The AMD Mantle Thread

[desprado]

Ah ok.....when you give percentages, I tend to think FPS. "Performance" is a weasel word and cannot be measured.

Yup...............Agree to that.
My guess is i am getting 120fps on bf4 and after is shift to mantle i get 150fps which useless for me but those users which are getting 25 to 30fps without mantle that are real beneficiaries.

 

[ShintaiDK]

Also be careful what get measured. As we already seen with Tomb Raider example.

Prescripted benchmark 800Mhz Haswell vs 4800Mhz Haswell. 40% difference.
Ingame playing, 400% difference.

In something with a very low CPU load, like a prescripted benchmark and comparing to AMD drivers without command list in DX. Its not hard to get better performance when Mantle is to brign what AMD have been unable to implement in their DX drivers. Unlike nVidia.

For example a 2Ghz FX8350 is as fast as a full speed version with Mantle. What they maybe didnt tell is, that it might have been just as fast in the prescipted benchmark with a ~4Ghz X2. And nobody will tell us. Mantle is a classic example of conflict of interest.

 

[blackened23]

nevermind

 

[selni]

20% on a new tool with a few months of use instead years and years of use?

Not bad.

Wait until they actually know what they are doing.

If the idea of mantle was to make PC development more "console like" then presumably they already do?

 

[Minkoff]

If current games do not receive increase in FPS that much, can it be different for latter ones?
If, CPU overhead is greatly reduced and multithreading receives additional improvements that leaves the CPU idling or actively assisting the GPU with…let’s say better in-game physics? What I mean is…a bullet hitting a wall and the collapsing wall (physics) is better suited for the GPU, but the flying bullet and the structural damage made to the wall and the decision whether or not the wall should collapse is better suited for the CPU. Correct me if I am wrong…
So, games without Mantle, will see a massive hit in FPS, due to the CPU already busy…, and the games with Mantle will maintain stable FPS.
Can this be considered as gain in FPS?

P.S BTW, I thought that that the main increase in performance and price in Iris pro was due to the eDRAM?

 

[SiliconWars]

Ah ok.....when you give percentages, I tend to think FPS. "Performance" is a weasel word and cannot be measured.

If a game is already at 80 fps average I'd rather the extra performance was put into further draw distances etc instead of another 20 fps. There's really not much point in just pushing up the fps for the sake of it.

These guys are artists and would rather see eye candy. That's a lot more impressive than a fraps number at the top of the screen.

 

[Gloomy]

For example a 2Ghz FX8350 is as fast as a full speed version with Mantle. What they maybe didnt tell is, that it might have been just as fast in the prescipted benchmark with a ~4Ghz X2. And nobody will tell us. Mantle is a classic example of conflict of interest.

Well since the whole point of that test was to showcase how Mantle removes CPU bottlenecks, then that 4GHz X2 would prove their point just as well.

 

[krumme]

If current games do not receive increase in FPS that much, can it be different for latter ones?
If, CPU overhead is greatly reduced and multithreading receives additional improvements that leaves the CPU idling or actively assisting the GPU with…let’s say better in-game physics? What I mean is…a bullet hitting a wall and the collapsing wall (physics) is better suited for the GPU, but the flying bullet and the structural damage made to the wall and the decision whether or not the wall should collapse is better suited for the CPU. Correct me if I am wrong…
So, games without Mantle, will see a massive hit in FPS, due to the CPU already busy…, and the games with Mantle will maintain stable FPS.
Can this be considered as gain in FPS?

P.S BTW, I thought that that the main increase in performance and price in Iris pro was due to the eDRAM?

Gaming in bf4 what counts most is winning , on 64 player maps, consistently high perf, with no or very small fps dips, is of outmost importance for most bf4 gamers - they tend to lean to the serious side.
If mantle can remove those dips or minimize them its a huge advantage that means more than average fps. If it comes from gpu or cpu does not matter.

The same goes for many rts. Worst performance in the game tends to shape experience.

The thief ppt. remark about mantle beeing able to hinder some of the typical stalls with reloading textures is imho of outmost importance in my world for real word gaming experience.

 

[GaiaHunter]

If the idea of mantle was to make PC development more "console like" then presumably they already do?

Some of it, yes.

But the console never had as much resources as high end PCs have.

Also the PS4 and Xbone are fairly new.

 

[frozentundra123456]

Yes, various sources have said 20%-50% improvements. There is a video with 3x higher performance. The BF4 chief architect has said Mantle will have better performance and better graphics.

What more do you want?

Maybe some solid numbers from something besides a hand picked demo that was obviously designed as a best case for mantle?

That is the only case where specific numbers have been mentioned. I honestly don't have a clue where you are getting that 20 to 50 percent number? If you have a link where someone states that, please provide it. Otherwise you are just guessing like everyone else. You are certainly entitled to do that, just don't keep quoting it as a solid number unless you can document it.

 

[SiliconWars]

Maybe some solid numbers from something besides a hand picked demo that was obviously designed as a best case for mantle?

That is the only case where specific numbers have been mentioned. I honestly don't have a clue where you are getting that 20 to 50 percent number? If you have a link where someone states that, please provide it. Otherwise you are just guessing like everyone else. You are certainly entitled to do that, just don't keep quoting it as a solid number unless you can document it.

http://forums.anandtech.com/showpost.php?p=35745311&postcount=2072

 

[frozentundra123456]

http://forums.anandtech.com/showpost.php?p=35745311&postcount=2072

An undocumented source from AMD supposedly quoted in Fudzilla? OK???........ And they dont say what the improvement is actually, Is it FPS or some other metric.

