The AMD Mantle Thread

[Skurge]

Check out @AMDFX's Tweet: https://twitter.com/AMDFX/status/426386005130633216

An amd twitter account asking followers to ask mantle questions. More hype for an imminent release?

Most people asking when will it be released.

I don't understand how twitter works. Are there answers to these questions?

Also seems like they are just drumming up hype. Doesn't look like there will be any worthwhile answers.

 

[smackababy]

Ok, like I said, you've never done software development, because you don't know what a build is.

No point in going any further down this path with you.

Lol. Okay buddy. Don't tell my employer this, they might wonder what they are paying me for then.

I don't understand how twitter works. Are there answers to these questions?

Also seems like they are just drumming up hype. Doesn't look like there will be any worthwhile answers.

Generally, twitter Q&A will have to be compiled. Otherwise, you can look at the official twitter and only view their tweets (they should be responding to the questions). If you have it show "all mentions" or some crap, it will be flooded with everyone who does the @AMD or whatever and basically worthless.

I can't view twitter at work, nor would I even want to, but I do hope some valid, technical questions get asked and some site compiles them with the answers.

This does seem like a hype generation move though, as 99.9% of twitter users aren't capable of understanding anything outside of marketing slides.

 

[Cloudfire777]

Mantle first time introduced September 26th 2013.
Date today: January 23rd.

4 months have passed.

Mantle API which was said to be easy to integrate in to games. Not even out on a single game yet.

For each day that pass, the more we know that Mantle will become Bulldozer. A bust. Marketing hype at its worst.

 

[bystander36]

Mantle first time introduced September 26th 2013.
Date today: January 23rd.

4 months have passed.

Mantle API which was said to be easy to integrate in to games. Not even out on a single game yet.

For each day that pass, the more we know that Mantle will become Bulldozer. A bust. Marketing hype at its worst.

I am not excited about the possible downsides to having an extra API to maintain, but 4 months is a VERY short time in terms of game development.

I do agree that all the marketing hype is annoying, but we still have to wait and see how it works out.

 

[smackababy]

Mantle first time introduced September 26th 2013.
Date today: January 23rd.

4 months have passed.

Mantle API which was said to be easy to integrate in to games. Not even out on a single game yet.

For each day that pass, the more we know that Mantle will become Bulldozer. A bust. Marketing hype at its worst.

To be fair, if a new API were to be released today and 4 months later there are games utilizing it, that is a pretty good turn around. The API isn't finished, or even past alpha it seems, so expecting games to just be using it is a bit of a stretch. However, if they were developing it with Dice, it makes little sense to announce it and hype it up so far out from it being ready.

I wouldn't have as much of a problem with them announcing Mantle in Sept and it being in the current state if they would stop putting out ridiculous marketing slides like "up to 45% performance increases". I'm sure in some specific test on their synthetic benchmark, it does achieve that, but let's be honest, that is atypical of overall performance.

 

[Cloudfire777]

The API itself is pretty much done isn`t it?
Its more the fact that their Mantle showcase, aka Battlefield 4, which likely have all AMDs attention with many working on it, still haven`t got the Patch that enables the API ingame. How much extra time will Mantle require to be enabled in future games?

Even if they could chop it down to 1-2 month after initial release for DX, it will work as a disadvantage for Mantle. Who would want to wait 1-2-3-4 months for a 10% increase in FPS?

 

[7stars]

I am not excited about the possible downsides to having an extra API to maintain, but 4 months is a VERY short time in terms of game development.

I do agree that all the marketing hype is annoying, but we still have to wait and see how it works out.

4 months? :biggrin:
are years that they're developing it... at least since 2011-2012...but possibly even before...
the API, if you mean a specific game ok...but i think that some devs already knew about that before September, of course the Frostbite Mantle games are in development since time ago...

 

[Erenhardt]

Even if they could chop it down to 1-2 month after initial release for DX, it will work as a disadvantage for Mantle. Who would want to wait 1-2-3-4 months for a 10% increase in FPS?

