T
Tomas
I am running the (statistical software) application Splus
6.1 on dual hyperthreaded Xeon 2.8G. When I run 1 instance
of Splus, CPU utilization is about 35-40% and only one of
the 4 logical CPUs- seems to be engaged (according to the
Task Manager). My job runs for let's say 5 minutes. Now,
when I run in parallel a second instance of Splus and run
on it a job identical to the 1st one, the CPU utilization
goes up to 80-100% and all 4 logical CPU-s are engaged but
it still takes 10 minutes (instead of the expected 5) to
complete the 2 identical jobs on the 2 separate instances
of Splus???
Any insight of what is going on and how to "properly" take
advantage of the "extra" CPU capacity?
6.1 on dual hyperthreaded Xeon 2.8G. When I run 1 instance
of Splus, CPU utilization is about 35-40% and only one of
the 4 logical CPUs- seems to be engaged (according to the
Task Manager). My job runs for let's say 5 minutes. Now,
when I run in parallel a second instance of Splus and run
on it a job identical to the 1st one, the CPU utilization
goes up to 80-100% and all 4 logical CPU-s are engaged but
it still takes 10 minutes (instead of the expected 5) to
complete the 2 identical jobs on the 2 separate instances
of Splus???
Any insight of what is going on and how to "properly" take
advantage of the "extra" CPU capacity?