Replies to post #42723 on Advanced Micro Devices Inc (AMD)

[chipguy]

I was going by memory. I'm pretty sure I've read of Itanium price cuts before. In fact I seem to recall some very severe Itanium price cuts last year?

Wrong. Intel introduced lower performance and lower cost I2's
in the 14 months since the 1.5 GHz Madison 6M was introduced
but existing models stayed at the same price. BTW what has
happened to Opteron 844 pricing over the past 14 months? Has
it stayed the same?

Ooops I guess not:

June, 2003 - $2149

http://marketwatch-cnet.com.com/AMD+expands+Opteron+chip+line/2100-1006_3-1022167.html

August 2004 - $698

http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_609,00.html

I think you mentally misplaced the "very severe price cuts" for
Opteron and attributed it to IPF. Wishful thinking perhaps? ;-)

[wbmw]

Re: HPQ is the biggest backer of the Itanium (only significant OEM)? HPQ reported that it was having earnings problems due to slow "big Iron" sales. There have been reports that a significant number of HPQ customers were swithching from HPQ to IBM/SUN for their "Big Iron" because they didn't want Itanium. Ergo, a price cut in Itanium would seem to make sense to increase demand.

If you said that HP's customers moved away from HP "Big Iron" because they either re-evaluated the need for scale-up systems, or because they were disenchanted with HP's support roadmap, then I might be inclined to agree. But if you said that HP's customers moved to IBM or Sun for the same "Big Iron" infrastructure, then I'd find that hard to believe. IPF has huge performance improvements over UltraSparc and huge cost improvements over Power. It has the best price/performance of any "Big Iron" architecture, and it's also catching up in terms of software development. At last count, IPF ported applications numbered about 2,000.

If HP is failing to execute on IPF, then that's a big problem for Intel, but not one they can fix with the CPU. I think Intel should push harder for IBM and Dell to adopt Itanium, rather than depending on HP. The tier-2 OEMs have some pretty nice systems, too (NEC, SGI, Unisys, Bull, Hitachi, and soon Fujitsu), but the volumes don't compare to the big MNCs. Intel's best chance of wooing Dell would have been a new fully featured chipset, but if rumors of Bayshore cancellation are true, then Dell is effectively out of the picture, and it really will depend on HP and the tier-2s (or more accurately, HP).