[PCWORLD] Intel approached AMD about access to Mantle

F

Final8ty

Gawd
Joined
Jun 13, 2006
Messages
1,001
PRIME1 said:
Of course AMD refused because it's not open.
Click to expand...

Stop taking things out of context.

"I know that Intel have approached us for access to the Mantle interfaces, et cetera," Huddy said. " And right now, we've said, give us a month or two, this is a closed beta, and we'll go into the 1.0 [public release] phase sometime this year, which is less than five months if you count forward from June. They have asked for access, and we will give it to them when we open this up, and we'll give it to anyone who wants to participate in this."
Click to expand...
 
It's inaccessible till it is not, promises by tech companies should be taken with a gain of salt. Time and time again tech companies are pretty unreliable with promises.
 
If anyone can gauge developer interest it would be Intel. Although I think this whole article is much ado about nothing, it does say developer interest is high enough that EVERYONE is paying attention.

Well done AMD.
 
Semantics said:
It's inaccessible till it is not, promises by tech companies should be taken with a gain of salt. Time and time again tech companies are pretty unreliable with promises.
Click to expand...

Indeed nothing is definite so lets not start twisting what was said and spreading comments as if they are.
 
Intel's quote shows that confusion surrounding what is, and what isn't, an open standard isn't limited to forum users.

AMD's quote is itself confusing, however. Do they intend to open Mantle within one or two months or within five months? The former is inclusive in the latter, but not the latter in the former.
 
Final8ty said:
Indeed nothing is definite so lets not start twisting what was said and spreading comments as if they are.
Click to expand...

Until the claim stops being a bullet point on a power point slide, it's just a claim. Nothing about mantle is open yet.

I give AMD the benefit of the doubt about access long term based on past performance, but I also question if it will be genuinely useful or beneficial to anyone but AMD. Their reticence to let anyone else look into the guts of it yet increases that apprehension.
 
wonderfield said:
AMD's quote is itself confusing, however. Do they intend to open Mantle within one or two months or within five months? The former is inclusive in the latter, but not the latter in the former.
Click to expand...
The official timeline is before the end of 2014.
He is saying giving them a few more months to work on it and they should have a specific release date ready.
 
raz-0 said:
Until the claim stops being a bullet point on a power point slide, it's just a claim. Nothing about mantle is open yet.

I give AMD the benefit of the doubt about access long term based on past performance, but I also question if it will be genuinely useful or beneficial to anyone but AMD. Their reticence to let anyone else look into the guts of it yet increases that apprehension.
Click to expand...

The point is about what they said there intention are, whether people believe them or not does not mean that its ok to twist the said intentions out of context to fit what one believes to imply they said something else.
 
Semantics said:
It's inaccessible till it is not, promises by tech companies should be taken with a gain of salt. Time and time again tech companies are pretty unreliable with promises.
Click to expand...
I would like to mail you a few grains of salt, you okay with that? :D
 
LordEC911 said:
The official timeline is before the end of 2014. He is saying giving them a few more months to work on it and they should have a specific release date ready.
Click to expand...
The comment would make more sense within the context of Intel asking for a release date, but that's not what Intel was after.

In any case, "end of 2014" is AMD parlance is "December 31, 2014". DirectX 12 should be entering into its early access period prior to that. Unless there's real momentum in Mantle on Linux, it's going to be very increasingly difficult to justify its usage in the face of an API not limited to only one vendor.
 
wonderfield said:
The comment would make more sense within the context of Intel asking for a release date, but that's not what Intel was after.
Click to expand...
Of course that isn't what Intel was after, they were after preferential treatment to get early access before any other hardware vendor, after AMD has told them multiple times they would get access the same time as everyone else.

wonderfield said:
In any case, "end of 2014" is AMD parlance is "December 31, 2014". DirectX 12 should be entering into its early access period prior to that. Unless there's real momentum in Mantle on Linux, it's going to be very increasingly difficult to justify its usage in the face of an API not limited to only one vendor.
Click to expand...
Before the end of 2014 is AMD parlance for sometime in Oct/Nov.
Too bad a majority of the devs on board don't feel the same way.
 
Mantle is a good thing. Once they get the major kinks worked out (it IS still in a beta testing phase), they say they will release it to everyone for free. I believe AMD on this because they NEED Mantle to succeed.

Mantle will reduce the CPU load in the more intense games. AMD has been moving towards APUs for some time now, with the major problem that the CPU portion has lagged behind Intel. Mantel will help them by making that lag not such a big deal to most people (most people being the vast hordes who do not read this site - and really do not care to read it).

For Mantle to succeed, it needs to be open and free to all. AMD wants Mantle to succeed and they claim it will be open and free to all. I trust AMD's self interest above all other things, so I trust they will make it open and free to all.
 
The 33% figure is way, way off...;) Basically, AMD and nVidia combined sell 100% of the discrete desktop gpus sold--Intel has no dog in that race at all.
 
Question: What good is Mantle to intel?

