AMD/ATI for Xbox Next?

Ok gotcha, wrong train of thought there.

Btw, considering all these issues we are mentioning, maybe 1 of the console makers will take to the "Clouds" and the box be a cheap dummy box (maybe even 360 HW inside!) with a host of interface and output potential for future HW/peripherals. If at most 50% of console owners use a console at any one time you have the ability pack more into the HW, as well as upgrades down the road. Even pitch the console as a service instead of a product...

I know BW is an issue, as well as BW caps and latency, but it would seem this would be a way to end the hardware cycle, significantly invest in your "blockbusters" etc. As a popular consumer device they could viably have servers in all the major hubs to reduce latency issues as much as possible.

MS has a large Live install base. If they launch in 2012 would they expect more than 3M console sales? They could sell consoles in parallel to capacity and target high bandwidth early adopters first, with more mainstream consumers coming in 2014+. Obviously they would be betting on better online infrastructure in 5 years to accomodate 20M+ consumers streaming game content.

Always the possibility with an online service to toss everything on its head. Imagine a $20/mo service where games were cheap ($10--or even free) and part of the service fees were directed to devs/pubs based on play and play time. May sound odd, but if the ave. tie ratio is about 10 games per console a generation, even assuming these are all $60 purchases and no platinum, that is $600, of which a lot goes to retailers. $20/mo. over 5 years is double that.

Maybe not feasible at all, but there is an opportunity for someone to "go broke" or really do a paradigm change. There are a lot of hurdles, it will be interesting how these are tackled. For MS I bet software and services, not so much HW, is at the center.
 
Ok gotcha, wrong train of thought there.
Actually is to be seen if xenon using SOI can be implemented on TSMC 40nm process.
I never found that that vahalla rumor made sense. If we're speaking on multiple chips on the same package my opinion is different.
Btw, considering all these issues we are mentioning, maybe 1 of the console makers will take to the "Clouds" and the box be a cheap dummy box (maybe even 360 HW inside!) with a host of interface and output potential for future HW/peripherals. If at most 50% of console owners use a console at any one time you have the ability pack more into the HW, as well as upgrades down the road. Even pitch the console as a service instead of a product...
I know BW is an issue, as well as BW caps and latency, but it would seem this would be a way to end the hardware cycle, significantly invest in your "blockbusters" etc. As a popular consumer device they could viably have servers in all the major hubs to reduce latency issues as much as possible.

MS has a large Live install base. If they launch in 2012 would they expect more than 3M console sales? They could sell consoles in parallel to capacity and target high bandwidth early adopters first, with more mainstream consumers coming in 2014+. Obviously they would be betting on better online infrastructure in 5 years to accomodate 20M+ consumers streaming game content.

Always the possibility with an online service to toss everything on its head. Imagine a $20/mo service where games were cheap ($10--or even free) and part of the service fees were directed to devs/pubs based on play and play time. May sound odd, but if the ave. tie ratio is about 10 games per console a generation, even assuming these are all $60 purchases and no platinum, that is $600, of which a lot goes to retailers. $20/mo. over 5 years is double that.

Maybe not feasible at all, but there is an opportunity for someone to "go broke" or really do a paradigm change. There are a lot of hurdles, it will be interesting how these are tackled. For MS I bet software and services, not so much HW, is at the center.
It 's possible but that manufacturer would restrain consistently the market it can aim at. But I would like to see cloud rendering implemented on a "per game basis"( as running server has a cost). For example for games intended to be played only on-line it could be a possibility. It would make sense to run a significant part of an open world simulation remotely. The same could be done to solve "viewing distance", only a local part of the screen would be rendered locally while a "skirt" computed remotely would be applied to completely the screen. But I can see some serious synchronization issue that would limit the usefulness of the tech to only the people with a good Internet connection. "Online" should not become a barrier imho manufacturers should focus at getting more people on-line and in a more seamless manner. Bandwidth is not a meaningful representation of buying power, one can be rich and still live in an area where ISP connection are not good enough to support could computing/rendering.
It would be really a risky business, the less bothering implementation would be to run the world sim remotely. Say in an open word where your actions have real effects on the ongoing of the virtual world. The amount of data to get back would be minimal compared to a stream at 720p or more.
 