One could argue that the DX was still not released. We have early access up until they fix the game.
They worked on dx bf4 for years, yet it doesn't work properly - says alot

 

[Cloudfire777]

I wouldn't have as much of a problem with them announcing Mantle in Sept and it being in the current state if they would stop putting out ridiculous marketing slides like "up to 45% performance increases". I'm sure in some specific test on their synthetic benchmark, it does achieve that, but let's be honest, that is atypical of overall performance.

Pretty much.
I think the +45% in BF4 comes from a scenario where you use a weak APU and a strong GPU. Mantle reduce CPU overhead and removes a lot of the bottleneck from the weak APU, which makes the FPS boost pretty good because the GPU can finally operate like it should without getting chocked.

With a strong APU/i7, I don`t think we are nearly close to +45% boost.

 

[OCGuy]

Are we any closer to knowing what part of Q1 we may actually see Mantle in the wild?

 

[tential]

To be fair, if a new API were to be released today and 4 months later there are games utilizing it, that is a pretty good turn around. The API isn't finished, or even past alpha it seems, so expecting games to just be using it is a bit of a stretch. However, if they were developing it with Dice, it makes little sense to announce it and hype it up so far out from it being ready.

I wouldn't have as much of a problem with them announcing Mantle in Sept and it being in the current state if they would stop putting out ridiculous marketing slides like "up to 45% performance increases". I'm sure in some specific test on their synthetic benchmark, it does achieve that, but let's be honest, that is atypical of overall performance.

Rather I wouldn't be annoyed if they would STOP talking about release dates then missing them. AMD does this a lot. They should just STOP guessing when things are going to be released. Saying a date then not delivering just annoys consumers. I'd rather have no date, than a date that I know is a lie.

 

[VulgarDisplay]

Mantle first time introduced September 26th 2013.
Date today: January 23rd.

4 months have passed.

Mantle API which was said to be easy to integrate in to games. Not even out on a single game yet.

For each day that pass, the more we know that Mantle will become Bulldozer. A bust. Marketing hype at its worst.

I'm pretty sure in one QA they said mantle is still currently in alpha. That's why they are only allowing access to a few devs at the moment.

Also, DICE/EA weren't very forthcoming with gamers about how broken their game was prior to release. They for sure didn't share that information with AMD either. All the other mantle games we know about haven't even been released yet on DX. Mantle is supposed to make them be released 9 months ahead of schedule?

Be serious.

 

[Cloudfire777]

I`m saying that if Mantle is so easy to incorporate in to games, why haven`t anyone who made games from 2012-2013 stepped up and updated the games with a patch that enables it? It would be the perfect solution to showcase Mantle because the games are already finished and all the devs from the studio and AMDs team could work on a game that isn`t bugged like BF4.

My guess because it actually does take an incredible amount of manpower to recode it in Mantle API which they won`t do for free on a game thats 1-2 years old. Or AMD can`t pay for it. Or they have talked about it but the devs see very little gains by doing all this work instead of sticking to DX.

How many are working on BF4 bugs? How many of DICE team and AMD`s team on just Mantle? I think its a somewhat bad excuse that since BF4 is so broken, they haven`t got any time to work on Mantle. I bet they have hundreds of people just working on Mantle, but it does take a shitload of time to incorporate it in the game.

Which makes me question if Mantle is ever gonna take of if it takes so many people and so much time to just get it working. Because it shure as hell aint gonna replace DX, which will have the priority.

 

[smackababy]

Pretty much.
I think the +45% in BF4 comes from a scenario where you use a weak APU and a strong GPU. Mantle reduce CPU overhead and removes a lot of the bottleneck from the weak APU, which makes the FPS boost pretty good because the GPU can finally operate like it should without getting chocked.

With a strong APU/i7, I don`t think we are nearly close to +45% boost.

It could come from a scenario with eyeFinity 4k screens where your average FPS is like 8, and a 45% increase would make it 12 FPS, both being extremely unplayable. Who knows?

I don't think the API is anywhere close to being finished. Especially, close to being a deliverable product to the "public". AMD chose a few developers willing to use their alpha to help flesh out what real needs would be desired in an API like this, which is a good thing. How good it ends up being? Well, let's just say I am not entirely on board with any kind of software AMD releases.

