It turns out that the best performance we get
is running the server on the sparc10 while
running the clients on each of our GX's repectively.
The MP's perform equal to running a server on a sparc2,
and the Sparc GT (graphics tower with 21" monitor)
runs slower than both a client and server running on
a sparc 1 (GO FIGURE!). Has anyone experiemnted
around with any optimizations using
threads packages under Solaris or SunOS running
on an MP or sparc10 MP? Any other optimization
advice would be greatly appreciated. We're
pushing the xpilot limits with customized side
effects and "pressure-so-high-your-palms-sweat"
server setups.
Greg
Dave.