I mean, Mantle reduces CPU overhead, but Intel's products aren't CPU limited in any way (they're pairing EXTREMELY powerful processors with comparatively weak integrated graphics).
Also, intel putting weight behind Mantle would legitimize an API designed to help overcome a shortcoming of their competators' products, without doing much for their own products. Why do that?

I'm just not seeing why Intel would want to get involved with it, beyond taking a peek under the hood.
 
tthey still need to write graphics drivers.

Mantle is a peek into the future of DX12.
 
Unknown-One said:
Question: What good is Mantle to intel?

I mean, Mantle reduces CPU overhead, but Intel's products aren't CPU limited in any way (they're pairing EXTREMELY powerful processors with comparatively weak integrated graphics).
Also, intel putting weight behind Mantle would legitimize an API designed to help overcome a shortcoming of their competators' products, without doing much for their own products. Why do that?

I'm just not seeing why Intel would want to get involved with it, beyond taking a peek under the hood.
Click to expand...

Well clearly there is a reason why Intel is investing R&D into low level APIs.
 
LordEC911 said:
Well clearly there is a reason why Intel is investing R&D into low level APIs.
Click to expand...
Well yeah, they have to support DX12 at some point.

But I still don't see why they'd ever implement Mantle in any released product.
 
LordEC911 said:
About... "it's going to be very increasingly difficult to justify its usage in the face of an API not limited to only one vendor"
Click to expand...
You believe that developers currently using Mantle do not feel as though it will become increasingly difficult to justify as an expenditure of development time in the face of an API that, though specific to Windows platforms, should offer at least a relatively strong subset of Mantle performance and flexibility but available to a much larger subset of their customers?

Putting aside the fact that I have my doubts the majority of these developers — if any — have actually commented as such (what we're talking about here is fairly specific), do you genuinely believe this is the case? DX12's architectural similarity to Mantle doesn't exactly kill Mantle in the crib, as performance is obviously not the only determinator for API choice, but widespread hardware and software support can and, if history serves, will.

As I said, if AMD can't sufficiently capitalize on the ability to move to other platforms (read: Linux), what they will offer in the face of DX12 does not appear particularly compelling. Only at this current time are developers able to achieve with Mantle what they cannot do, or are simply unwilling to do, with other APIs.

The Mac said:
tthey still need to write graphics drivers. Mantle is a peek into the future of DX12.
Click to expand...
Who needs to? Intel? Assumably, yes. NVIDIA, however, already has working D3D12 drivers.
 
i suspect it's more for their atom line that could easily be cpu limited. but it may also benefit designs like a future iris pro.
 
Last edited:
Ever since this ubiquitous delusion about how console gaming takes advantage of this amazing "low-level bare-metal" programming and that this explains how an off-the-shelf Colecovision can out-do your Quad-Titan setup, people have been totally losing their minds with the hype.

There are no good technical reasons for all these idiotic APIs. These things come about for reasons of marketing and vendor lock-in, two things that we know are right up Intel's alley so there's no doubt they're creaming their drawers just imagining getting in on this racket. I'm sure they're coming up with their own API so they can make more crazy claims about how their GPUs stomp nvidia's or whatever. And of course since they have absolutely no software competence over there, they'd love to lift some code from Mantle. (Scary thought: referencing AMD drivers for ideas.)

It's hard to not believe that this is all leading towards an API for each vendor: again, for marketing reasons. I know many will wash their hands of it saying, "ohh it's no big deal because the engines will support them all and developers don't have to worry about it." But that doesn't change the fact that it's still going to force on the consumer lock-in at either the platform or hardware level.
 
One major advantage for Mantle that seems to always be glanced over is the fact AMD packaged the API similar to console's SDK (which AMD must have very good knowledge of by now). Having developers in an environment they know saves time and money on training.
 
WaltC said:
The 33% figure is way, way off...;) Basically, AMD and nVidia combined sell 100% of the discrete desktop gpus sold--Intel has no dog in that race at all.
Click to expand...

Laptops and ultrabooks maybe? But still, anyone who buys a PC and assumes they'll want to do a bit of gaming on it I'd think would make sure they had a discrete GPU solution. Of course, there's always people like my father in law, who buys an ultrabook and then ask me to install Assassin's Creed 4 on it...
 
wonderfield said:
You believe that developers currently using Mantle do not feel as though it will become increasingly difficult to justify as an expenditure of development time in the face of an API that, though specific to Windows platforms, should offer at least a relatively strong subset of Mantle performance and flexibility but available to a much larger subset of their customers?

Putting aside the fact that I have my doubts the majority of these developers — if any — have actually commented as such (what we're talking about here is fairly specific), do you genuinely believe this is the case? DX12's architectural similarity to Mantle doesn't exactly kill Mantle in the crib, as performance is obviously not the only determinator for API choice, but widespread hardware and software support can and, if history serves, will.

As I said, if AMD can't sufficiently capitalize on the ability to move to other platforms (read: Linux), what they will offer in the face of DX12 does not appear particularly compelling. Only at this current time are developers able to achieve with Mantle what they cannot do, or are simply unwilling to do, with other APIs.


