by INQUIRER staff: Wednesday 26 October 2005, 11:41
THAT NICE ANALYST CHAP Nathan Brookwood has made some comments about the u-turns Intel made last Monday. While the first three paragraphs are complete Double Dutch to us as they mention someone called Casey Stengel who appears to play a game called rounders, Nathan gets into his stride about 64-bit Xeons and Itaniums soon enough.
He points out that Intel really really needs to have an on-chip memory controller, but, he says: "Intel's phobic reaction to on-chip memory controllers is somewhat understandable. Both of its earlier attempts to incorporate this technology in its processors (the 386-SL and the stillborn Timna) ended in disaster."
He also says Intel only sells a few hundred thousand Itanium processors and its "long term survival seems increasingly challenged".
He says that Intel won't be able to field a server processsor with an on board memory controller until 2009. And that, he says, means "little likelihood that Intel will be able to claim performance leadership." µ
Here's something to echo what you have been saying for some time, the words of Brannon Batson ex-DEC/Intel Architect in a reply to Linus Torvalds about Intel's current Server debacle.
>"Never ascribe to malice, that which can be explained by >incompetence".
Massive, gargantuan, unrepentant incompetence in an unfocused and constantly shifting management hierarchy. I'm willing to bet that upper management and most of middle management still does not understand the root cause of these problems. And the problem feeds on itself in an organization like Intel--because there will always be a dozen opportunistic people feeding upper management convenient and easily digested (but utterly false) explanations: "all you have to do is put me in charge of 'foo'". This causes more shifts in the power structure, inserts even more layers insulating upper management from reality, which leads to more bad decisions, which leads to more failed projects.
It's a self-correcting system, though. Eventually they'll either run out of money or go back to basics. The latter mostly involves putting decision-making back into the hands of people who understand the technology. Unfortunately, the management hierarchy on the server side is so overwhelmingly polluted with mediocrity now that it's going to take them a long time to figure out who that is. I wouldn't bet money either way.
News 
Market Data 
Discover 
