ARM public key benchmark

[Torbjorn Granlund]

Richard Henderson <rth at twiddle.net> writes:

  On 2013-04-04 06:51, Niels Möller wrote:
  > And it's no use to even think of porting the loop mixer to arm without
  > access to cycle-accurate timing.
  
  Looking around the web it seems that what most folks do is write a
  minimal kernel module that toggles the bit that allows userspace
  access to the cycle counter MSRs.  They're generally supervisor only
  because then userspace *could* set up cycle counter timer interrupts.
  
  I havn't fiddled with this myself, because the A15 includes a real
  high-resolution clock, which is accessed via CLOCK_MONOTONIC.
  
  But it might be worth adding some code that looks for the MSR.  Just
  set up a SIGILL handler and try the read.
  
I tried these kernel modules, in fact I bravely tried several suggested
variants.  Not of them made cycle counters be accessible on my systems

-- 
Torbjörn