AMD Mantle API [updating]

[silent_guy]

If BF4 Mantle has the same graphics as BF4 DX11, then 17% or even 10% faster in the highest-end configuration Ultra setting is huge: it may not be a game changer for the players, but seeing how GPU benchmarks results are judged by the slightest percentage differences, this can mean a major win for AMD in many cases.

I think Mantle is a really big deal just for that. The even better lower-end performance improvements are just gravy.

 

[CarstenS]

I really wonder, if this will tempt Nvidia to open up Physx to be completely available on the CPU, asking reviewers to compare "their codepath" with the "normal codepath".

 

[overclocked_enthusiasm]

Mantle, being the huge game changer AMD expect it to be, was thrown out the door like an ordinary driver update. Hell, I even remember nVIDIA making a fuss about a "bing bang" driver update that was nothing special some years ago. Mantle surely deserved a more polished presentation? Including reviewers numbers right of the bat.

Your tone is too combative dude...chill a bit. I for one would have liked to see a better roll-out of both Kaveri and Mantle. Here are the missteps I am critical of:

1. The "Dual Graphics" page still doesn't list Kaveri yet http://www.amd.com/us/products/technologies/dual-graphics/pages/dual-graphics.aspx#3

2. The latest driver update is December 18 http://support.amd.com/en-us/download/desktop?os=Windows+7+-+64

3. Kaveri was hyped with Mantle by AMD. Keep in mind the "45% faster" Mantle slide by AMD was specifically noted with a 290x. This implied dual graphics compatibility to many and resulted in mass confusion until the 240/250 dGPU clarification was made by AMD. Yet the AMD site still does not list Kaveri or the 240/250 optimization. Throw in the confusion of DDR3 vs GDDR5 dGPU compatibility and you have a VERY confused retail consumer AND some very confused review sites not knowing which way to go. The result? Buyers sitting on their hands or buying Nvidia.

What AMD is doing is leaving the interpretation of the benefits of their hardware or API to the conclusion of others rather than taking control of the narrative with a consistent and compelling message. The resulting confusion is RAMPANT. It could have and should have been handled better...especially with a product as big as Kaveri. AMD needs to be more consistently excellent when they roll out MUCH hyped hardware and associated drivers/APIs etc. ESPECIALLY at this fragile state in the company's history. This is a marketing issue...plain and simple.

My apologies if off topic...

 

[Picao84]

What AMD is doing is leaving the interpretation of the benefits of their hardware or API to the conclusion of others rather than taking control of the narrative with a consistent and compelling message. The resulting confusion is RAMPANT. It could have and should have been handled better...especially with a product as big as Kaveri. AMD needs to be more consistently excellent when they roll out MUCH hyped hardware and associated drivers/APIs etc. ESPECIALLY at this fragile state in the company's history. This is a marketing issue...plain and simple.

Exactly! Thank you for probably having said it better than me.
I am afraid my combative attitude is general and not specific to AMD or Mantle. Hard to solve

 

[overclocked_enthusiasm]

If BF4 Mantle has the same graphics as BF4 DX11, then 17% or even 10% faster in the highest-end configuration Ultra setting is huge: it may not be a game changer for the players, but seeing how GPU benchmarks results are judged by the slightest percentage differences, this can mean a major win for AMD in many cases.

I think Mantle is a really big deal just for that. The even better lower-end performance improvements are just gravy.

I agree. Double digit gaming increases in the most popular games can, in and of itself, be the reason gamers will choose a particular brand of dGPU.

 

[Alexko]

If BF4 Mantle has the same graphics as BF4 DX11, then 17% or even 10% faster in the highest-end configuration Ultra setting is huge: it may not be a game changer for the players, but seeing how GPU benchmarks results are judged by the slightest percentage differences, this can mean a major win for AMD in many cases.

I think Mantle is a really big deal just for that. The even better lower-end performance improvements are just gravy.

Well, it depends. If it's 10~20% in 10% of the big games, that's just 1~2% overall, i.e. well within the margin of error of most benchmarks. If it's 10~20% of 50% of the big games, then yes, it's a major competitive advantage.

 

[xDxD]

cut

Under 7970+8350, the difference is even bigger. Notice how the GPU is a straight line under Mantle, compared to DX:

DX:

Mantle:

seems like Mantle have been designed to bring more benefits to the AMD' cpus than to vgas...

 

[Tamlin]

seems like Mantle have been designed to bring more benefits to the AMD' cpus than to vgas...

To me, it looks like Mantle have been designed to bring benefits to the whole pipeline, when you see GPU frame time being a straight line like this.

 

[homerdog]

I shall post test results from my HD7950 + i5-3450S rig ASAP. When will I be able to get the Mantle driver?