 

[smackababy]

I`m saying that if Mantle is so easy to incorporate in to games, why haven`t anyone who made games from 2012-2013 stepped up and updated the games with a patch that enables it? It would be the perfect solution to showcase Mantle because the games are already finished and all the devs from the studio and AMDs team could work on a game that isn`t bugged like BF4.

My guess because it actually does take an incredible amount of manpower to recode it in Mantle API which they won`t do for free on a game thats 1-2 years old. Or AMD can`t pay for it. Or they have talked about it but the devs see very little gains by doing all this work instead of sticking to DX.

The answer to this really comes down to money, plain and simple. Once a game is released, the majority of programmers are laid off and the lucky ones switch projects. There is a very small budget dedicated to maintenance releases and sustainment. That is nowhere near enough to do a completely new build with the Mantle API.

Also, the API isn't finished. Once AMD actually releases it, we might see some companies pick it up as an afterthought, patch it in, and let people see the differences. AMD approached teams that are currently in development and gave resources to add this in, while helping make the API past alpha.

It hasn't even hit beta yet (which would mean it was close to feature complete).

 

[MutantGith]

...

It hasn't even hit beta yet (which would mean it was close to feature complete)

...
.

Right, but I guess my immediate question in response to that statement is "Why were people promising it was going to be released in December for a shipping product if it isn't even out of beta yet a month later?"

I completely agree, it probably isn't really out of what most people would call beta yet. I think that AMD, releasing the information when they did, in the way they did (as suggestive yet technically vague as possible) was calculated to maximize the impact AMD cards would have on the holiday season, potentially disproportionately to their ability otherwise. It seems relatively obvious that there might have been hints that they would have to be playing fast and loose with boost clocks, yields, and shipping board volume to retailers based on the 2xx series of cards. The recent miningg craze has only magnified that yield/volume issue. This meant that there were definitely people that were waiting for a 290 or 290x with a custom cooler that weren't going to be able to get one pre-holidays. Thos same people might decide to get something else if they couldn't get that card. Releasing a huge amount of suggestive material about Mantle made sure people who were on the fence about buying for the end of the year would have another reason to wait to make a decision, or just flat out buy AMD in order to ensure that they had access to this next, great thing. There have been a number of posts on the board here about people that were thinking of building and asked for advice about weather to hold off or grab a 290x in order to ensure they could use Mantle.

While the mean attention span on the buying public might be really low, there does come a point where the people that are likely to be able to appreciate the value of features you (over?)promised are the same people that remember that you promised them months ago. There's a tipping point in the swing of public opinion, even for the tech minded. At some point, they're really going to find themselves in a pickle, especially if they end up seriously revising the ballooned expectations many have for the API.

My two cents.

 

[Paul98]

What most said was targeting a certain part of December or January. Not that it WILL be released then, only targeted that part. It didn't surprise me to see it pushed back.

 

[OCGuy]

What most said was targeting a certain part of December or January. Not that it WILL be released then, only targeted that part. It didn't surprise me to see it pushed back.

"Most" said December because AMD said December, with no qualifier like "target."

 

[smackababy]

"Most" said December because AMD said December, with no qualifier like "target."

Yeah. When you say "In this month: this event", I generally don't think you mean "maybe in this month this will happen, if we can find it in our hearts to make it so."

 

[Paul98]

"Most" said December because AMD said December, with no qualifier like "target."

Look at the quotes form the developers

 

[moonbogg]

So February then? Just saying.

 

[OCGuy]

Look at the quotes form the developers

Did you provide some? I don't see any. I'm not sure why they wouldn't match their presentation in any case.

 

[Paul98]

Did you provide some? I don't see any. I'm not sure why they wouldn't match their presentation in any case.

http://www.hardwareluxx.com/index.p...battlefield-4-expected-for-late-december.html

 

[OCGuy]

http://www.hardwareluxx.com/index.p...battlefield-4-expected-for-late-december.html

Yes, that is when they moved the date again to "late December." Since then it has been changed to Q1. Originally the same person said likely shipping in Nov:

 

[Paul98]

But who really cares what they said, it will get here when it gets here. Considering the state BF4 was released in, and the state it still is in, I am kinda glad it has been pushed back a bit.