Lawrence, in connection with your point about NC's (if they get off the ground, my ed.) requiring Multi-user NT servers, I got my first inkling of the power projected for the P7, aka Merced. It came in HP's workstation flyer called "FYI". The P7 is the 64 bit chip that has been getting more in the news lately. Microsoft and Intel announced this week that they are jointly developing NT for it, e.g.
Now, bear with me for a minute while I get some numbers out. The HP FYI is touting their latest workstation based on the PA-8000 RISC chip as taking the lead for world's fastest workstation, , etc., etc. They have a graph of processing power vs. time, 1986 to 1996 and beyond. Anyway, the one they're touting now is at about 400 SPECint92. In what looks like next year (small graph), they get to about 500 with the PA-8200. Then, they have a discontinuity in the vertical, or performance axis. After that, they show no more PA---chip based workstations. Just Intel Merced. The reason they had to do a discontinuity in the vertical axis is that the performance, with the Merced, jumps to over 2000 SPECint92 in 1998. That would have put it off the graph! That's a 4X performance jump in one year!
So, here's the "PC microprocessor chip maker" coming out with a chip in 1998 4X faster than the best HP had a year before that.
If all this works out re performance and schedule, HP has one heck of a joint venture going for their workstation market. and Intel has one heck of a chip. Of course, who knows what Sun has in development. But, if they don't have something similar, they better have some strong Java brewing.
Tony
|
