This is the mail archive of the glibc-bugs@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug nptl/4274] Performance issue: pthread_cond_signal() causes three context switches instead of one


https://sourceware.org/bugzilla/show_bug.cgi?id=4274

Torvald Riegel <triegel at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
                 CC|                            |triegel at redhat dot com
         Resolution|---                         |OBSOLETE

--- Comment #13 from Torvald Riegel <triegel at redhat dot com> ---
A new condition variable implementation has been committed, which as different
performance properties.
The statements made in comment #9 do not really consider what is actually going
on (e.g., waking another thread does not generally require the kernel to
context switch to it immediately; not being able to acquire a mutex does not
mean the respective thread has to block using futexes right away; the scenario
plays out differently on a multiprocessor that is not oversubscribed; ...). 
The new condition variable implementation contains code comments that discuss
related aspects in some more detail.

Therefore, I'm closing this bug.  If you observe a potential performance
problem with the new condition variable, please open a new bug.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]