AMD's Future Chips & SoC's: News, Info & Rumours.

[InvalidError]

First: a failed Ryzen launch would have been AMD's failing, not Intel's succeeding with whatever their 10nm strategy actually was. Completely different.

Intel successfully executing on 10nm would have finished AMD off.

Like I said, AMD's patent portfolio and x86 licensing rights make them way too sweet a property to overlook. Or to simply take apart: the licensing rights alone make them far more valuable as a whole than in pieces.

AMD's x86 license from Intel is non-transferable, so it would be worthless to any would-be buyers.

 

[drea.drechsler]

....

AMD's x86 license from Intel is non-transferable, so it would be worthless to any would-be buyers.

That's right, and the reason Intel executing on 10nm could not possibly have finished them off. There's just to many investors willing to line up for a piece of the only competition in town for the very lucrative x86 market space which Intel had been milking for so long. With that license, and AMD's patent portfolio, all you really need is a business plan to scarf up the money to begin development of a CPU. But AMD had Ryzen in the lab...and a real CEO with a vision...and Jim Keller in engineering; a solid foundation for the business plan.

Considering how good Ryzen actually was, and is, you'd have to concede they'd also have a dandy of a presentation demo to clinch the deal too. Intel executing on 10nm, however, turned out to be pure fantasy and marketing, their go-to whenever they get in a pinch.

Now why...or how... did they screw it up. That's still the question.

 

[InvalidError]

There's just to many investors willing to line up for a piece of the only competition in town for the very lucrative x86 market space which Intel had been milking for so long.

VIA, NatSemi and a few others have x86 licenses too and they chose to bail out of the x86 space. Investors aren't particularly eager to pour money into companies just because they have an x86 license and a few related patents.

 

[drea.drechsler]

VIA, NatSemi and a few others have x86 licenses too and they chose to bail out of the x86 space. Investors aren't particularly eager to pour money into companies just because they have an x86 license and a few related patents.

I believe VIA's x86 cross-licensing agreements, which they got along with an acquisition, expired sometime in the early 2010's. Whatever they do now uses clean-sheet alternate design approaches. Don't keep up with it enough to know for certain. I'm not familiar with Nat Semi but they've long expired and I'd have to imagine any agreements along with them.

But more important is the reasons for both of them exiting the x86 market space which had far more to do with Intel's well known illegal anti-competitive and monopolistic business practices. The same tactics that hurt AMD to the point that a mis-step similar to any of the ones Intel's made several times in the past became nearly impossible to recover from (in addition to some pi55 poor management by a former CEO.)

AMD's licenses and patents alone wouldn't be reason to invest, sure. That's where the "business plan" part comes in; an actual product, a strategy to bring it to market and a team of talented individuals that can make it happen. The business plan provides the will and the product, along with all necessary licenses and patents that erase impediments, provide the means. You gotta have both.

Oh yeah, and Intel's under court orders on three continents, now, in an attempt to reign in their anti-competitive tactics. That would have to have been the subject of a really interesting slide or two in AMD's pitches to their white knight investors back in the day.

 

[gamerk316]

First: a failed Ryzen launch would have been AMD's failing, not Intel's succeeding with whatever their 10nm strategy actually was. Completely different.

Second, even if Ryzen wasn't as well received as it was it would have just made AMD a peachy takeover property at worst. There would have been a buyer, maybe one of the majority stake holders. Like I said, AMD's patent portfolio and x86 licensing rights make them way too sweet a property to overlook. Or to simply take apart: the licensing rights alone make them far more valuable as a whole than in pieces.

They would have lived on in one way or another.

I'm also talking prior to Ryzen...bulldozer/excavator. But even after Ryzen proved a success and IF 10nm was all Intel wanted it to be the same crowd that hung in before would still hang in there. The performance alone wasn't enough before...the performance wouldn't have been enough after.

And besides, this is all purely hypothetical. Intel failed, AMD succeded. That's indisputable now. I'm just curious why Intel failed.

I'm pretty sure we concluded a few years back when we were debating the state of AMD that it's license was non-transferrable if they ever got totally purchased; the whole licensing thing between Intel/AMD is a huge can of worms.

But yeah, people forget that prior to Ryzen launching, AMD had debts about to come due in excess of the companies total value, and were loosing an uncomfortable amount of money to boot. Ryzen was legit AMDs last shot, and to their credit they absolutely nailed it. Even then, their financial situation still isn't what I'd consider "good" [they still have a lot of outstanding debt], but at least they're out of "immanent bankruptcy" territory.

As Invalid pointed out, if Intel's bet with 10nm had paid off we'd be seeing quite the competitive CPU market right now. But it didn't, and combine that with the fact Intel really can't squeeze more out of the -Bridge architecture Intel's in a bit of a tight spot. At this stage, once they "finally" get their fabs up to snuff they should seriously start designing something new.

