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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Scumbria who wrote (91828)	2/6/2000 3:00:00 PM
From: kapkan4u	Read Replies (1) \| Respond to of 1577125

<If it is being done by the Nexgen team, it will probably be junk.> This is just a bogus statement. In case you didn't know the Sunnyvale team designed the Athlon FPU. Need I say more? Kap

To: Scumbria who wrote (91828)	2/6/2000 3:01:00 PM
From: Charles R	Respond to of 1577125

Scumbria, <The complexity of trying to service multiple address streams on a single datapath is too high. It is much easier to provide a single datapath per stream, including a dedicated cache and bus.> Though I have lost the touch to estimate the impact of the tradeoffs I can believe that. Especially in the context of MHz. <Multiported caches tend to become very complex and slow. > This is what the original topic was about - alternative to multiported cache is a cache running at a higher MHz sufficient to service requests from multiple execution units. Especially since we are talking about L2 caches, queueing theory might work in its favor. Now, there is a cache size/MHz issue and I don't intend to trivialize it. Chuck

To: Scumbria who wrote (91828)	2/6/2000 3:20:00 PM
From: Charles R	Read Replies (1) \| Respond to of 1577125

Scumbria, <If it is being done by the Nexgen team, it will probably be junk.> I heard somewhere that Fred Weber is a pretty sharp cookie. I am not sure but he may actually be one more of the DEC boys. <- Do you see a reason to support higher MHz than Athlon can do for 2000? Willamette.> I think Wilamette is being overestimated just like Athlon was overestimated. It will take a while for Wilamette to ramp on the desktop side. It will take longer for Foster to ramp up on the server side. It will be even longer for Wilamette derivatives to ramp on the laptop side. Assuming Intel outexecutes AMD with Wilamette, this whole story could playoff something like this: 2000: Q1: We are already here and AMD will do north of $1 EPS this Q so I guess no one is worrying about this quarter. Q2: Athlon penetraion improves. Some business SKUs. Q3: Athlon comes to laptops and servers. Athlon in a a lot of business SKUs. Q4: Wilamette ships some units at the high-end. AMD starts kicking PIII butt in mobile space. Server penetration increases. 2001: Q1: Wilamette kicking Athlon butt high-end. Athlon reduced to a highly-profitable midrange consumer/business space as Windows 2000 and Windows Millenium ramp. Athlon portables continue to kcik PIII butt. Q2: Wilamette continues to kick Athlon butt. Athlon servers are everywhere. Foster starts getting closer to reality. Q3: The story begins all over with Sledgehammer and Wilamette and the players switching roles. In summary what I see is a leap frog game that does musical chairs around segmentation boundaries. Based on this it is hard for me see either party losing. Only thing is that AMD will substantially outperform Intel because it will gain market share in areas it has no presence to date (corporate systems, high-end laptops, servers) I see AMD being very well positioned going into 2001. The issue that AMD faces is one of execution. Chuck