Question AMD Laying off 4% of worldwide workforce to "focus more on AI"

Page 5 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

poke01

Platinum Member
Mar 8, 2022
2,581
3,409
106
what makes a design server-derived? If it’s finally moving to mobile focused does this this mean better idle power consumption?
 

DrMrLordX

Lifer
Apr 27, 2000
22,184
11,889
136
No? It's the opposite with APUs moving towards chiplets instead.

Huh. Well so far, from Zen1-Zen5, mobile designs have been all monolithic (aside from the Ranges). If they're moving away from that and towards a chiplet-based design for mobile Zen6 then that would be a bit unexpected.

If it’s finally moving to mobile focused does this this mean better idle power consumption?

Compared to current desktop designs? Maybe. Compared to current monolithic mobile parts? Nah, probably not.
 

LightningZ71

Golden Member
Mar 10, 2017
1,910
2,260
136
To keep up on the bleeding edge, they almost have to, at least for leading designs. Using the ever increasingly expensive leading nodes for their current market cost support range is going to be impossible. They need to limit that to critical parts of the processor, so chiplets it is. Part of what Strix Halo was getting the tech together to successfully do that in mobile and not just be a detuned desktop processor. Plus, the same design has to work on desktops as well.
 

LightningZ71

Golden Member
Mar 10, 2017
1,910
2,260
136
There's more that can be done while keeping the designs based on the same or highly similar logic blocks. The server targeted CCDs can be more heavily optimized for the targeted frequencies at the physical level as pushing higher frequencies requires more transistors that can be better spent elsewhere on functional logic. The finflexing can also be different for the N3 and smaller nodes. You can make design level decisions that favor SMT gains over ST gains, all while retaining the overall core design, more or less. You can even have them on completely different nodes as some are more optimized to different use cases.

Previously, AMD didn't have the complete volume to justify such a thing. Now, they've hit critical mass on DC/server volume that it can be justified. On client, they're to a point where they are moving far more mobile than previously, enough to justify different core layouts like Strix has. Desktop is an afterthought at this point.

So, going forward, it'll be server and mobile as optimization targets and desktop as a "let's see what we can get our mobile design to do" goal. There is solid precedent for making that work, just look at the Apple M series. Their desktop performance is top notch. There's no reason AMD can't make that formula work for them.
 

inquiss

Senior member
Oct 13, 2010
250
354
136
The most amd can do is to slash AVX512 for client, that's it
Not much for 'Client' title
There's loads they can do and some really good points above my post.

Targeting different frequencies. Different core mix, different use of fin options, different GMI links, different IO dies for memory interfaces, and as you said different AVX widths
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
26,389
15,513
136
The most amd can do is to slash AVX512 for client, that's it
Not much for 'Client' title
Client and server are made with the same chips, except different billing. That would be a foolish change. For both client and server.
 

LightningZ71

Golden Member
Mar 10, 2017
1,910
2,260
136
Client and server are made with the same chips, except different billing. That would be a foolish change. For both client and server.
Today... Today they're made with the same CCDs.

Not tomorrow. Server gets it's own CCDs and mobile/desktop merge and do their own thing.
 

DrMrLordX

Lifer
Apr 27, 2000
22,184
11,889
136
4% does not seem too noteworthy to me
Its not a lot, but when you compare AMD's headcount to Intel's, it's a bit concerning that all AMD can think to do is to follow the usual Harvard Business School MBA(tm) strategy of cutting jobs in the face of increasing revenue/profits. That would have to be some serious deadweight to really justify the job cuts.
 

OneEng2

Senior member
Sep 19, 2022
259
357
106
Today... Today they're made with the same CCDs.

Not tomorrow. Server gets it's own CCDs and mobile/desktop merge and do their own thing.
Possibly; however, I think it more likely that AMD will offer up SOME of the same CCD's for both. Today, desktop CCD's top out at 8. Turin Dense has its own 16 core Zen 5 CCD's. I believe that we will see a 16 core CCD for desktop with Zen 6 giving a possible 32 core, 64 thread beast. That CCD will likely be shared with server. I doubt AMD will be putting a 32 core CCD on desktop though.

I think that monolithic designs are going the way of the Coocoo bird. Just too darned expensive as cutting edge process nodes continue their march into exponential cost.
Its not a lot, but when you compare AMD's headcount to Intel's, it's a bit concerning that all AMD can think to do is to follow the usual Harvard Business School MBA(tm) strategy of cutting jobs in the face of increasing revenue/profits. That would have to be some serious deadweight to really justify the job cuts.
I have a feeling they are trimming up the waist line in anticipation of an acquisition or other large investment.
 

LightningZ71

Golden Member
Mar 10, 2017
1,910
2,260
136
On the "leaked" slide that was circulated here a while back, they were showing an expected total performance uplift for Zen6 over Zen 5 of about 10% or so. I don't think going from 16 to 32 cores on desktop would give so little.

I just don't think we'll see server CCDs on normal AM5 processors. Aside from what a few of our more trusted posters have said in the past, there's no competitive need for server CCDs with full blown AVX512 implementations on desktop, especially where they will hit a brick wall with memory throughput.
 
Reactions: Tlh97 and OneEng2

GTracing

Member
Aug 6, 2021
168
396
106
On the "leaked" slide that was circulated here a while back, they were showing an expected total performance uplift for Zen6 over Zen 5 of about 10% or so. I don't think going from 16 to 32 cores on desktop would give so little.

I just don't think we'll see server CCDs on normal AM5 processors. Aside from what a few of our more trusted posters have said in the past, there's no competitive need for server CCDs with full blown AVX512 implementations on desktop, especially where they will hit a brick wall with memory throughput.
I don't know why anybody puts any weight in that roadmap. It says Zen6 is coming in 2024. If it was ever accurate (which I doubt) it is way out of date now.
 

LightningZ71

Golden Member
Mar 10, 2017
1,910
2,260
136
I don't know why anybody puts any weight in that roadmap. It says Zen6 is coming in 2024. If it was ever accurate (which I doubt) it is way out of date now.
We don't have any reason to believe that it will be greater than the Zen 5 improvement over Zen 4, and that was famously struggling to get to 10% in most cases, requiring a windows patch that also helped Zen4, and an AGESA patch to help MT. Zen 5 was supposed to be a complete overhaul. Zen 6 is supposedly a fleshing out.
 

Gideon

Golden Member
Nov 27, 2007
1,842
4,379
136
Bad timing, should have reserved the name for X3D.
They obviously had that planned or you think they didn't know their own pipeline, less than a year from then?

It's just such a retarded thing to call L3 cache, that I'm quite certain it would have gotten the same backlash and memes.
 

ondma

Diamond Member
Mar 18, 2018
3,124
1,597
136
Didn't answer the poll, as there wasn't a response "I hate AI with a passion, and hope (futilely) that it dies an early death."
That would have been my answer.

Except for possibly the 9xxx x3D chips (obviously the best gaming chips, but IMO, a good but not great improvement from the last gen), the last generation from both AMD, and especially Intel have been a serious disappointment. With the emphasis on AI, I see this trend of lower performance improvements in non AI apps, especially gaming, continuing.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |