SI
SI
discoversearch

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 : DELL Bear Thread -- Ignore unavailable to you. Want to Upgrade?


To: divvie who wrote (2201)10/26/1998
From: Bilow  Respond to of 2578
 
Hi divvie; You're right that a lot of what happens next boils down to what the difference is between a desktop and a server. I feel that they are on a continuum, and that the integration revolution will move up from low end desktops, then to high end desktops, then on to low end servers, and will finally get to high end servers. So I see the markets as being connected now, and being even more connected in the future. I suppose I should illustrate this belief with an example from the past...

There used to be a major distinction between "personal computers" and "workstations." Typically, the workstations ran Unix, and had much better graphics. They also cost a lot more, and it seemed like the two markets would stay separate forever.

Back in 1985, I had a job with Burroughs to visit companies that designed processors, and to scope them out to see if Burroughs might contract with them to design a machine for Burroughs. (I think I have the year correct. It was the year that Burroughs merged with Sperry, in any case.) I saw presentations by close to a dozen design houses. The one that impressed me the most, rather than a famous place like MIPS was a tiny little company that did nothing other than design specialty processors. The computers they were doing all their simulation on were PC-ATs. They used code they wrote themselves, and relied on a sort of top-down register level simulation of the machine to verify functionality of the processor. This was at a time when a handmade high-end 32-bit processor required multiple gate arrays. Their technique was beautiful, in that it meant that they could put multiple engineers on the project, without modifications of the ALU by one engineer messing up the simulations of the TLB by another engineer. Essentially, they did this by inheriting test vectors down through an RTL model of the computer (processor plus main memory and I/O). I learned the basic fact that the better an engineer is, the less power is required in his workstation to get the same job done.

Anyway, from that time on, I expected PCs to take over the workstation market. It took a lot longer than I expected...

It probably started with MMI (?) providing the famous program "PALASM" for designing PALs on the IBM-PC. Gradually more and more complicated engineering functions were ported to the PC. First code for board layout was ported to DOS. Then the FPGA makers, like ALTR and XLNX wrote code for windows. Now, I think, you can get pretty much any (common IC) engineering code to run on PC workstations, using NT or 95.

The fundamental reason for the convergence of those two markets was that the PCs were so much cheaper than the specialized workstation hardware. And all the things that workstations needed to do extremely well, were also things that it was nice to have a regular PC do extremely well. So the markets converged. Advances in the workstation market were copied over to standard PCs.

As of now, the fundamental differences between a high end desktop PC and an engineering workstation have been erased. I expect to see the differences between a desktop PC and a server to also be slowly erased over time. Just don't hold your breath...

-- Carl

P.S. While I like to hold stocks for 15 minutes or less, I am getting the impression that we are seeing a daily chart top in stocks. So I predict that DELL will be lower on Wednesday close, than it is on Monday open this week. (Of course, the Fed could make me a chicken little...)