[Octave-bug-tracker] [bug #52858] Consider using C gnulib clock_gettime() for Octave cputime()

[Dan Sebald]

Follow-up Comment #1, bug #52858 (project octave):

Continuing from this post:

https://savannah.gnu.org/bugs/?52809#comment22

I've implemented the example loops with cputime and tic/toc as functions:


function tdiff=bm_for_loop()
a = 1; b = 1; t0=cputime; for i=1:1000; for j=1:1000; a = a + b + 123.0; end;
end; t1=cputime; tdiff = t1 - t0



function t1=bm_for_loop2()
a = 1; b = 1; t0=tic; for i=1:1000; for j=1:1000; a = a + b + 123.0; end;
end;
t1=toc(t0); t1


As results, I'm seeing:


octave:9> tl=zeros(20,1); for i=1:20; tl(i) = bm_for_loop; end;
tdiff =  2.7720
tdiff =  2.8040
tdiff =  2.7960
tdiff =  2.8000
tdiff =  2.8000
tdiff =  2.8000
tdiff =  2.8000
tdiff =  2.8000
tdiff =  2.8160
tdiff =  2.8000
tdiff =  2.8000
tdiff =  2.7960
tdiff =  2.7960
tdiff =  2.8000
tdiff =  2.8000
tdiff =  2.8040
tdiff =  2.7960
tdiff =  2.8160
tdiff =  2.7960
tdiff =  2.8000
octave:10> tl2=zeros(20,1); for i=1:20; tl2(i) = bm_for_loop2; end;
t1 =  2.7467
t1 =  2.8181
t1 =  2.8147
t1 =  2.8200
t1 =  2.8073
t1 =  2.8180
t1 =  2.8186
t1 =  2.8184
t1 =  2.8127
t1 =  2.8237
t1 =  2.8149
t1 =  2.8078
t1 =  2.8130
t1 =  2.8135
t1 =  2.8145
t1 =  2.8168
t1 =  2.8243
t1 =  2.8114
t1 =  2.8069
t1 =  2.8138
octave:11> mean(tl)
ans =  2.7996
octave:12> std(tl)
ans =  0.0085987
octave:13> mean(tl2)
ans =  2.8118
octave:14> std(tl2)
ans =  0.016044
octave:15> mean(tl2) - mean(tl)
ans =  0.012160


I don't think I can conclude there is a statistically significant difference
from that, given there could be slightly different CPU usage between
invocations.

I do believe though that there could be a difference in the above link,
statistically speaking, i.e.,


octave:1> bm_for_loop
ans =  4.6840
octave:2> bm_for_loop
ans =  4.6640
octave:3> bm_for_loop
ans =  4.6240
octave:4> bm_for_loop2
t1 =  4.4830
octave:5> bm_for_loop2
t1 =  4.5830
octave:6> bm_for_loop2
t1 =  4.4962


I wonder if the issue there is that, in fact, the test is physically taking
less time to operate than what cputime() reports because cputime() might
include a combination of CPU usage from more than one processor core.  That
is, perhaps some small aspect of this test can be shifted to a second CPU in a
multi-threaded environment.

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?52858>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/