He was not "responsible" for Trinity, let's not recycle ideas thrown out into the air in sensationalistic purposes. Which is not to say he wasn't a good guy and someone you may want to keep around.
The AMD Execution Thread [2007 - 2017]
- Thread starter overclocked_enthusiasm
- Start date
-
- Tags
- amd
- Status
He was not "responsible" for Trinity, let's not recycle ideas thrown out into the air in sensationalistic purposes. Which is not to say he wasn't a good guy and someone you may want to keep around.
So, Bob Feldstein just left for the green(er) pastures.
AMD has a serious problem with retaining talent. Corporations keep key people with two things -- boatloads of money, and stock. No-one, not even employees, place all that much value on AMD stock anymore, and they just don't have the cash to keep people. So now whenever someone really distinguishes himself at AMD, he gets an offer he doesn't want to refuse.
AMD has a serious problem with retaining talent. Corporations keep key people with two things -- boatloads of money, and stock. No-one, not even employees, place all that much value on AMD stock anymore, and they just don't have the cash to keep people. So now whenever someone really distinguishes himself at AMD, he gets an offer he doesn't want to refuse.
How are you sure that those people were cut? They might have chosen to leave on their own.
DK
I've been trying to get a handle on the scope of Keller's contributions to AMD, particularly with regards to the K7 and K8.
He was employed at AMD from 98 to 99.
It sounds like he could have been involved in the bringup of K7, although the design would have probably been quite final by 98, being that it was released in 99.
It sounds like he had a role in the early parts of the K8 development cycle, but even if he were working on K8 from day one he would have been gone for the majority of its development.
If we consider the lack of a competitive x86 core a significant factor in AMD's woes, Keller is a remnant of a happier time, but he doesn't have the bona fides of the standard bearer of a new Athlon. Athlon was mostly done before him, and K8's core did not hop far from K7.
His most recent experience doesn't look to turn around the weak core problem, and he'd have to contend with the fact that just as he has probably changed in over a decade, so has AMD.
It's not like he's getting the hit band together again, not unless AMD is also rushing to hire back the engineers and designers Ruize and Meyer lost or removed since then.
That said, his SOC and networking experience is probably more relevant to AMD in the few niches it still lingers in. That's not strong comfort for the desktop market. For servers, his most recent experience screams more Seamicro board than powerful Xeon blade.
He does have experience in working to put together disparate system elements and putting together protocol. This might be good for HSA, although it would be more likely that AMD didn't just hire the guy to put together the initiative it's already announced and outlined. The heterogenous elements are about the only thing left to put together, as we are a far cry from the days when the IMC, PCI, and even the L2 were off-chip.
He's coming back just as one of his fingerprints has begun to fade, as we see hypertransport falling out of favor in the latest AMD consumer chips.
One old friend might be that warty old crossbar that AMD put in on the first K8 chips, which interfaced with the HT links and had a spot waiting for that first dual core.
Even though the core he knew didn't quite last the full duration of his absence, he might be in on the ground floor of a replacement for that base element of the uncore he might have helped spec out.
He was employed at AMD from 98 to 99.
It sounds like he could have been involved in the bringup of K7, although the design would have probably been quite final by 98, being that it was released in 99.
It sounds like he had a role in the early parts of the K8 development cycle, but even if he were working on K8 from day one he would have been gone for the majority of its development.
If we consider the lack of a competitive x86 core a significant factor in AMD's woes, Keller is a remnant of a happier time, but he doesn't have the bona fides of the standard bearer of a new Athlon. Athlon was mostly done before him, and K8's core did not hop far from K7.
His most recent experience doesn't look to turn around the weak core problem, and he'd have to contend with the fact that just as he has probably changed in over a decade, so has AMD.
It's not like he's getting the hit band together again, not unless AMD is also rushing to hire back the engineers and designers Ruize and Meyer lost or removed since then.
That said, his SOC and networking experience is probably more relevant to AMD in the few niches it still lingers in. That's not strong comfort for the desktop market. For servers, his most recent experience screams more Seamicro board than powerful Xeon blade.
He does have experience in working to put together disparate system elements and putting together protocol. This might be good for HSA, although it would be more likely that AMD didn't just hire the guy to put together the initiative it's already announced and outlined. The heterogenous elements are about the only thing left to put together, as we are a far cry from the days when the IMC, PCI, and even the L2 were off-chip.
He's coming back just as one of his fingerprints has begun to fade, as we see hypertransport falling out of favor in the latest AMD consumer chips.
One old friend might be that warty old crossbar that AMD put in on the first K8 chips, which interfaced with the HT links and had a spot waiting for that first dual core.
Even though the core he knew didn't quite last the full duration of his absence, he might be in on the ground floor of a replacement for that base element of the uncore he might have helped spec out.
If I remember correctly, the K7 was mostly developed by the design team acquired from NextGen. Keller was working on a successor to K7. The FPU from this successor was then bolted onto the K7 late in development (and the successor cancelled).
I'd imagine he had some inputs on the system level design of K8, with its integrated memory controllers and HyperTransport.
Well, prior to being acquired by Apple, PA Semi did develop a potent PPC core with good performance, and outstanding performance/watt (which saw multiple military design wins).
Cheers
Who's stupid enough to buy AMD?
itsmydamnation
Veteran
How much debt do they currently have?
Somebody who wants to accrue in some serious fire-power in a patent war, perhaps. At this particular point this is one of AMD's primary uses...if somebody else wants their engineers it's pretty obvious that they need but show up and inquire, and I doubt that many would care much about their management structure or other ancillary departments. On the other hand, AMD's patent portofolio is quite respectable. However, I don't place much faith in this particular rumour...yet.
About $3 billion, I think. But that's factored into their market capitalization, which is why it's just under $3.1 billion.
Considering the IP they have, AMD is not a bad buy, just not something I could see Qualcomm making much use of (unless they really need some of their patents). Samsung, on the other hand, might be interested.
If we assume, for the sake of argument, that all of this is true, then it's no wonder that they're only valued as low as they are. Because not a lot of people seem to care about the compute abilities of GPUs, especially those of AMD.zorg said:
To be fair, GCN-based FirePros were only released a couple of days ago. Earlier VLIW cards were much less compute-friendly, so this is kind of new for AMD.
Actually, I'd say the largest intangible asset that AMD has is the ability to effectively design and validate x86 CPUs. There is only one other company in existence with that ability.
And no, ARM will never buy AMD. ARM makes IP. AMD makes chips. ARM doesn't wish to compete with its customers. Samsung would make much more sense.
DK
And no, ARM will never buy AMD. ARM makes IP. AMD makes chips. ARM doesn't wish to compete with its customers. Samsung would make much more sense.
DK
And what's the 3rd one that can do high performing GPUs besides AMD and nVidia?
- Status
