Some Thoughts on the Impact of AMD's Layoffs on the Future of AMD GPUs
by Anand Lal Shimpi on November 4, 2011 8:18 PM ESTHere's something I'm concerned about. AMD's Carrell Killebrew, part of yesterday's announced layoffs, was a Director of Product Planning for AMD's GPU division. His job, at least as he explained it to me so many times in the past, was to figure out what the next 3 - 5 years of AMD GPUs were going to look like. He's still technically with AMD today, although that will change in the not too distant future.
Carrell is a dedicated guy, he works hard and generally seems to know the right move for AMD in the GPU space. His track record as of late is a good one. The verdict isn't out on AMD's 28nm GPUs, but the last three years of AMD GPU releases have been excellent. They've been competitive and well executed.
When reducing workforce to cut costs, you don't go after your product planners - unless their vision and your vision don't line up. We all know what Carrell wanted for the future of AMD GPUs (as I wrote before, he wanted to deliver a first generation "Holodeck" by 2016), but what does AMD's new CEO want that conflicts with this goal?
Carrell's vision saw the continued growth of the high-end GPU. On November 9th we're supposed to hear more about Rory Read's new strategy for AMD. I am concerned that it may conflict with Carrell's vision. Maybe I'm reading too much into all of this. What do you all think?
Facebook
Twitter
RSS
48 Comments
View All Comments
rtfg - Saturday, November 5, 2011 - link
== ( http://www.hapous.com )====== ( http://www.hapous.com )====
== ( http://www.hapous.com )====
Casper42 - Saturday, November 5, 2011 - link
One of the guys laid off is my local AMD Rep who deals mostly with Server chips.20 year AMD Vet and they laid him off 2 weeks before Interlagos was released to consumers?
YOU SO STUPID
B.Wallach - Saturday, November 5, 2011 - link
AMD has been making up the costs of the fabs it ditched and is actually doing well, this layoff is more personal that practical and I think they are heading in another direction that may be we see the last of them in the main stream cpu/gpu world.If this happens we can all sit nicely and cut our own throats, the ones that is that like cheap cpu's and gpus.
I thought their mix of the two was a great direction if they kept it up to where they merge middle class cpu's with good gpu's because the both offest the other's weakness's.
If the merged the Bulldozer with the next gen GPU they would have one killer product if the software was there to take advantage of it. In non gpu apps it could double as the FPU where the BD is weak and use the cpu to help out with some of the physic the gpu games need. This is just fast guessing but they had a good direction and are making money and this layoff isn't about profits and I am really regretting they changed CEO's. At least Myer's was dedicated full time to AMD and I really wonder just where the new ones dedications really are grounded at.
If we loose AMD cpu's (which the hardware sites like this one and others have done their best to talk people out of buying them before software has been out to show just how well they can really do, too much early reviews that don't mesh with that new of a cpu structure) and no one around seems to remember the days when Intel was alone in the 286 cpu days then you will get a taste of it if AMD moves out of this market because this is a fact you WILL be paying as much for a cpu as your paying for you whole computer and this isn't counting what Nvidia charges for discrete GPU cards if ATI goes down with AMD. Computer's used to be pretty expensive before AMD and Cyrix came on the market and most people ran Atari and Commador and Apple computers. Paying 2,000 or more for a 286 was common and even not counting inflation it's more than a good one costs today. Wave bye bye to cheap computers.
I'll forget about using computers or use one for a very very very long time because that one will be the last one affordable to buy.
AMD needs to sell good server CPU's and good high end graphics cards for the high end, not our high end but cad servers and higher as there is big money in those area's.
They have been falling short in the high end GPU range lately maybe thanks to the merger with AMD. I wish they never would have merged because now they are all in one target range to really screw things up if they fail.
Oh well, it was fun for a while.
It is way to obvious they can not keep up with Intel and they know they can't and were trying for the mid range market because a R&D budget is very high and getting a lot higher as the size shrinks. AMD has worked with IBM and others to offset this somewhat but will never be able to match Intel so don't expect their cpu's to come close to their top of the line cpus but more in the mid range where with a good production run they can do ok and focus on the long term merge of the cpu/gpu market where the future for them is bright for them and the main stream users. I just hope they have the time and this new CEO doesn't really screw things up like it sure sounds like he is doing.
If they can't sell enough or make enough to make their R&D budget the game is over.
softdrinkviking - Sunday, November 6, 2011 - link
and scary that they would let him go.it seems like AMD is going to blaze a trail into budget chips and fade away into irrelevancy.
i wonder what's about to happen to AMD's stock value? could this be a good thing for investors?
