16 mutext lock/unlocks per query?

Rick Jones raj at cup.hp.com
Wed Dec 5 21:52:49 UTC 2001



> TSI is my generic term for "test and set interlocked", which means
> some instruction or sequence of instructions that's guaranteed to
> succeed for only one thread regardless of cache pipelines or number of
> processors.

OK, for PA-RISC that would be load anc clear word (LDCW). There is
either a test and set or fetch and add or somesuch in IPF, which I'm
confident is being used by the pthread_mutex_lock calls.

> > The profile should show the syscall stubs for when those things were
> > called. As I recall, that was rather rare.
> 
> so the lost processor time is all function call overhead unless your
> TSI has to empty the cache pipelines rather than snooping them.

I'm reasonably certain that the instruction does not have to flush the
caches.

rick
-- 
Wisdom Teeth are impacted, people are affected by the effects of events.
these opinions are mine, all mine; HP might not want them anyway... :)
feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...


More information about the bind-workers mailing list