Things look bad at AMD

[jebo_4jc]

It's looking like AMD might be getting out of the high/mid range market completely, leaving only stuff like Lano and ARM competition.

I would be shocked, since AFAIK their server division is pretty healthy. I'd have to dig out some numbers, but I'm lazy ATM.

Crazy to think, though, that Phenom and Radeon as we know it would be going away.

Here's some of the discussion from another thread, for those who don't make it out of the DC subforum

http://mobile.theverge.com/2011/11/3/2536299/amd-lays-off-1400-employees-including-some-senior-execs

Possible faildozer fallout.

There seems to be a great deal of talk about AMD refocusing on the low-power markets. I suppose it makes good business sense seeing as how well ARM is doing, but I can't help but feel very disappointed.

That means we can wave good-bye to any reasonable competition to intel in the desktop, enthusiast and server segments.

Apparently Carrell Killebrew is one of the people chopped, a VERY big player in their GPU division and the man responsibly for Eyefinity. I really think this is terrible news, they are cutting out all the people who have helped them to this point.

GPU division actually lost money last quarter and only made ~26 million this quarter (which granted is not bad by any means).

Im assuming the new CEO just did a standard cost / effect analysis and chopped those who lost, regardless of past with the company.

http://techreport.com/discussions.x/21954

(Layoff email notice: http://pastebin.com/FGC10TSB)

According to TechReport, the biggest hit was in the public relations department. All but one person who deals with their graphics card division was removed. I am thinking that JF-AMD was laid off as well since he is the Director of Product Marketing for Server, Embedded, and FireStream products at AMD. Until he says otherwise or there is another blog on AMD from him, I think he was included unless Marketing and Public Relations department are two separate entities in AMD.

This also does not portend good news for a lot of us who wish to buy a desktop AMD CPU. If going by the email linked above and what TechReport mentions, AMD will probably devote more focus on the mobile and APU division than the desktop and server CPU division.

Key growth areas? Does that mean APU (mobile and desktop) and graphics cards?

I'm not sure how Trinity will be successful though. It is part Piledriver (Bulldozer with 10 to 15% improvement sans L3 cache), and part Radeon graphics. The Piledriver part is worrying given that recent tests and customer complaints about the Bulldozer CPU shows that the module, shared resource design is flawed in nearly every aspect. How can AMD expect that to succeed? Would it perform faster than current Llano? Or will it just barely be faster or match it? Or even slower?

They do admit in some ways that they consider their competitors a "serious and significant" threat.

Everything in the email suggests to me that we may see less funding devoted to desktop CPUs, and more focus towards APUs and possibly their graphics cards divisions.

Is this an indication that Intel may now be the more dominant provider of mid-range to high-end CPUs with what looks like AMD giving up entirely in that area? It makes me think that Intel will have more competition in the mobile and integrated CPU+GPU market, and practically none whatsoever in the mid-range to high-end market. Not like Intel had much competition in that area to begin with...

It really sounds like an admittance of defeat.

AMD: "You won Intel, we give up. We're going to focus our resources and skillsets now in other areas."

http://www.anandtech.com/show/5050/...-workforce-reduction-carrell-killebrew-is-out

Yup, Killebrew is definitely out. It'd be crazy he gets picked up by Nvidia. NV Surround implemented using a single card instead of two? I think AMD just made a big mistake, and I don't think Killebrew should be one to take blame for any losses AMD may have gotten in the past fiscal year.

As for John Fruehe, his last blog was November 1st. No word anywhere or even on his Twitter page whether he was let go as well.

Icrontic has listed more people who were let go outside of Killebrew.

http://tech.icrontic.com/news/the-axe-falls-at-amd-as-layoffs-sweep-through-the-company



I'm still surprised they laid off their Senior Engineer. With the way AMD FX was receiving poor reviews, and future products on the table, as well as upcoming 7000-series Radeon GPUs, wouldn't that be a bad idea to do?

The outlook for future AMD desktop CPUs (non-CPU+GPU integrated ones) and now their GPU products isn't looking so bright now. The email linked above and on Icrontic, and whatever is going on in Read's screwed up thinking at the moment is really pointing AMD at a different direction.

We may possibly see AMD FX be the very last mid-range or high-end CPU to be released from AMD unless someone from AMD says otherwise. This is again, sounds like an admittance of defeat from AMD. We'll probably just see a more increased focus on desktop and mobile APUs as the email alludes to, as well as other reports on the internet.

Hey guys, Brian from Icrontic here. I just wanted to let you know we've added a few names to the list that we originally published, and I'm getting emails from other former AMDers confirming their own layoffs as well.

Added to the list: John Swinimer, who also said,

Just so you know there’s only one guy left in the AMD Markham PR office to support all of AMD’s graphics PR efforts – Dave Erskine.

The landscape at AMD has dramatically changed, almost overnight.

He really just killed AMD GPU's. Sure smaller things like Trinity will be fine, but AMD is OUT of the high end graphics market after this. Its over. Expect it in the announcement on the 9th. This is amazingly sad.

