V-6.1 ZAPHOD Single Priority Array O(1) CPU Scheduler

Version 6.1 of the ZAPHOD single priority array scheduler patches for the 2.6.10 kernel are now available for download and evaluation from:

<http://prdownloads.sourceforge.net/cpuse/patch-2.6.10-spa_zaphod_FULL-v6.1?download>

This contains the extra per runqueue CPU statistics for improving GA fitness functions that we discussed.

struct runq_cpustats {

#ifdef CONFIG_SMP
	unsigned long long timestamp_last_tick;
#endif
	unsigned long long total_delay;
	unsigned long long total_rt_delay;
	unsigned long long total_intr_delay;
	unsigned long long total_fork_delay;
	unsigned long long total_sinbin;
};

total_delay - is total time spent by tasks waiting for CPU on this runqueue
total_rt_delay - as for total_delay but only for real time tasks
total_intr_delay - is total time spent by tasks waiting for CPU on this runqueue after being woken to service an interrupt total_fork_delay - is total time spent by tasks waiting for CPU on this runqueue for their first time slice after forking

Peter

Peter Williams                                   [email protected]

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