<< [...] You also have to realize that there's more to high-performance graphics workstation than just a fast processor.>>
I realize this. I write this surrounded by SGI and NT equipment -- SGI's ranging from an Onyx with two infinite reality pipes to Indigo's and Pentium Pro NT systems with everything from Matrox cards to Glyder/MP adapters with 40MB of texture memory.
<<Even though Intel is addressing the situation with PentiumII, it's going to take a while for Wintel vendors to evolve system designs that will have high-performance graphics and video I/O.>>
While Intel is certainly improving their chips as you would expect given Moore's law (doubling of transistors every 18 months), and is even improving some architectural issues (e.g., AGP (Advanced Graphics Port), etc.) -- I think this is only interesting when you're talking about $5K-$6K and less workstations. When you're talking about a $6K-$10K workstation, you have to look at what's going on with the 3D chip vendors and chart their path in terms of potential threat to SGI.
<<If you look at Intergraph, they've found it pretty difficult for the past 3-4 years trying to come up with NT workstation designs that will kill SGI. The O2, Indigo2 and Octane are competing quite successfully against NT in their respective price ranges. >>
I certainly believe that Intergraph, who is trying to apply a vertically integrated strategy to the PC world, has not been successful.
At the same time, I don't understand the market segments you are referring to with respect to PC/NT versus SGI competition. Anecdotal evidence (including that of a company I chair which does visualization engineering work) is that for a growing number of applications, PC/NT with something like SoftImage is a much more cost-effective solution on a TCO (Total Cost of Ownership) basis due primarily to software availability and prices and service and component prices. I'd love to understand what I'm missing here -- so if you have any data, even antecdotal data, please let me know.
<<The main problems for SGI had been self-inflicted: glitches in the transition to new products, lack of control over expenses, and over-ambitious expansion plans & acquisitions. These are MANAGEMENT problems, not technology problems.>>
I totally agree that many of SGI problems are self-inflicted. As a person who has witnessed SGI's bumbling attempts at product sales, upgrades, and service to the university I spend time at I absolutely believe you've nailed a huge issue. If SGI were executing on all cylinders, I think they'd be much more like Sun -- you'd see excellent shareholder return even with longer-term threats. But like Sun, I'd still be worried about them on a long-term basis -- because SGI seems to me to be dominant in a shrinking niche. |
