The AMD Execution Thread [2007 - 2017]

Status
Not open for further replies.
I wouldn't be suprised if they decide to alter their core business and get rid of something large. For example the discrete GPU business in general seems to be spiraling down.

The graphics division is profitable, and should get better with increasing market share in pro markets (has already gone from 9% to 18% in a few months, according to Charlie). After announcing a new architecture clearly aiming for better compute performance (GCN) getting rid of this division would be foolish.

Mainstream APUs (Brazos and Llano) bring in most of the revenue, and server processors bring in a large part of their profit. Plus, Bulldozer was obviously designed with servers in mind. So it would be quite dumb to ditch either.

There's really nothing to get rid off; nothing I can think of, anyway. If anything, it looks like AMD is trying to expand to new, more mobile markets.
 
I wouldn't be suprised if they decide to alter their core business and get rid of something large. For example the discrete GPU business in general seems to be spiraling down.

So basically keep the division that is losing you money. CPU.

And jettison the division that has consistently made you profits for the last year+? GPU.

Regards,
SB
 
So basically keep the division that is losing you money. CPU.

They have the x86 license, which in the long run is probably worth quite much so I think they will keep their CPU business as is. But regarding the discrete GPUs, it may take few years but eventually, I think, the APUs will completely replace the discrete GPUs. To be more exact, I think the consumer discrete GPUs will mostly vanish, like previously has happened with sound cards.

But on the professional side of the fence the discrete GPUs will still have their place, but aimed more to general computing.
 
Anyone know if there was any impact on the OpenCL team at AMD? Hopefully they weren't impacted by the massive layoffs :(

As far as I know (may not be very far), this cull is primarily about cleaning out quite a bit of...err...redundancy (see who's getting cut). Their SW devs aren't really subject to that IMHO, they're under-staffed enough as it is. It would appear that the important people are left untouched.
 
I wouldn't say Carrell Killebrew was fluff. As some on other forums have pointed out, without his extensive input we would not have the move from monolithic dies to the "sweet spot" strategy that worked so well for the 4xxx range.

How is OpenGL Guy, Wavey and gang doing? I have heard the driver guys, marketing and PR were hit with the latter two departments being hit hardest of all.
 
I think, the APUs will completely replace the discrete GPUs. To be more exact, I think the consumer discrete GPUs will mostly vanish, like previously has happened with sound cards.
I can see them becoming de-facto standard in non-gaming PCs but anyone that does any serious gaming will quite definitely still have a half-decent discrete GPU. Sound cards dissappeared because the computation requiremetns it needed was easy to provide in a CPU + a simple chipset. When will integrated GPUs be good enough for gaming? Consider that we'll be getting next generation of consoles soon so we won't be able to play the latest console ports that are basically optimized to run on 5y old hardware.
 
Why would they hit the driver team hard?
AMD's perennial problem has been software, a problem that has only worsened with the fragility of designs like Bulldozer.
The GPU section's investment in drivers that trickled down to Zacate and Llano. Llano especially would have been utterly uncompelling without it.
Do they think Llano's weak-ass cores would sell with drivers and performance that matched Sandy Bridge's GPU?

AMD's x86 side already knows what happens when you cede leadership in a market in terms of mindshare and the endless jerking around trying to divine the goals of a competitor (BD again, or even why OpenCL resembles CUDA so much).
Are they going to water down the GPU side and then wonder why nobody makes a game that caters to their graphics (less so than the rather poor showing they already have), or why none of the standards seems to be moving their way anymore in a few years?
 
It does sound crazy. But I've read that mostly PR and marketing were affected. Have the driver guys really been significantly touched?
 
If they went after the driver guys then they are insane. Unless they are replacing some b/c they are upset with their results and think someone else can do a better job. If they are just cutting then...
 
While I would certainly agree that the engineering departments are the lifeblood for any technology company, and that losses there would be even more detrimental to AMD on the whole, the losses among PR and marketing staff are clearly much more than just a healthy cleaning up of dead wood.

I don't think it is unfair to say that AMD needs either more of all of these, or better ones. Because regardless of whether it is in their product engineering, platform development, developer and press relations, design wins, retail presence, or a plausible coherent vision for the post-PC future - I can see few bright spots for the company at the moment.
 
Well, GCN will be easier to write drivers for. And if the VLIW4/5 drivers are already in a "good enough" state... But I certainly hope hey didn't touch that department.
 
We've been hearing time and again the driver compiler for the VLIW architecture was good enough, as well as the same attribute - being easy to make drivers for - was cast upon Nvidias architecture since G80. Somehow I doubt that it's going to be much easier in the future.
 
With all the new applications for and technologies and features involved in GPUs (everything compute-related, Eyefinity, video decoding, tessellation, power management, all the new anti-aliasing techniques, etc.) sometimes I think it's a miracle anything even works.
 
I wouldn't say Carrell Killebrew was fluff. As some on other forums have pointed out, without his extensive input we would not have the move from monolithic dies to the "sweet spot" strategy that worked so well for the 4xxx range.

How is OpenGL Guy, Wavey and gang doing? I have heard the driver guys, marketing and PR were hit with the latter two departments being hit hardest of all.

Most of this mythos is based on a piece by Anand. Myths are nice, but they tend to be...err...creative. Whilst I am convinced he is a nice guy and all, and truly wish him all the best, the Internet implosion happening over his dismissal is just the Internet being itself over a hero it had created. Mr. Killebrew wasn't lead architect for anything major coming out of ATI (no, Eyefinity does not count, and even there he wasn't lead architect or an important designer beyond generally defining some functionality, AFAIR).
 
I really didnt like that Anandtech 4850 editorial. These products are the work of many teams not just a few managers. That piece made a few managers sound like superheroes.
 
Wokring for a large corporation where an entire areas have been downsized, removed or outsourced I know how de-moralising it can be for the ones that remain. Here is hoping to the well being of the entity known as AMD - it certainly is now entering a completely new chapter. I hope they do well and rebound - I think the target of the board is still to increase shareholder value.

When Dirk was unceremoniously released from his position AMD stock was at around $9, even accounting for the recent stock market hits AMD is still not heading in the right direction as perceived by Wall Street et al.
 
It does sound crazy. But I've read that mostly PR and marketing were affected. Have the driver guys really been significantly touched?
No one outside of the executives and HR know the complete picture of who was let go, but it wasn't all marketing and PR. People know about those groups because they talk to the press.
 
Status
Not open for further replies.
Back
Top