tech-kern archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: x86: per-cpu tmp vas



hi,

On Sat, Jan 28, 2017 at 6:42 PM, Maxime Villard <max%m00nbsd.net@localhost> wrote:
I've updated my patch. The previous one was completely wrong. However, I'm not
going to commit it now.

For the record: in fact, I cannot measure the performance because of two issues
 - our PMCs do not support SMP, which is problematic since false-sharing occurs
   with at least two CPUs. I can't get the interesting measurement MSRs on my
   AMD.

tprof used to work fine on smp when i was using it actively.
no idea about your cpu model though.
 
 - time-based benchmarking does not work either for me; when performing 5x10^8
   calls to pmap_page_zero on two CPUs, several driver errors get printed on
   the screen midway and the system freezes. It seems to be related to my
   hardware, because I've already seen this kind of things in the past on that
   particular machine, regardless of the OS running on it.

Maxime



Home | Main Index | Thread Index | Old Index