Aries1470 - Sunday, November 6, 2011 - link
they may not foade in to irrelevancy, but might go the way of VIA... they still do x86 but not much is heard about them. They now do low power cpu's... for embedded systems mostly.I just hope they can do what they were suppose to do.
Ditch the automation, do a cycle every 18-24 months with high specced Cpu's, not just their FPU's of which use to be great, and then start competing with Intel again.
They should not have sold their ARM part off... but hey, if they really want to, they can get a licence and start again, and do their own design. nVidia is doing that already.
Hindsight is great... but getting in to the game, and having the guts for tough decisions is even more tougher especially when you have to stay afloat.
Personally, I would think of diversification... try and get most markets. They already have a good portfolio of products, they jsut need to optimise that portfolio.
Have at least 3-5 items going, with 2 of them in the front and the other in the back with a slightly more "slacker" time frame, so they can still be in the game, and be merging them every 1.5-3 cycles.
An example would be:
Bulldozer + 7xxx series, merge the mid-series, every 6-8 months, create a new skew with the same or slightly more optimized version of the CPU with an updated graphics core and adding a suffix.
When the GPU is not needed for "gaming" or "boinc" or some other high demand, then to power gate / throttle down and jsut use a portion of its available power. Same as with the CPU. When you browse or use social networking with flash available, you do not need the full power, at this power state, the APU (CPU+GPU) can go in to a very low state, and have a power usage of about 15-30 Watts, then when you go in to another task, rev up to the 100+ Watts.
SouthBridge etc:
Make it the most power efficient it can be.
Make the Skew's scalable, all M/B's do not need 12-20 USB ports!
Chop down lanes, not features! i.e.:
Keep Sata, but not all applicaitons need 6+ ports, or all the usb ports or even PCIe / PCI lanes!
Example:
Low End Mid High End
PCIe* 8/4 16/8 32/8
USB** 2/4 4/8 6/12
SATA***4/0 6/2 8/4
PCIe = Video/ Remainder for other ports like a 4 lane PCIe port or 2 * 1x ports etc. High End can be 2 * 16x lanes + 1 * 4x + 2 * 1x lanes
USB = 3.0 / 2.0. The 3.0 chips are still more expensive to make and also for routing on the board.
SATA = Int / Ext
Backburner:
SoC = Make server on Chip.
SoC = Also System on Chip for Mobile / Tablet. Nearly everything is on the chip, including LPDDR2!!! This is becoming common practise for ARM. This has EVERYTHING ON IT. Just peripherals are added. Ideal for "closed" systems as phones and tablets. no SB needed.
Anyway, these are just my thoughts.
jabber - Sunday, November 6, 2011 - link
It's just not going our way.Give it another 2-3 years and most folks will be doing their computing on low power, light devices and probably mobile to a degree too.
10 years ago everyone wanted desktops because laptops were a compromise (slow/missing features and ports) and they cost far more.
Now none of my customers want desktops. Its laptops, laptops and more laptops. Yeah I still call them laptops, big whoop wanna fight about it?
Now I'm getting asked about tablets and smartphones. It's all about getting smaller, lighter for normal folks, not bigger and faster.
to 90% of the computing public Bulldozer and anything over an i5 is pretty pointless.
Once we get smartphones and tablets up to a 2GHz dual core platform with a GPU built in (and it will happen) then that's when AMD/Intel need to worry. Especially if they haven't got anything in the cupboard that fits that bill.
What we have to remember is that the world of computing no longer hangs on the word of the PC enthusiast anymore (if it ever did). We are largely irrelevant to the guys at AMD/Intel etc.
As long and mom and pop can browse the web and look at their photos and Joe Corporate can work on his spreadsheet in a hotel room or in a airport lounge that's as far as it needs to go.
It seems pretty obvious to me from the actions of game developers that if you want to play games then get a console.
I don't like it either but then I guess the blacksmiths and coach-makers didn't like it when the automobile came along.
MadAd - Sunday, November 6, 2011 - link
what we need is another killer app for the pc, one that you just cant do on wind up toy pc tabletsdfsd - Sunday, November 6, 2011 - link
==== Something unexpected surprise ====[ http://www.hapous.com ]
very good web,believe you will love it.
fdfsxcvdcfdh - Monday, November 7, 2011 - link
yegrt7eytg8rfdfsxcvdcfdh - Monday, November 7, 2011 - link
eghfydhgvt