Sometimes I just wish AMD would die already so we could all stop talking about this.

I got a friend at AMD, and he says that the 45 nm chips are having all kinds of problems. He doesn't have much faith in them yet, but in 6 months they possibly may have the kinks ironed out.

I dunno. Sometimes I think these guys have their heads in the sand. The competition is hitting on all cylinders right now.

I dislike the job cuts but they are most-likely the consequence of the bad start of the Phenom. Such things can happen, just like a burning fab for example, and people got told about the TLB-bug from the beginning. That was responsible. So now after the "medium glitch" it is all back to normal with a clear goal, some spring cleaning and a puffed-up talk.

I hope that they keep bringing those clear CPU designs even when the "true" quad-core design turned out to be more risky than they could possibly handle. AMD found one of it limits here but that happens when you want to push your limits. Its just sad to see those job cuts :-/

Well, they really need some nice upgrades up their sleeves aswell, besides the shrink. Intel were out of sync with their Netburst tech, but the juggernaut has made the turn back on course. A 1-cycle SSE implementation that seems to do something (with todays compilers anyway, I guess), faster units where it matters, fast caches, shared L2 cache, etc, etc.

The Phenom's were supposed to have a 1-cycle SSE kind of implementation aswell, but it is barely noticable in practice as far as I can tell, and this is probably the biggest loss when it comes to the performance numbers. Power efficiency isn't all that good either.

I guess they're back at keeping their head over water, keeping what they have as competative as possible, like Intel were before. But they won't be able to afford doing that as long as Intel did. So I hope a shrink isn't all they got..

"Robert Rivet said AMD is on track to kick off 45nm production "at mature yields" in the summer, and AMD COO Dirk Meyer added that volume 45nm product shipments are scheduled for the fourth quarter of this year."
I keep reading that.
If the process is running that great, then why are they not ramping it up?
they claimed the same thing 6 months ago.
What AMD needs the most, are DualCores with high clockrates.
65nm brought no more than a shrink and some energie saving, it's time the clockrate surpasses the 3GHz.
IBM did it.

Actually, I would guess GCC or MSVC. And I don't think they include different codepaths depending on it being Intel processor or not, like the Intel compiler apparently does. Or did, at least.

But it would be nice to see some kind of comparison between different processors and different compilers (like PGI for instance).

First off, AMD's last 6 quarter have not had "spectacular losses". AMD has been cash flow positive on operations in the last 3 quarters, including Q1'08. The only thing "spectacular" have been the goodwill write downs based on the acquisition cost of ATI. Since the purchase price is always higher than the book value, goodwill has to be written down at some point, and AMD chose to do it earlier rather than later.

Second, Intel may be "hitting on all cylinders" but it's been losing marketshare pretty steadily since 2002 per mercury data:
1997 7.3
1998 11.9
1999 13.6
2000 16.7
2001 20.2
2002 14.9
2003 15.5
2004 15.8
2005 18.2
2006 22.9
2007 22.1*
Figures are a full year average - AMD finished Q4'07 with a 23.1% share. If anything the core2duo and TLB issue may have slowed the growth, but hasn't stopped it.

AMD's only "issue" is trying to absorb ATI's overhead. Apparently Intel was ATI's biggest customer buying something like 1/3 of their products - but then after the acquisition, Intel bought nothing. ATI was also re-tooling in 2007 so, ATI simply wasn't contributive to AMD's core business from a profit point of view.

As far as whether ATI was the right idea - I don't think AMD had a choice. As things progress, the big guys (hp and dell) want turn key solutions, especially on notebooks, and it took ATI's experience to deliver that. It appears to be working because the Puma notebook platform has over 100 design wins (twice that of Turion), and reportedly consists of 70% of all new notebooks in design right now. AMD traditionally has had a much smaller presence in notebook than desktop.

So, to conclude, with B3 server chips now shipping and apparently seeing wide adoption, the Puma platform shipping in June (in time for back to school), and some serious belt-tightening from AMD I see it emerging 2008 as a very serious, very lean competitor.

Plus

#22: I will never forget the Al Pacino character in a movie refering to "Jack Daniel's" as "John Daniel's" - and when another character went to correct him, he said, "When you've known him as long as I have, you call him John!"

Whatever you call AMD, they absolutely have to start writing with black ink within 2-3 quarters.

I'm cautiously optimistic. I want to see them succeed, but they've dug themselves a pretty big hole that they need to climb out of. If they can execute properly on these plans, they should be able to pull it off... but that's a pretty big "if".

I know it's been said in many different ways by lots of different people, but, I'd be more optimistic about AMD if Ruiz was no longer involved. I don't think they're going to disappear any time soon, but I also don't think some of the higher ups are really paying attention.

You want to cut operating costs, give the execs reasonable salaries instead of the millions a year they might get.