Skip to contentSkip to conversation

self tuning scheduler

PATCH Jim Houston 2 participants 3 messages Mar 11, 2003 – Mar 12, 2003

lore
Share & tools
lore
#0 lore raw

Hi Mike

I made a bit of progress on understanding the irman problem with my scheduler change. When I run irman and top, the processes end up with priorities like:

irman parent 36
irman child 21
process_child 31-33 (group of 9 processes)

Since I expanded the range of priorities (to 0-79) these are quite favorable priorities. They are all have MAX_SLEEP_AVG bonus equivelent of nice +10.

It's a priority inversion problem. The irman child is waiting for a read. The process_child processes are happly running as a group at approximately the same priority. The irman parent is starved because it is at a lower priority. It is at a lower priority because it uses more cpu on each pass. It is doing the gettimeofday calls while the child only does the pipe read & write. The parent gets an occasional boost from the fairness_update() code so it doesn't totally starve.

I'm contemplating making synchronous wakeups share the run_avg between the processes so that groups of cooperating processes would clump at the same priority.

I also wonder about trying to detect cycles of synchronous wakeups. It seems that a group of processes passing a token should be treated as compute bound.

I'm still playing with the "make -j 30". I can adjust the priority range where I start enforcing interactive behavior. I may wire it into the rq->prio_avg. I assume that you can tolerate a bit more timing jitter when doing a "make -j 30".

Jim Houston - Concurrent Computer Corp.

#1 lore raw
Quoted 1 line · Jim Houston

I'm still playing with the "make -j 30".

Just a note: if you play with it in a ram poor environment like mine, you may need to adjust the number of tasks considerably... gcc > 2.95.x makes size "Godzilla" footprints.

-Mike

#2 lore raw
In reply to Jim Houston
Quoted 14 lines · Jim Houston

Hi Mike

I made a bit of progress on understanding the irman problem with my scheduler change. When I run irman and top, the processes end up with priorities like:

        irman parent    36
        irman child     21
        process_child   31-33   (group of 9 processes)

Since I expanded the range of priorities (to 0-79) these are quite favorable priorities. They are all have MAX_SLEEP_AVG bonus equivelent of nice +10.

Quoted 8 lines

It's a priority inversion problem. The irman child is waiting for a read. The process_child processes are happly running as a group at approximately the same priority. The irman parent is starved because it is at a lower priority. It is at a lower priority because it uses more cpu on each pass. It is doing the gettimeofday calls while the child only does the pipe read & write. The parent gets an occasional boost from the fairness_update() code so it doesn't totally starve.

I forgot to mention something odd I encountered wrt the parent's priority. I can set the irman child SCHED_RR, and it shows up in top with priority -2. The parent remains at priority 36 and remains starved. Max renice doesn't help either.

-Mike