Question. Wouldn't MS need a x86 license from Intel in order to have any rights to a custom x86 AMD design?

I could see MS strictly buying custom chips from AMD. But wouldn't that defeat the route chosen by MS for the 360, to license the technology to avoid the circumstances that existed with intel and nvidia over the xbox1.

Licensing at least provides the option of using another vendor if you have problems with the current manufacturer. And I am under the impression that only Intel can provide a license for x86 chips to make that possible.
 
Exactly what i wassaying a page back, as far as I know AMD are in no position to sell a complete x86 ip anybody, due to their complex licensing situation with Intel.
 
Exactly what i wassaying a page back, as far as I know AMD are in no position to sell a complete x86 ip anybody, due to their complex licensing situation with Intel.

But MS isn't "buying" IPs. e.g. MS doesn't own a PPC license or all of ATI's vaaaaaaaaaaaast graphics IPs just because they were in Xenos. They own the large scale "design" and the right to shirnk it, merge it with other chips, and to utilize bc without fees (ala to NV as in the case with 360 BC).

If AMD can get away with making chips outside of GF (big if still) I see no reason why MS cannot own a "design." Infact, if AMD cannot make MS a custom chip then AMD is screwed because bulldozer is aimed at being highly scalable to meet various user needs. I can imaging the annoyance a customer would have ordering a special BD core (e.g. 12 cores with a special HTT configuration and cache setup) only for AMD to turn around and sell it to the competitors cost free in terms of design.

Btw, Bulldozer's concept sounds like a strong pitch to companies like MS (wanting flexible design characteristics, e.g. we want A# cores, B# L2 cache on each core, and C# L3 cache for the die. We Want X# memory controllers and Y# frequency" etc.
 
Actually is to be seen if xenon using SOI can be implemented on TSMC 40nm process.
I never found that that vahalla rumor made sense. If we're speaking on multiple chips on the same package my opinion is different.
Does xenon use SOI for 65 nm today?

I traced the valhalla romour back to this source:
http://blog.seattlepi.com/digitaljoystick/archives/132014.asp?from=blog_last3

It looks kind of believable if you assume Valhalla is to be implemented on 40-45 nm process. I really doubt they can keep the GPU chip(s) as it is today through another shrink.
 
But you're not getting an order for 20 or 50 mil CPU's, you're getting a CPU design contract with possibly some small royalties on a limited allotment.

Good point, I forgot MS takes care of fabbing the chip themselves.

That said, the only way imho they can even win that design contract is with the interconnect between the CPU and GPU. Otherwise, they just can't compete against with a Power7 variant unless they're willing to do it for free or close to it.

Care to elaborate on why do you think this is the case? I'm curious.
 
Any AMD CPU will have to fabbed at GF, so even then, it will help AMD's balance sheet. After all, AMD owns a large part of GF.

But yeah, the money will be in the design contract ( dunno if this bit will lead to profits) and the subsequent royalties.
 
If AMD wants to see their silicon inside of the Xbox720 they will have to build a custom CPU indeed. IIRC all the code running on the xbox is encrypted, shaders running on the GPU are not but I imagine MS will want them to be for the next gen hardware as well. So regardless of whether MS decides to go with an x86 arch this time the CPU has to be custom built.

OTOH if you are going to order 20 million CPUs I don't think AMD will have a problem building a custom part for you, just like they did for the GPU.
Are you sure of that? I've never read something about (custom hardware within the CPU) that from Ms or IBM. And I don't feel like you need that to have a secured system.
 
Oh, mmendez is right. Xbox360's security is a mighty castle, nigh impossible to breach. Ask those who did it (actually booted linux on it) and they will tell you better... :)
 
It is all relative. The Xbox was so bad I would hear technical noobs in clan matches talking about modding their Xbox1 with large HDDs with dozens of pirated games. This doesn't appear to be the norm with the 360.
 
