[BusinessWeek] Plugged In: AMD bets future on bold vision - Page 4

Not true. The original launch was delayed, but R&D never stopped.

In fact, its currently shipping. (now in its 3rd generation)
http://www.anandtech.com/show/6265/intels-xeon-phi-in-10-petaflops-supercomputer

Sandy Bridge was supposed to have Larrabee as its GPU but due to the delay Intel stuck with its older GenX GPU. Skylake was always planned to fully integrate the gpu and cpu, allow for memory sharing, etc. But now its also the first chip that will feature larrabee.
Edited by erunion - 9/14/12 at 7:05pm

That's beneficial to average consumers how??

Xeon Phi != Larrabee, though it does take many elements from it

source on your claims regarding skylake

Because its going to be in skylake in 2015, following right behind HSA in 2014.

http://semiaccurate.com/2012/06/12/what-comes-after-knights-landing/

Edit: I should add that Intel has been talking about heterogenous compute since Larrabee was announced. (I was wrong earlier, it wasn't 2009 it was 2008) Larrabee was created for future heterogenous cpus, just because it hasn't existed (yet) as a discrete gpu doesn't mean Intel scrapped their entire roadmap.


http://www.anandtech.com/show/2580/14
Edited by erunion - 9/14/12 at 7:40pm

HSA may eventually take off, but a future dominated by AMD? That's just too funny! TI and Samsung are involved in HSA and you think AMD will be running things?

Nothing stopping intel from doing HSA as well you know.

HSA isn't about money, it's about politics. Proprietary vs. open standards. The consumer benefits more from the latter, assuming it has sufficient backing.
Edited by Homeles - 9/14/12 at 8:17pm

Well that's a gamble for intel. If HSA is the future it becomes a race that amd might be further ahead than we know and could beat out intel. Or maybe HSA is garbage and now intel invested a ton of money for nothing. So intel probably will just be keeping a watchful eye for a while.

Not really.

To me, it almost sounds (or does sound) as if this were some form of trying to aim at software developers and let them know that there is another way to program for CPU's. In short what I'm saying is, this is kind of dumbed down from what I'm used to from "developers." This instance is kind of lacking in structure as far as an information standpoint. Not only does this lead to more questions, but also leads to more issues as far as what is expected. The main issue here is that programmers do not want to learn or even HEAR about another language, only a form of programming language that makes their job easier, and this by far is not easier.

I'm weary of AMD trying to weave their way into the CPU market not, instead of "earn" their way in. I believe they WERE on the right track many years ago, but that is not today. I fear more than am happy of the views of today. More and more want to try to compete, instead of actually compete. In another view, more players wish to sit on the bench, rather than play the real game because it is less work and more money. In my eyes, AMD is going backwards and not forwards as a company. It's like feigning laziness for success. One can only believe for so long that they are the best but in actuality, are only riding off anothers success.

To me, the real pain here lies in the fact that AMD SHOULD have been a leader in the industry, but today STILL fall short of Intel's dealings. I, for one, wish that AMD would instead of publishing articles, show that they are a true leading amongst the industry. I have been a long time follower of AMD, but tell me why it is that I cannot aim to purchase AMD architecture when I really want to? Intel has always had the "best" processors in the industry and I feel that if AMD does not do BETTER (in terms of their strategy,) then they are doomed to fail as all that came before them. Regardless, if AMD does not take into account what is happening/what is in their best interests, then they are certainly doomed to fail.