Then I shall swap out the 3450S with my old i5-3550 (got a 3770K for free in my main rig) and see what gains I get. Should be interesting

Then I shall test with my 3770K + GTX670 rig... will be interesting to see if the AMD rig takes the lead over the 670 which has been faster since day 1. Let the games begin!

Should I even go so far as to pair the 7950 with the 3770K and see if Mantle takes better advantage of Hyperthreading? So many options..

Would like to test with an AMD CPU but the only one I have is a S939 Athlon X2 3800+

 

[xDxD]

To me, it looks like Mantle have been designed to bring benefits to the whole pipeline, when you see GPU frame time being a straight line like this.

right, I mean in terms of pure performance

 

[3dilettante]

seems like Mantle have been designed to bring more benefits to the AMD' cpus than to vgas...

Was the entire swath of presentations, dev and AMD statements, and everything else about getting around the DX11 API and driver model not indicative of that?

I thought the information was pretty consistent in that the big problem was CPU and driver bottlenecks that left the GPU underutilized.
There were additional benefits beyond that, but they pounded the API/driver issue pretty hard.

Is it the case that BF4 is a port to Mantle, and thus not able to leverage it as much, or has its role as the early adopter given DICE at least some opportunity to massage the design?

 

[DavidGraham]

I have the impression that actually most of them are presented with 4xAA?

No. All benchmakrs are run with Ultra and FXAA or 1XMSAA .. which means no MSAA. all except the CF Dice benchmark.

Then I shall test with my 3770K + GTX670 rig... will be interesting to see if the AMD rig takes the lead over the 670 which has been faster since day 1. Let the games begin!

The patch also improved performance on DX path, it also made specific NVIDIA optimizations that were previously absent, but present on AMD hardware since day one.

-Tweaked and improved CPU multi-core utilization to try and avoid stuttering and low framerate due to stalls that happened on some CPU and OS configurations.

-Enabled tile-based compute shader lighting optimization on Nvidia for improved GPU performance (already active on AMD GPUs)

 

[Deleted member 13524]

seems like Mantle have been designed to bring more benefits to the AMD' cpus than to vgas...

No, it's been designed to bring more benefits to cheap and lower-performing CPUs than to GPUs, and every single PC gamer without a limitless budget for gaming hardware (and laptop users) should be happy with that.

 

[xDxD]

Was the entire swath of presentations, dev and AMD statements, and everything else about getting around the DX11 API and driver model not indicative of that?

I thought the information was pretty consistent in that the big problem was CPU and driver bottlenecks that left the GPU underutilized.
There were additional benefits beyond that, but they pounded the API/driver issue pretty hard.

Is it the case that BF4 is a port to Mantle, and thus not able to leverage it as much, or has its role as the early adopter given DICE at least some opportunity to massage the design?

Well, when you present your new flagship (R9 290x) and at the same time your new revolutionary Api, the message seems to be that the performance of the vga will benefit most from these Api.
But in fact the emphasis on overcoming the bottleneck of draw calls, had to be understood that it is perhaps the most studied for the Apus than anything else

p.s.: i hope my english is sufficiently clear..

No, it's been designed to bring more benefits to cheap and lower-performing CPUs than to GPUs, and every single PC gamer without a limitless budget for gaming hardware (and laptop users) should be happy with that.

ok ok, nothing wrong with that, but sounds like "if you can not beat Intel hw side, do it with the sw".

 

[xDxD]

edit double

 

[DavidGraham]

ok ok, nothing wrong with that, but sounds like "if you can not beat Intel hw side, do it with the sw".

Nothing worng with that too, they deserve to reap the rewards for their efforts on the APUs field, which got them into becoming the dominant hardware provider for consoles, which opened the door for them to create a new API. if Intel did all that, then I bet they would do the same.

 

[Tamlin]

Starswarm demo is out now on Steam for those that want to take it for a spin while waiting for Mantle driver.

 

[Kaotik]

Starswarm demo is out now on Steam for those that want to take it for a spin while waiting for Mantle driver.

For the lazy
http://store.steampowered.com/app/267130/

 

[silent_guy]

Well, it depends. If it's 10~20% in 10% of the big games, that's just 1~2% overall, i.e. well within the margin of error of most benchmarks. If it's 10~20% of 50% of the big games, then yes, it's a major competitive advantage.

If a lot of people buy a new GPU for just a few of flagship games (BF4 probably qualifies?) and those can be convinced to use Mantle and be a bit faster, that may be enough.

 

[overclocked_enthusiasm]

If a lot of people buy a new GPU for just a few of flagship games (BF4 probably qualifies?) and those can be convinced to use Mantle and be a bit faster, that may be enough.

My hardware decisions were solely based on WoW for almost 4 years...

*hangs nerdy head in shame....