 

[InvalidError]

Intel really can't squeeze more out of the -Bridge architecture

Haswell, Broadwell and Skylake were new architectures too. Skylake is the one Intel got stalled on while trying to sort out its 10nm and failing to do so to the point of now having to back-port Ice Lake and Xe to 14nm for Rocket Lake.

 

[drea.drechsler]

...they'd might as well go after VIA instead [who still has a valid x86 license as far as I'm aware].

But yeah, people forget that prior to Ryzen launching, AMD had debts about to come due in excess of the companies total value, and were loosing an uncomfortable amount of money to boot. Ryzen was legit AMDs last shot, and to their credit they absolutely nailed it. Even then, their financial situation still isn't what I'd consider "good" [they still have a lot of outstanding debt], but at least they're out of "immanent bankruptcy" territory.

As Invalid pointed out, if Intel's bet with 10nm had paid off we'd be seeing quite the competitive CPU market right now. But it didn't, and combine that with the fact Intel really can't squeeze more out of the -Bridge architecture Intel's in a bit of a tight spot. At this stage, once they "finally" get their fabs up to snuff they should seriously start designing something new.

According to all I've read, VIA's license was term limited at the time acquired and that term expired in the early 2010's. Unless they got Intel to renew it...but why would they?

I'm not doubting that if Ryzen hadn't clicked AMD would be a vastly different company. How it remains and operates is something for others to talk about but I do find it hard to believe their console business and the only other viable GPU mfr. could be allowed to completely dissapear all at once. Did you, and the other internet guru's, discuss whether AMD gets to retain their licenses when placed in receivership? And what happens to the AMD licenses (it's actually cross-licenses) Intel also need?

BTW:
"Devinder Kumar, chief financial officer of Advanced Micro Devices, said in a statement last week that the company could enter into joint ventures, mergers or acquisitions (M&A) agreements without fearing of termination its cross-license pact with Intel Corp. "

Key word is "merger". Seems to me there's ample room to construct a way for AMD to continue making x86 CPU's even after receivership.

I only maintain that Intel's 10nm succeeding on time would not alone have pushed them into failure. The x86 marketspace is too lucrative for investors to ignore an opportunity and AMD, even as debt loaded as it was and is, are the only opportunity in town. Ryzen made it easier.

 

[Eximo]

Samsung is moving onto Gate All Around for their 3nm solution. TSMC looks like they are sticking to FinFET through '3nm'. Intel supposedly has a department for Gate All Around, but no node announcements. I imagine they will be pretty secretive so as not to spoil their stock price and not make any more bold claims they can't follow through on.

Do you think you could cause aneurysms by walking up to Intel engineers and just repeating Skylake endlessly?

 

[BogdanH]

About Intel's troubles with 10nm and below... I would say, nobody here has complete insight. And so, we're all allowed to guess and make conclusions from what we can read in media.

I wouldn't be surprised if part of the reason why Intel took extra risks with process development was to give AMD a chance to survive so Intel could avoid becoming a strictly regulated monopoly after AMD's bankruptcy.

-that is HIGHLY unbelievable.
If that would be the case, then Intel would start showing his "actual" progress by now -after certain reputation was lost, Apple decided to ditch Intel, few CEO's were replaced, etc.
Of course, it's important for Intel that competition is alive (avoiding monopoly), but if AMD would start to have problems (at that time), there are "other ways" how to "help" AMD to start breathing again -I'm quite sure Intel knows that.

My impression is, Intel made few strategic mistakes in past (maybe they invested more into keeping stock holders happy, than into technology development). Time goes very fast when talking about technology and I think Intel was caught at taking a nap, while others were working.
Don't get me wrong, it's not that I would think it's over with Intel -far from that!! I just think Intel will remember that and learn a lesson.

Just sharing my thoughts.

 

[Order 66]

AM5 has launched. with the 7950x with 16 cores and 32 threads. The 7800x3d is currently (arguably) the fastest gaming CPU. I am trying to start a much-needed discussion (on this thread) on AM5 since AM5 thoughts need to be shared.

 

[COLGeek]

AM5 has launched. with the 7950x with 16 cores and 32 threads. The 7800x3d is currently (arguably) the fastest gaming CPU.

So the question is, "What is next?".

 

[Order 66]

So the question is, "What is next?".

8000 series, unless AMD decides to change their naming scheme (which I really hope they don't) I am really hoping that all the 8000 series have base clocks of 5+GHz and boosts of 6GHz+. I think that is feasible considering the current 7600x is only 300MHz away from a 5GHz base clock. Heck, even the 7950x has a base clock of 4.5 GHz on all 16 cores.

