Sun, OPL and Rock
This article by Ashlee Vance sheds light on something, that kept my mind busy the last month: How to explain the dense pack of IV+ 2.1 GHz, OPL and Rock. To make one thing clear, the conclusions are mine, not the one of Sun.
UltraSPARC IV+ reaches higher speeds than planned and the development of Rock is vastly more successful than anybody has anticipated (Well, engineers are conservative in their time schedules . And grown up with Scotty as a role model, their say “3 days” to make it in “3 hours” ;) )
So, what´s the rationale of OPL?
Well, OPL was and is to a part a contingency plan for problem with IV+ and Rock. You don´t run a multi-billion company without having a PLAN B to at least F in your desk. What would happen when IV+ had hit the wall at for example 1.3 GHz ? Or when there were severe problems with the Rock processor ? Sun would be essential toast. You don´t do that!
At second OPL has an advantage: It´s made for running old-wold applications at the maximum speed. Only a few threads, but these must run on maxium speed. Every MHz counts. Application that can´t take advantage of many cores . They get fewer with every iteration of our Multicores and with every iteration of new applications, but … welll … they still exist. And so OPL makes perfectly sense. Even in the advent of Rock. The issue: With Niagara on one side and the multicore wave at x86, these applications are really rare.
Now you have to take one thing into considerations: Compared to other platforms, migrations under Sparc/Solaris 10 are incredibly boring. It´s the same Solaris 10, it´s SPARC system … so you don´t loose anything by using OPL …. oehm … this night, and Rock tomorrow.
At the end there is only one thing that counts. Even the media believes, we are back. Or as Ashlee Vance of the Register wrote:
For the first time in many years, we're feeling bullish about Sun's high-end system prospects.