Oh, mmendez is right. Xbox360's security is a mighty castle, nigh impossible to breach. Ask those who did it (actually booted linux on it) and they will tell you better... :)
I'm not questioning that 360 security was unbreakable or actually tough to break. I'm questioning the fact that they were a part of the CPU dedicated to its support.
I remember reading (from Granmaster actually) that Ms should have crypt the DVD player firmware for example.
 
You know I mentioned the built-in security in Xenon awhile back in this thread. ;)
Figures people ignore the mod. :devilish:

It was one of the things they mentioned at the start of the generation. Pretty sure the IBM site mentioned it too.
 
You know I mentioned the built-in security in Xenon awhile back in this thread. ;)
Figures people ignore the mod. :devilish:

It was one of the things they mentioned at the start of the generation. Pretty sure the IBM site mentioned it too.
I missed it :(
Anyway I tracked back the white paper from IBM and I can't find such a thing.
the presentation is here I'm under the impression that test and debug parts of the chip are confused for what they are not (see last paragraph).
I already heard something alike on another forum but I could not dismiss the claim but I never found anything to support it either. People think Xenon include something like "palladium" because they were contemporary but once again I found no proof.
 
But MS isn't "buying" IPs. e.g. MS doesn't own a PPC license or all of ATI's vaaaaaaaaaaaast graphics IPs just because they were in Xenos. They own the large scale "design" and the right to shirnk it, merge it with other chips, and to utilize bc without fees (ala to NV as in the case with 360 BC).

MS isn't buying the IPs, it buys licenses that provide certain rights over the technology in the 360. MS and IBM do have a licensing agreement over the cpu.

http://www.ibm.com/developerworks/power/library/pa-fpfxbox/?ca=dgr-lnxw961XBoxDesign

MS has one with ATI as well.

http://www.beyond3d.com/content/articles/4/2

If AMD can get away with making chips outside of GF (big if still) I see no reason why MS cannot own a "design." Infact, if AMD cannot make MS a custom chip then AMD is screwed because bulldozer is aimed at being highly scalable to meet various user needs. I can imaging the annoyance a customer would have ordering a special BD core (e.g. 12 cores with a special HTT configuration and cache setup) only for AMD to turn around and sell it to the competitors cost free in terms of design. Btw, Bulldozer's concept sounds like a strong pitch to companies like MS (wanting flexible design characteristics, e.g. we want A# cores, B# L2 cache on each core, and C# L3 cache for the die. We Want X# memory controllers and Y# frequency" etc.

AMD owns a Intel license for x86. Intel even threatened to pull the license over the GF spinoff. Problem is Intel was bluffing because the AMD license is a cross platform licensing deal and intel chips have AMD tech inside. But AMD can't just license out x86 tech even if its a custom build because Intel owns the overall IP.

http://contracts.corporate.findlaw.com/agreements/amd/intel.license.2001.01.01.html

3.2. Intel License to AMD. Subject to the terms and conditions of this
--------------------
Agreement, Intel hereby grants to AMD a non-exclusive,
non-transferable ***** worldwide license, without the right to sublicense, under Intel's Patents to:


If MS just wanted AMD to supply the chips, sure, but MS wants certain rights that can't be extend by AMD. MS has to get a x86 license from Intel to get the such rights as to use any fab of their choice.
 
I don't see how the next gen is going to be x86 based while the current gen isn't. Consoles don't have to worry about compatibility with 20 yearold business apps and no need to use the x86 with its extra transistors. If anything they'll have to be ppc based RISC for back compatibility. What would interest me is a gpu based on cell with more spu s and some other gpu hardware like a tesellator and rasterizer
 
I don't see how the next gen is going to be x86 based while the current gen isn't. Consoles don't have to worry about compatibility with 20 yearold business apps and no need to use the x86 with its extra transistors.
I think what everyone means when they say "x86" is a kick-ass OoOE Intel CPU. ;) I doubt the "x86" ISA itself actually takes up much space.

What would interest me is a gpu based on cell with more spu s and some other gpu hardware like a tesellator and rasterizer

Larrabee... ;)
 
Back
Top