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: Matt Webster who wrote (33012)	5/29/1998 12:21:00 PM
From: Jim McMannis	Respond to of 1570363

Matt, What you say makes too much sense. What we are really doing here is pulling for Luke Skywalker as he goes into battle against the Evil Empire and Darth Vader. Jim

To: Matt Webster who wrote (33012)	5/29/1998 1:39:00 PM
From: Ali Chen	Respond to of 1570363

Matt, <talk about alternative ways AMD could become a profitable company.> Your speculations are based on questionable ideas, IMO. Let me elaborate: <Jerry Sanders has a vendetta with Intel..With a new management team, they may again be open to new ideas.> This is not completely true that the sole driving force behind the war against INTC is Jerry. Although he is the main figure in the field, the whole PC industry is backing his efforts to break Intel monopolistic chains around PC manufacturers. Without this unadvertised support, AMD would be dead long time ago. Without the "Sander's factor" the outcome would be the same. Following your capitulative sentiments, everyone should give up in desktop market and leave the field clean for Intel, right? <Instead of wasting time on 3DNow!, AMD could have been working on a K6 derivative that had large on-chip cache (L1 and L2), parallel processing, etc.> First, as far as we all know, they are working on it (K6-3 or K6-3D+, due this fall I think). Second, your assessment of technical directions is too superficial, softly speaking. There are principal technical reasons for not going to much larger L1 caches; on-chip L2 caches reduce yields significantly due to larger die sizes; there is not too many tasks that can be effectively run on parallel processors in regular desktop applications. <This was a strategic mistake ... server CPU's cost more, what is the problem?> The problem is that the better CPU makes a server better. The better server means that it CAN SERVE MORE CLIENTS!!! This means that for every single "better" server you have an expanded market for low-end clients! This is the first idea you overlooked. The second idea is that the crappy x86 ISA (and the Intel system architecture itself)is no good for server workload whatsoever. This is not under a dispute, and the proof is the EPIC/Merced itself. In the contemporary client-server model of computing, the instruction set compatibility is not an issue - connections occur via computer- independent high-level logical protocols. When you surf the net, you never notice what kind of computer is on the server side - SGI, Alpha, Pentium, or else. The Alpha problem is price. <I'd much prefer they charge $1,000/chip and make a profit on a smaller number than charge $100/chip and make a loss on a large number.> These are clearly the two market strategies. Let's wait 2 more quarters and see which one works. Could be both.