Originally Posted by wingless
Hindsight is always 20/20....
AMD and ATI should have waited to get their next generation chips out BEFORE merging at such a critical time in their product cycles. Thats an error on the part of AMD/ATI's operations management plain and simple. The health of the R600 and Barcelona chips would have most likely been better due to lesser fabrication process complications on the CPU and GPU sides of things. Also ATI probably would have been a billion dollars cheaper if they had launched the R600 as poorly as they did (its doing pretty damn good now with the new drivers though).
It wouldn't have mattered what the corporate side of things was doing.
R600 and Barcelona were had problems on the engineering side.
They would have been delayed either way.
The name on the company doesn't make yields tank, chips leak, or create speed path issues.
It is also the case that AMD was running out of time.
Intel's hints for adding graphics modules to the CPU or CPU package predated the acquisition.
Let me ask all of you a question about brand loyalty. Why did Intel manage to stay afloat just fine being second to AMD for nearly 4 WHOLE YEARS with tons of people buying worthless Pentium 4's/D's when AMD can barely make it a year without the press and resulting public opinion destroying them for being 6 months late?
There's market inertia, better dev relations, better engineering (guess whose compiler AMD uses to get peak scores), more reliable supply, better platforms (with a few notable exceptions), some possible exclusionary practices (K7 era), brand recognition, and (relevant to the topic of the thread) better marketing.
Intel could afford the big ad campaigns and all the surrounding fluff.
The Centrino platform was itself a master stroke.
Intel even has a catchy jingle.
What about the other company?
Why don't you AMD m--sorry, I just forgot what the subject of this sentence was.
Intel could market itself to most of the desktop market.
It could supply the OEM market.
It could rely on the slow upgrade and verification cycles of the server market (somewhat).
With Pentium M, it kept growing in the lucrative mobile market.
With Intel's clout, it could intimidate or bribe companies to impede AMD's progress.
This likely happened in the K7 period, going by the rumors going around and how terrified the motherboard makers were of showing any Socket A product.
Why is that? What made all these AMD users and fans jump ship for Intel?
Because most rational buyers are fans of performance, not corporate cheerleaders.
Personally I do believe AMD has a part that will compete with PENRYN clock for clock (in non SSE4+ applications) easily and will be a pretty decent chip to have in your box with the upcoming new chipsets and socket types.
I can, clock for clock, match Penryn. Trouble is, I operate at .000001 Hz (I'm groggy in the morning).
AMD fans need to keep faith in a company that didn't take FOUR whole years to get a worthy competitor to the market after the competition outran them.
Welcome to the world of CPU design.
You've pointed out the approximate length of time it took to design a high-end x86 processor, one that is actually partially related to an earlier design.
(Recall that the Pentium M was in high demand. People were begging to put it in a desktop. Intel had a competitive product prior to Core2)
How about AMD, which hasn't released a totally new architecture since the year 1999?
Bulldozer's the new one, and it won't be out until 2009.
It will be a decade before something new comes out.