Who needs to? Intel? Assumably, yes. NVIDIA, however, already has working D3D12 drivers.
Click to expand...

Mantle still allows them to do things that DX12 simply cannot.
The fact that Devs came to AMD to do something like Mantle shows there is demand for Mantle, enough to disregard the "widespread hardware and sofftware" deficiency for the foreseeable future.

I'm sure Nvidia's D3D12 drivers are just as good as their launch D3D10 drivers.

jwcalla said:
Ever since this ubiquitous delusion about how console gaming takes advantage of this amazing "low-level bare-metal" programming and that this explains how an off-the-shelf Colecovision can out-do your Quad-Titan setup, people have been totally losing their minds with the hype.

There are no good technical reasons for all these idiotic APIs. These things come about for reasons of marketing and vendor lock-in, two things that we know are right up Intel's alley so there's no doubt they're creaming their drawers just imagining getting in on this racket. I'm sure they're coming up with their own API so they can make more crazy claims about how their GPUs stomp nvidia's or whatever. And of course since they have absolutely no software competence over there, they'd love to lift some code from Mantle. (Scary thought: referencing AMD drivers for ideas.)

It's hard to not believe that this is all leading towards an API for each vendor: again, for marketing reasons. I know many will wash their hands of it saying, "ohh it's no big deal because the engines will support them all and developers don't have to worry about it." But that doesn't change the fact that it's still going to force on the consumer lock-in at either the platform or hardware level.
Click to expand...

Blah, blah, blah. Rant, rant, rant. Drivel, drivel, drivel.
Your post has been already been proven wrong numerous times.

Mantle is only a vendor lock-in if Nvidia/Intel choose it to be that way.
 
Last edited:
Yakk said:
One major advantage for Mantle that seems to always be glanced over is the fact AMD packaged the API similar to console's SDK (which AMD must have very good knowledge of by now). Having developers in an environment they know saves time and money on training.
Click to expand...
You're being silly mantle has jack to do with how ps4 and xbox one interact with developers.

Both Sony and Microsoft put into GNM/DirectX11 low level access because it's a console and you always have the same parts.

AMD has very little to do with consoles SDK, Sony and Microsoft control that shit extensively. There is no mantle on consoles, it would also require mantel for ps4 to be ported to freebsd as that is the system os.
 
LordEC911 said:
Mantle still allows them to do things that DX12 simply cannot.
Click to expand...
Assumably, DX12 is not the end of all non-Mantle APIs.

LordEC911 said:
The fact that Devs came to AMD to do something like Mantle shows there is demand for Mantle
Click to expand...
Yes, there was, and to some extent still is, demand for Mantle. Once again — and if you had grokked my posts merely than aimlessly blasting through them — I'm not disputing current and past demand for Mantle. Note, however, that developers allegedly came to AMD for a low-level solution prior to the DX12 announcement. Prior to any announcements for any future version of DirectX.

LordEC911 said:
I'm sure Nvidia's D3D12 drivers are just as good as their launch D3D10 drivers.
Click to expand...
They were seemingly good enough to run the early Forza port at the announcement. Whether that's only an indicator of them only having been good enough at the time for that particular demo or whether they're actually well-baked drivers I don't know.
 
wonderfield said:
Yes, there was, and to some extent still is, demand for Mantle. Once again — and if you had grokked my posts merely than aimlessly blasting through them — I'm not disputing current and past demand for Mantle. Note, however, that developers allegedly came to AMD for a low-level solution prior to the DX12 announcement. Prior to any announcements for any future version of DirectX.
Click to expand...
Talk about grokking... The information we have points to a majority of those +40 Devs are still wanting AMD to improve, develop and support Mantle once DX12 is out.
 
I'm being accused of not grokking things you've not even alluded to yet?

Right. I think we're done.
 
mt2e said:
Intel dosent want this, they just want to knowlege it brings
Click to expand...
Intel doesn't need to make a whole nother api just to get their cpu's to work in games. But intel didn't become number 1 by not doing their due diligence.
 
have you missed the "Will be open" comments?

its in CLOSED BETA.

you do know what that means right?

when 1.0 is released, and its not open, THEN your arguments might hold weight

Right now you are just being pedantic, malcontent, and argumentative.
 
mt2e said:
Intel dosent want this, they just want to knowlege it brings
Click to expand...
I don't see it. Mantle would benefit them in certain scenarios, but it's not clear to me what knowledge they could derive from Mantle they could not similarly derive for themselves (or through D3D12).

The point was made earlier that Mantle really shines on low-end CPU/higher-end GPU platforms, and that doesn't really describe what Intel does. They're kind of in the inverse position that AMD is in, with ordinarily very powerful CPUs coupled to typically weaker iGPUs. What Mantle would do for Intel is not particularly wonderful.
 
Dont forget intel doesnt use a raster model for their igpus like amd and nvidia

If i had to guess, they want to take a peek and see what all the fuss is and see if they can leverage any ideas.
 
You must log in or register to reply here.
Back
Top