Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: Joe NYC who wrote (97256)	6/13/2003 2:17:16 AM
From: Petz	Read Replies (3) \| Respond to of 275872

We both agree that, for now, Intel's main goal is to keep the desktop market from moving towards 64 bits until their "low cost" 64-bit solution is ready. I still think that will be Itanium IA-64, not some new Yamhill IA-32-64. Unfortunately for Intel, the "affordable" 65 nm IA-64 will perform like a Celeron compared to 65 nm dual Opterons and be roughtly equivalent in cost and performance to a 90 nm Opteron. Itanium's performance per transitor is just low. I do think the Prescott with extra pins that comes out mid-2004 probably has some kind of segmented >32-bit addressing. Or maybe the extra pins are just direct DDR and DDR II interfaces, i.e., memory controller on-chip. That would take away one of Opteron/Athlon 64's advantages. <I think Intel will have to reveal its cards in September, to undermine adoption of AMD64.> They will probably yell a little louder about higher clock speeds, and higher memory bandwidth, at least vs. Athlon 64. They can give big incentives for software companies not to optimize for AMD-64. And they can make server chipsets that support 40 bits of memory address space with 32-bit segments. Microsoft must already be supporting something like this for Xeons. (or maybe it'll be "in-the-chip"} <I wouldn't be surprised if [in September] they started emitting some vapor about the IA-32-64 beast, so that third parties don't invest in AMD64 support. I am surprised that Intel has not done it yet.> The fact that they haven't, or that we haven't heard a peep about anyone writing code for a Yamhill instruction set is the main reason I dis-believe it is currently in their plans. But you are right -- they could announce it in September to kill support for AMD's solution. Of course, they would promise chips for 2004 and they would actually arrive in 2005 at the earliest. But Mickey and Carly would probably believe them. If we don't hear anything about Yamhill around AMD's September launch, then I think you'd have to agree that Yamhill is dead, dead, dead. Petz

To: Joe NYC who wrote (97256)	6/13/2003 4:01:32 AM
From: mozek	Read Replies (1) \| Respond to of 275872

Joe, I think you've characterized Intel's necessary countermove, which I suspect wouldn't feel so good right now if you were in Otellini's shoes. As far as your call of a September unveiling of the strategy, that makes complete sense. Now that Microsoft's committed to support AMD-64 and AMD and Intel have diverged their 64 bit architectures, I believe the stakes will get much higher by next year. It's good to see today that AMD's R&D has strong legs under it because they're going to need them over the next few years. As far as software/applications support for native 64 bit, AMD64's assembly language looks like an experienced IA32 compiler writer's dream, while IA64 will likely provide good optimizers with job security for years to come :-) Mike