[Cbe-oss-dev] [PATCH] spusched: ix variable length timeslice calculation

Luke Browning lukebr at linux.vnet.ibm.com
Wed May 30 02:58:05 EST 2007


I said, "8 seconds is a very long time, particularly if you have to wait
for N-1 jobs to be scheduled first.  There may also be benefits in the
concurrent scheduling of SPEs and PPC threads, particularly, if we ever
implement a function offload library model.  You want SPE and PPE
threads to be time sliced together."

I didn't really mean to say function offload.  I meant to say that there
may be a benefit in having the PPE and SPE time slices expire concurrently
so that related PPC threads and SPE contexts are scheduled together.  There 
may be data flow considerations.  If the PPE thread is time sliced before
the SPE, there may be extra queueing, locking, signaling, errors, ... It
may affect the other SPEs also.

Luke




More information about the cbe-oss-dev mailing list