[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Objectivity Parallel Write Tests
I have rerun the parallel write tests described this morning with 15,000 and
100,000 events on each of 1-24 processors and 1-12 processors, respectively.
(Scratch disk space problems interfered with attempts to run the 100,000-event
case on more than 12 processors.)
I have also added a little more analysis, and aggregate throughput numbers
from both system and user perspectives; the former is a more optimistic view
of system performance than the one discussed this morning.
The page is still accessible from the GCA Meetings page; you may need to
refresh your browsers, though.