most people going for 120hz in BF4 couldn't care less about AA.

 

[overclocked_enthusiasm]

Darn... http://www.techpowerup.com/197425/amd-catalyst-mantle-driver-delayed.html

$100 says they take the weekend to tweek it and release it on Monday. Why does this feel like DICE released the Mantle patch without AMD's knowledge...or AMD wasn't ready and DICE just decided "sorry guys...we got a patch to release!". Either way it is sloppy and unprofessional to not have a driver ready when the debut of your in-house api is released!

 

[CarstenS]

It was promised that the Mantle patch would be out in Dec... then January. This has been fulfilled thus.

 

[Zeross]

While waiting for a true reference manual of the API, at least we can console ourselves reading the functions names
http://pastie.org/8682650

 

[3dilettante]

Total fail by AMD for not releasing Catalyst 14.1 yesterday AT THE SAME time as the BF4 patch came out.

What are they thinking ?

It was likely planned to hit the same day, then for some reason it couldn't hit the deadline.

It's not the case that AMD is selecting to futz things up because it feels like it.

 

[homerdog]

I think loekf's point still stands. Makes it look like AMD couldn't coordinate a shit to hit the inside of the toilet bowl.

 

[3dilettante]

That's a given.
Asking "what were they thinking" indicates that they thought about what they wanted and then chose to mess it up.

 

[Picao84]

That's a given.
Asking "what were they thinking" indicates that they thought about what they wanted and then chose to mess it up.

No but, at the least, the BF4 patch should have been launched only after the driver was released. That would have avoided all this mess.

 

[overclocked_enthusiasm]

That's a given.
Asking "what were they thinking" indicates that they thought about what they wanted and then chose to mess it up.

I don't think anyone is accusing the AMD driver team of sabotage...let's just agree this is yet another AVOIDABLE missed opportunity to change the perception about AMD launches/support.

 

[LordEC911]

No but, at the least, the BF4 patch should have been launched only after the driver was released. That would have avoided all this mess.

You do realize the patch had other things in it than just Mantle, right?

 

[silent_guy]

Not that I disagree about the avoidable release sync issues, but in the grand scheme of things, this is really not such a big deal: AMD has always had a bad reputation about drivers (deserved or not) and this won't make it worse.

In fact, it's probably the best option: delaying a few more days may displease a few thousand of whiny entitled kids, but the alternative may have been much worse: crashes, crashes everywhere.

 

[Picao84]

You do realize the patch had other things in it than just Mantle, right?

Yes, but what bad could possibly come out of delaying it 1 or 2 days, so AMD could iron out the driver?