 

[86 5.0L]

Intel wants AMD to stay and WILL pay AMD to keep them around. They don't want the Feds knocking on their door with an antitrust lawsuit

 

[MacLeod]

This is like the 4th thread we have on this subject here. But not one to pass on a chance to increase my post count, Ill say what Ive said on those others -

I dont think AMD is going anywhere. They make plenty of money on their server products and their GPU line so Bulldozer being a flop will hurt them but not cripple them. Theyve still got plenty of money makers to hold them over while they get this shit straightened out and back on track.

I would think if they were going to jump ship on the mid-high end market, it wouldnt just be marketing people that were getting the axe. Theyd be lopping off heads in all the engineering departments as well I would think.

This looks to me like theyre just shaking things up and getting a fresh start.

 

[86 5.0L]

Yup cutting out the fat and getting back on track since bulldozer was a flop, they are changing things up

 

[Cerulean]

Well, either way.. the future is technology, so GPUs for gaming are going to exist and continue existing one way or another. This may open doors for newer competition, or perhaps a change in gameplan but ultimately everything will be fine.

 

[systemviper]

gpu's are still rocking and now they have sudo 8core chips out there.

it's like Mac's in the dark days, they will survive!

 

[metallicafan]

Dont forget that Intel did something very similar (layoffs, cutting fat, reorganization) after the Pentium 4/Pentium D era. The result was the whole tic-toc architecture planning and structure that they have had since the Conroe era. I think people are always just a little quick to proclaim that the sky is falling. Too bad about all the laid off work force though. The economy doesn't really need that at the moment.

 

[BreezeDM]

AMD has won the next gen consoles, so they have that. Marketing isn't really needed and they haven't been able to compete with Nvidia's machine, so waste of time imo. Market with your hardware performance. Server is where they should be worried. ARM is entering with low power and lots of cores, where AMD made more sense. Now all AMD has is a server chip that is only good with highly threaded applications that have a lot of INT calculations, but power isn't that great. Only thing good about Bulldozer is the turbo core is awesome, but I don't think that will save it. They will survive, but CPU will be rough for a year or two.

 

[MacLeod]

I dont think marketing is a waste of time, at least not for desktop CPU's. The average consumer is the main buyer of prebuilt PC's and they go to Best Buy and browse thru the aisle and they see Intel and recognize it immediately from the commercials. They see AMD and theyve never heard of it. Now if you were gonna buy something, would you buy a brand you recognize or one youve never heard of?

Now thats not so say that AMD should go out and sponsor the Super Bowl or anything extreme like that but maybe a commercial or two here and there would help.

 

[BreezeDM]

AMD has always said they don't care about consumers recognizing their processor brand. They think the amount of money is not worth it. They are trying to court OEMs to use their fusion processors with the promise of a "better overall experience"

 

[MacLeod]

I know, and it hasnt worked out too well for them either.

Stop any guy at a Best Buy browsing the computer aisle and I guarantee you that the vast majority, if not all of them have never heard of AMD. There is no way you can convince me that consumers never hearing of you doesnt matter.

If Intel branding didnt matter, they wouldnt be putting INTEL INSIDE stickers all over everything.

This is even starting to creep into the cell phone market. Walk in to your local Verizon store and youll notice the smart phones are all advertising how fancy a CPU theyve got. 1 GHz or Dual Core stickers are showing up there too.

 

[Mtnduey]

crazy stuff man, AMD still rocks the socks off of computational databases (SQL), not as good in transactional databases compared to intel tho (oracle/db2). But we've had excellent results with AMD based blades in our VMware environment. And last I checked AMD is still the preferred supplier for Oracle's x86/x64 based servers.

it does beg the question, why are they doing so well in the server market and so crappy in the mid/high pc market? Traditionally speaking, the things learned in the server market gets put to good use in the desktop market, I just dont understand how they are missing the mark like they are.

 

[pjkenned]

The big thing this that their strategy with Magny Cours forward was to make tons of cores. This works in the server world where you load up a decent core with tons of RAM and HT 3.0 and can host tons of VMs and such, but it is not as good as in desktops where you cover more than 90% of users with 4 cores. Plus, AMD is way behind just introducing 32nm chips. Intel is ramping on the next node so realistically, AMD needs to do something different for the mainstream world.

That being said, Intel does need AMD to stick around. Odds are the competition committee is looking how to keep them in the 15-20%+ market share range.

 

[jebo_4jc]

The big thing this that their strategy with Magny Cours forward was to make tons of cores. This works in the server world where you load up a decent core with tons of RAM and HT 3.0 and can host tons of VMs and such

I'm not sure they didn't screw something up though along the way. Performance with 8 "cores" at 4ghz is OK, but it's not where it should be. I'm afraid Interlagos is going to disappoint. Something tells me a 16 core Interlagos will not be 25% faster than a 12 core magny cours.

 

[Zink]

I'm not sure they didn't screw something up though along the way. Performance with 8 "cores" at 4ghz is OK, but it's not where it should be. I'm afraid Interlagos is going to disappoint. Something tells me a 16 core Interlagos will not be 25% faster than a 12 core magny cours.

This is 32nm so there is some advantage there but it really depends on how the chips handle low voltages. From what I've seen power consumption at stock isn't too bad but it sucks at high voltage. If we take this trend and go the other way maybe they can hit 3Ghz on really low voltage on the best binned Opteron chips. If power consumption could be kept in check on a 16 core 3Ghz chip then they might be able to hit 25%. That's what I hope at least, though I don't see myself every being able to afford G34.