Thursday, June 11, 2015

alleviation

Finally the problem might have been caught... might.  Dr.Remy and I ran a sample run with 60 threads and saw that the system average load was increasing to a max point of 8 or so. So the computation was stopped... by me so no prob... and reran the program but with 20 threads instead of 60. The average load didn't go past 5.. This was good. So the prevailing theory is that when the computer was running with 40+ threads, it was working as fast as it could on this one process. But when another program was run, by someone else on campus, the computer didn't could not think of what to do...so it stopped. When 60,70, and 100 threads where used the computer couldn't handle it so it would take longer to complete the task, or just stop altogether. So as of now, I'm going to work on the continuation class, and rerun all the mutations again but using 20 threads.

No comments:

Post a Comment