Massive performance loss from OS::sleep hack

[Daniel Eischen]

On Sat, 15 Sep 2007, Kris Kennaway wrote:

> Hi,
>
> I have been running the volano java benchmark 
> (http://www.volano.com/benchmarks.html) on an 8-core i386 system, and out of 
> the box jdk15 on FreeBSD performs extremely poorly.  The system is more than 
> 90% idle, and profiling shows that the ~800 threads in the benchmark are 
> spending most of their time doing short nanosleep() calls.
>
>
> I traced it to the following FreeBSD-specific hack in the jdk:
>
> // XXXBSD: understand meaning and workaround related to yield
> ...
> // XXXBSD: done differently in 1.3.1, take a look
> int os::sleep(Thread* thread, jlong millis, bool interruptible) {
>  assert(thread == Thread::current(),  "thread consistency check");
> ...
>
>  if (millis <= 0) {
>    // NOTE: workaround for bug 4338139
>    if (thread->is_Java_thread()) {
>      ThreadBlockInVM tbivm((JavaThread*) thread);
> // BSDXXX: Only use pthread_yield here and below if the system thread
> // scheduler gives time slices to lower priority threads when yielding.
> #ifdef __FreeBSD__
>      os_sleep(MinSleepInterval, interruptible);
> #else
>      pthread_yield();
> #endif
>
> When I removed this hack (i.e. revert to pthread_yield()) I got an immediate 
> 7-fold performance increase, which brings FreeBSD performance on par with 
> Solaris.
>
> What is the reason why this code is necessary?  Does FreeBSD's sched_yield() 
> really have different semantics to the other operating systems, or was this a 
> libkse bug that was being worked around?

It's certainly not a libkse bug, at least with scope process threads 
libkse does the right thinng.  For scope system threads and all libthr
threads, it probably depends on what scheduler you are using since
it's essentially a __sys_sched_yield().

On a side note, I think pthread_yield() is deprecated and not in the
latest POSIX spec.  sched_yield() is in the spec and is specified
to account for behavior in a threaded environment.

-- 
DE