I'm looking at the code and it seems that the function that rtclock
opcode calls and the one that one subsequently calls has comments
about resolution on Win95.  The time resolution stuff in this code is
all very strange.  (This is all in Top/csound.c).  It seems to be
coded in a way such that depending on what you have available on your
system, rtclock may report results in seconds or microseconds.

Now, in this code is a check for HAVE_RDTSC which I have not seen in
the codebase before. I don't know if that is a symbol automatically
defined by systems or if that is something that is supposed to be
checked by SConstruct and passed in.

Hopefully someone else could shed some light on the topic!
steven


On 9/8/07, Victor Lazzarini <victor.lazzarini@nuim.ie> wrote:
> I reckon the man is wrong. On Windows I am getting
> similar results to yours.
>
> Victor
>
> >
> >
> > hi -
> >
> > i tested the manual example for rtclock ...
> >
> > instr 1
> >    ; Get the system time.
> >    k1 rtclock
> >    ; Print it once per second.
> >    printk 1, k1
> > endin
> >
> > ; Play Instrument #1 for two seconds.
> > i 1 0 2
> > e
> >
> > ..and expected something like this (from the manual
> > example)
> >
> >   i   1 time     0.00002: 1018236096.00000
> >   i   1 time     1.00002: 1018236224.00000
> >
> >
> > but i get with output to dac:
> >
> >   i   1 time     0.00002:     0.01464
> >   i   1 time     1.00000:     1.01257
> >   i   1 time     2.00000:     2.01394
> >
> > and with output to file:
> >
> >   i   1 time     0.00002:     0.09404
> >   i   1 time     1.00000:     0.27993
> >   i   1 time     2.00000:     0.33443
> >
> > so it seems to give cpu time, right? is this the correct
> > behavior for   rtclock, or is this a bug (at least under
> > os x)? is there an other   way of getting the universal
> > time in seconds?
> >
> > thanks for help -
> >
> >     joachim
> >
> --
> Send bugs reports to this list.
> To unsubscribe, send email to csound-unsubscribe@lists.bath.ac.uk
>