I like this idea of a quick-and-dirty, self-contained benchmark to give a
rough idea of the relative (floating-point biased) processing power of a
machine. However, for readability and ease-of-comparison, how about quoting
it as the inverse, i.e.
%100 (6!:2) '%.200 200 [EMAIL PROTECTED]' NB. Looking only at CPU
which gives me a value of 40 on one machine at work and 16 on another.
Also, running dual simultaneous versions of this shows that the "40" machine
is dual core - each version returns a value of 34 - whereas the "16" machine
is single core - each version returns about 8.
A look at perfmon (MS Windows) indicates a substantial amount of page
faulting going on when I run this, so it's by no means a pure CPU test.
On 1/28/08, Arie Groeneveld <[EMAIL PROTECTED]> wrote:
>
> Following a wise advise: here's is an indication of the speed of my pc:
>
> 10 ts'%. 200 200 [EMAIL PROTECTED] 1000'
> 0.0865417 3933888
>
>
> @@i schreef:
> > JATK or let's say J4F.
> >
> > Generating a sequence of ordered perfect powers <= N
> > ref. OLEIS A001597
> >
> > I have a tacit (pmt) and a explicit (pmx) version both with the same
> > algorithm
> >
> > *nb* Without ext. prec.
> >
> > pmt 200
> > 1 4 8 9 16 25 27 32 36 49 64 81 100 121 125 128 144 169 196
> >
> > 10 ts 'pmt 2000000'
> > 0.006247 232448
> >
> > 10 ts 'pmx 2000000'
> > 0.0036827 275712
> >
> >
> > x:{: pmx 2e6
> > 1999396
> >
> > #pmt 1e6 NB. OLEIS A070428
> > 1111
> >
> >
> > Testing if a number > 1 is a perfect power:
> >
> > (1<[:+./[:{:__&q:) 220
> > 0
> >
> > (1<[:+./[:{:__&q:) 32
> > 1
> >
> >
> > Just anxious to know (JATK) what's possible. :-)
> >
> >
> >
> > =@@i
> >
> >
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
> >
> >
> >
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
--
Devon McCormick, CFA
^me^ at acm.
org is my
preferred e-mail
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