 

[COLGeek]

8000 series, unless AMD decides to change their naming scheme (which I really hope they don't) I am really hoping that all the 8000 series have base clocks of 5+GHz and boosts of 6GHz+. I think that is feasible considering the current 7600x is only 300MHz away from a 5GHz base clock. Heck, even the 7950x has a base clock of 4.5 GHz on all 16 cores.

Roadmaps can be insightful:

AMD Confirms AM5 Support For Ryzen 8000 Processors, Zen 5 with Navi 3.5

www.anandtech.com

AMD Confirms AM5 Socket Will Extend to 2026, Zen 5 to Use RNDA 3.5 GPUs

A new roadmap slide from AMD contains several intriguing updates to its future CPU and GPU plans.

www.extremetech.com

AMD's Zen CPU roadmap leaks - Zen 5 and Zen 6 designs detailed - OC3D

AMD x86 CPU roadmap has leaked, detailing the company's upcoming Zen 5 and Zen 6 CPU core designs. Expect big things from AMD!

overclock3d.net

AMD Desktop Roadmap: Ryzen 8000 "Granite Ridge" CPUs In Late 2024 & Threadripper "Shimada Peak" In 2025

AMD's next-gen desktop CPU families including Ryzen 8000 "Granite Ridge" & Threadripper "Shimada Peak" along with their release dates.

wccftech.com

 

[COLGeek]

AM4 lives on, as well, it seems:

News - Purported AMD Ryzen 7 5700X3D and Ryzen 5 5500X3D Specs Leaked

A leaker has tipped the impending arrival of the AMD Ryzen 7 5700X3D and Ryzen 5 5500X3D, as well as the key specs of these 96MB L3 cache boosted processors. Purported AMD Ryzen 7 5700X3D and Ryzen 5 5500X3D Specs Leaked : Read more

forums.tomshardware.com

 

[Order 66]

Roadmaps can be insightful:

AMD Confirms AM5 Support For Ryzen 8000 Processors, Zen 5 with Navi 3.5

www.anandtech.com

AMD Confirms AM5 Socket Will Extend to 2026, Zen 5 to Use RNDA 3.5 GPUs

A new roadmap slide from AMD contains several intriguing updates to its future CPU and GPU plans.

www.extremetech.com

AMD's Zen CPU roadmap leaks - Zen 5 and Zen 6 designs detailed - OC3D

AMD x86 CPU roadmap has leaked, detailing the company's upcoming Zen 5 and Zen 6 CPU core designs. Expect big things from AMD!

overclock3d.net

AMD Desktop Roadmap: Ryzen 8000 "Granite Ridge" CPUs In Late 2024 & Threadripper "Shimada Peak" In 2025

AMD's next-gen desktop CPU families including Ryzen 8000 "Granite Ridge" & Threadripper "Shimada Peak" along with their release dates.

wccftech.com

Very interesting, I didn't even know die stacking was a thing. My question is whether die stacking will limit clock speed like the 3d cache does now.

 

[Order 66]

Roadmaps can be insightful:

AMD Confirms AM5 Support For Ryzen 8000 Processors, Zen 5 with Navi 3.5

www.anandtech.com

AMD Confirms AM5 Socket Will Extend to 2026, Zen 5 to Use RNDA 3.5 GPUs

A new roadmap slide from AMD contains several intriguing updates to its future CPU and GPU plans.

www.extremetech.com

AMD's Zen CPU roadmap leaks - Zen 5 and Zen 6 designs detailed - OC3D

AMD x86 CPU roadmap has leaked, detailing the company's upcoming Zen 5 and Zen 6 CPU core designs. Expect big things from AMD!

overclock3d.net

AMD Desktop Roadmap: Ryzen 8000 "Granite Ridge" CPUs In Late 2024 & Threadripper "Shimada Peak" In 2025

AMD's next-gen desktop CPU families including Ryzen 8000 "Granite Ridge" & Threadripper "Shimada Peak" along with their release dates.

wccftech.com

how exactly does die stacking work? do you think it is possible to have die stacking and 3d cache on the same CPU?

 

[COLGeek]

how exactly does die stacking work? do you think it is possible to have die stacking and 3d cache on the same CPU?

See here: https://www.amd.com/en/technologies/3d-v-cache

 

[Order 66]

See here: https://www.amd.com/en/technologies/3d-v-cache

Yes, I know what 3d vcache is, But I was wondering how die stacking and 3d cache would work on a zen 4 CPU. The diagram on the AMD site shows the zen 3 version.

 

[Order 66]

I think it is interesting that zen 5 will have RDNA 3.5 graphics. These could be legitimately good for budget gaming rigs without a discrete GPU. That is assuming you can either build a rig with one for $300 or the RDNA 3.5 graphics provide more performance than something like an RX 580 build at a $500 price point. I put together a part list with a 580 and 5600 for $514. hopefully a CPU with RDNA 3.5 graphics will outperform this at the same or cheaper price point.