As far as I know AMD hasn't directly quoted any performance numbers, even in BF4, which is their showcase for Mantle. So I still prefer to wait for hard data, even from AMD, but more so from independent reviewers.

 

[VulgarDisplay]

AMD guys said nothing and if they know how much performance gain they would tell by now.

See this
As reported by tech report:
Andersson didn't bring up performance estimates, but other developers who discussed Mantle at APU13 did. Jorjen Katsman of Nixxes, the firm porting Thief to the PC, mentioned a reduction in API overhead from 40% with DirectX 11 to around 8% with Mantle. He added that it's "not unrealistic that you'd get 20% additional GPU performance" with Mantle.

Guys say u could get 20% performance although he did not mean at least so what mean to say he not sure but 20% is there target.

That say an additional 20% more GPU performance. It doesn't talk about overall increase in performance from adding the additional CPU performance in as well. I suspect the amount will vary greatly based on the type of game, and 20% will likely be the minimum improvement we see in games that aren't CPU limited.

 

[Paul98]

An undocumented source from AMD supposedly quoted in Fudzilla? OK???........ And they dont say what the improvement is actually, Is it FPS or some other metric.

As far as I know AMD hasn't directly quoted any performance numbers, even in BF4, which is their showcase for Mantle. So I still prefer to wait for hard data, even from AMD, but more so from independent reviewers.

Specific numbers aren't something they can come up with since the gains are going to depend on the developer and what the game needs. Depending on the game and the needs it could give a 20-30% gain or a 200-300% gain. We know there are a few things that all games will benefit from, there are other things it's going to mainly depend on the game and the developer.

 

[Minkoff]

Could this be real? ...and is it using Mantle?
"Kaveri APU A10 7850K, 512SP GPU, DDR3-2133MHz vs Battlefield 4, Ultra/High/Medium setting !"
http://www.youtube.com/watch?v=axyHkKn_e80

 

[Teizo]

Could this be real? ...and is it using Mantle?
"Kaveri APU A10 7850K, 512SP GPU, DDR3-2133MHz vs Battlefield 4, Ultra/High/Medium setting !"
http://www.youtube.com/watch?v=axyHkKn_e80

Not sure, but Mantle is going to make a huge difference for AMD APUs.

 

[Paul98]

Could this be real? ...and is it using Mantle?
"Kaveri APU A10 7850K, 512SP GPU, DDR3-2133MHz vs Battlefield 4, Ultra/High/Medium setting !"
http://www.youtube.com/watch?v=axyHkKn_e80

No it's not using mantle

 

[VulgarDisplay]

Isn't Kaveri the first APU using GCN? I thought the last iteration was till on VLIW4? It's still pretty impressive what they are showing there.

 

[Teizo]

I think it is also important to note the speed of the system ram in that video due to the reports of high speed ram making BF4 run allot better. Not sure you would get the same performance with 1600mhz ram.

Hopefully Linus and Slick get their comparative memory benchmark out soon to confirm/deny those findings.

 

[VulgarDisplay]

It's pretty much common knowledge that faster memory greatly increases performance on these apu's.

These will make great HTPC's when they hit the market if they can run BF4 like this. Perfect for tiny cases with very little noise. I can't wait to play with one.

 

[jackstar7]

It's pretty much common knowledge that faster memory greatly increases performance on these apu's.

These will make great HTPC's when they hit the market if they can run BF4 like this. Perfect for tiny cases with very little noise. I can't wait to play with one.

Sounds like a good ingredient for a steam box...

 

[Arkadrel]

Could this be real? ...and is it using Mantle?
"Kaveri APU A10 7850K, 512SP GPU, DDR3-2133MHz vs Battlefield 4, Ultra/High/Medium setting !"
http://www.youtube.com/watch?v=axyHkKn_e80

Yes that is real.

No its not useing Mantle (this is directx).

Its useing a non optimised (for BF4) early driver, they mention that in the video.

Once they start optimiseing drivers I suspect performance will go up a great bit than is shown there.

 

[SiliconWars]

No need for them to optimise it for DX, the Mantle version will be available just before Kaveri is released.

 

[Nintendesert]

Not sure, but Mantle is going to make a huge difference for AMD APUs.

Ultimately not just AMD but hopefully all of PC gaming. The cheaper PC gaming can become the bigger the market is and hopefully it means more developer focus vs. consoles. That's the hope with Intel's IGP and these AMD APUs. I don't think any of us would use either, but if that means less crappy console ports all the better. I'm really interested to see what Mantle will do.

 

[3DVagabond]

AMD guys said nothing and if they know how much performance gain they would tell by now.

See this
As reported by tech report:
Andersson didn't bring up performance estimates, but other developers who discussed Mantle at APU13 did. Jorjen Katsman of Nixxes, the firm porting Thief to the PC, mentioned a reduction in API overhead from 40% with DirectX 11 to around 8% with Mantle. He added that it's "not unrealistic that you'd get 20% additional GPU performance" with Mantle.

Guys say u could get 20% performance although he did not mean at least so what mean to say he not sure but 20% is there target.

Sorry, can you provide a link please. With your reading comprehension and skill with the Written English language, we need to read it for ourselves.

I understand English isn't your native language, and I'm not trying to pick on you. You do pretty good at communicating, overall. You completely misunderstood the post you were responding to, though.