[lkml]   [2004]   [Aug]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
Subject[PATCH] V-5.0 Single Priority Array O(1) CPU Scheduler Evaluation
Version 5.0 of the various single priority array scheduler patches for
the kernel are now available for evaluation.

This version adds NEW FEATURES to ZAPHOD and HYDRA schedulers.

1. [ZAPHOD] My proposed replacement scheduler which offers runtime
selectable choice between a priority based or entitlement based O(1)
scheduler with active/expired arrays replaced by a single array and an
O(1) promotion mechanism plus scheduling statistics with new simplified
interactive bonus mechanism and throughput bonus mechanism. This
version adds hard (/proc/<TGIG>/task/<PID>/cpu_rate_hard_cap) and soft
(/proc/<TGIG>/task/<PID>/cpu_rate_cap) CPU rate caps (in parts per
thousand of a CPU) and an unprivileged RT mechanism (which was inspired
by Con's SCHED_ISO patchs and will treat tasks that try to change policy
to real time but fail due to lack of privilege as pseudo real time tasks
and run them at MAX_RT_PRIO provided that their CPU usage rate is less
than the threshold specified in
/proc/sys/kernel/cpusched/unpriv_rt_threshold (in parts per thousand --
default 10 i.e. 1% of a CPU). Obviously setting this value to zero will
disable the mechanism completely.). Tasks that have a soft cap of zero
are treated much the same as Con's SCHED_BATCH tasks and will have their
time slices multiplied by the value in
/proc/sys/kernel/cpusched/bgnd_time_slice_multiplier (default 1):


2. [SC] Slightly modified version of Con Kolivas's staircase O(1)
(version 7.I) scheduler (includes SCHED_ISO and SCHED_BATCH
modifications) with active/expired arrays replaced by a single
array and an O(1) promotion mechanism:


3. [HYDRA] Runtime selection between staircase, priority based and
entitlement based O(1) schedulers:


Other schedulers are also available from

Your comments and feedback are requested.

Peter Williams

"Learning, n. The kind of ignorance distinguishing the studious."
-- Ambrose Bierce

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 14:05    [W:0.041 / U:4.072 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site