Date: Thu, 30 May 2013 12:27:17 -0500 From: Alan Cox <alc@rice.edu> To: Zbigniew Bodek <zbb@semihalf.com> Cc: freebsd-arm@freebsd.org, Alan Cox <alc@cs.rice.edu>, Aleksandr Rybalko <ray@freebsd.org> Subject: Re: pmap_copy for ARMv6/v7 Message-ID: <51A78BF5.5000901@rice.edu> In-Reply-To: <eb88997fc774ddc47ca0b44e3ef2dbd8@smtp.semihalf.com> References: <eb88997fc774ddc47ca0b44e3ef2dbd8@smtp.semihalf.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 05/28/2013 17:17, Zbigniew Bodek wrote: > Hello Everyone, > > I would like to share with you the patch adding pmap_copy() > functionality to pmap-v6.c. This is the last but not least of > the pmap improvements done as a part of Semihalf work on > Superpages support. > > We tried various technics to test the patch on our target. > Apart from obvious long term stability tests and others, > we used a simple benchmark to measure fork system call time and overhead. > > The program can be downloaded from here: > http://berlin.ccc.de/~packet/fork_test.c > > A single fork time has increased as suspected (the program is not measuring > a real-life fork situation where child process is actually doing something) > but the average system overhead has decreased. > For 10000 fork(), 10000 vfork() and 10000 pthread_create() calls we observe > 100000 less vm_faults and shorter overall test time. > > If you have any suggestions or have an idea for benchmarking this > feature please share. >From what you describe above, I think that you've done a reasonable job of testing. The performance results are in line with what I would expect. I wouldn't worry too much about obtaining further test results. > We will appreciate if you could test this patch on your ARM platforms and > send > us your review/feedback. > The patch looks fine. Alan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?51A78BF5.5000901>