Bug #663: Benchmark.measure outputs different result when executed using command line "ruby -e ..."

• Status: Rejected
• Assignee: -
• Target version: -

[greck (Artem Vorozhtsov)]

=begin
I've tested it on :

• ruby 1.8.6 (2008-03-03 patchlevel 114) [i686-linux]
• ruby 1.8.5 (2006-08-25) [i386-linux]

Sorry, I've not tested it against 1.9 and 1.8.7.

[artem@dev bm_test]$ cat a.c
int main() {
int i = 0, k = 0, n = 1000000000;
for(i = 0 ; i< n; i++){ k += i*i; }
return k;
}
[artem@dev bm_test]$ make a
cc a.c -o a
[artem@dev bm_test]$ ruby -rbenchmark -e "puts Benchmark.measure{ ./a }"
0.000000 0.000000 0.000000 ( 0.000011)
[artem@dev bm_test]$ echo 'puts Benchmark.measure{ ./a }' > b.rb
[artem@dev bm_test]$ cat b.rb
puts Benchmark.measure{ ./a }
[artem@dev bm_test]$ ruby -rbenchmark b.rb
0.000000 0.000000 4.660000 ( 4.652781)

Bugs:

1. "ruby -e" : "total" is zero
2. "ruby -e" : "real" is too small (should be ~ 4 seconds)
3. "ruby" : "total" is not equal to "user" + "system" in the last example.
   =end

[mame (Yusuke Endoh)]

=begin
Hi,

Thank you for your report.

2008/10/19 Artem Vorozhtsov [e-mail]:

Bugs:

1. "ruby -e" : "total" is zero
2. "ruby -e" : "real" is too small (should be ~ 4 seconds)

I think ./a is executed by your shell. If so, it's not a bug of ruby.

$ ruby19 -rbenchmark -e "puts Benchmark.measure { ./a }"
0.000000 0.000000 0.000000 ( 0.000000)

$ ruby19 -rbenchmark -e 'puts Benchmark.measure { ./a }'
0.000000 0.000000 7.540000 ( 7.540114)

3. "ruby" : "total" is not equal to "user" + "system" in the last example.

While "user" and "system" show CPU time that only the process comsumed,
"total" shows CPU time that both the process and its children consumed.
I don't know whether it is intended.

You can see the user and system CPU time that both consumed in the
following way:

$ ruby19 -rbenchmark -e 'puts Benchmark.measure { ./a
}.format("%10.6U %10.6Y %10.6t %10.6r\n")'
6.460000 0.000000 6.460000 ( 6.460097)

--
Yusuke ENDOH [e-mail]

=end

[shyouhei (Shyouhei Urabe)]

• Status changed from Open to Rejected

=begin

=end

[greck (Artem Vorozhtsov)]

=begin
2008/10/19 Yusuke ENDOH [e-mail]:

Hi,

Thank you for your report.

2008/10/19 Artem Vorozhtsov [e-mail]:

Bugs:

1. "ruby -e" : "total" is zero
2. "ruby -e" : "real" is too small (should be ~ 4 seconds)

I think ./a is executed by your shell. If so, it's not a bug of ruby.

$ ruby19 -rbenchmark -e "puts Benchmark.measure { ./a }"
0.000000 0.000000 0.000000 ( 0.000000)

$ ruby19 -rbenchmark -e 'puts Benchmark.measure { ./a }'
0.000000 0.000000 7.540000 ( 7.540114)

:)

I'm really sorry.

3. "ruby" : "total" is not equal to "user" + "system" in the last example.

While "user" and "system" show CPU time that only the process comsumed,
"total" shows CPU time that both the process and its children consumed.
I don't know whether it is intended.

You can see the user and system CPU time that both consumed in the
following way:

$ ruby19 -rbenchmark -e 'puts Benchmark.measure { ./a
}.format("%10.6U %10.6Y %10.6t %10.6r\n")'
6.460000 0.000000 6.460000 ( 6.460097)

--
Yusuke ENDOH [e-mail]

Thanks!

Then documentation should be fixed.
At http://www.ruby-doc.org/stdlib/libdoc/benchmark/rdoc/classes/Benchmark.html
one could read
"This report shows the user CPU time, system CPU time, the sum of the
user and system CPU times, and the elapsed real time. The unit of time
is seconds. "

=end