Csound Csound-dev Csound-tekno Search About

Re: [Csnd] Cpu usage of opcodes

Date2011-08-03 11:00
FromTarmo Johannes
SubjectRe: [Csnd] Cpu usage of opcodes
Great! Thanks!
One day would be really nice to have such info about all (or most) opcodes. Perhaps really some student or volunteer...
What do the negative values mean?
Tarmojpff@cs.bath.ac.uk kirjutas:>
>Found them
>Need some care in use!
>
>*.ak:           0       6       15
>*.ka:           0       6       15
>*.kk:           0       0       9
>+.aa:           0       6       16
>+.kk:            0      0       9
>/.kk:           0       0       9
>=.i             7
>=.k:            0       0       7
>init:           0       4       12
>arand:          52      -0.351226       188.512264
>balance:        38      2.351111        274.488889
>buzz:           75      -0.639775       430.385596
>cpspch:         106.368056
>delay:          4097    -0.340018       157.403009
>expon.k:        178.553846      0        9
>expseg.a:       669.121622     5.000011 37
>expseg.k:       669.121622     0.000031 26
>foscil:         81      29       72
>frac.i:         65
>gbuzz:          77      64.000105 125.499682
>int.i:          67
>line.k:         20             0.000762 8.992379
>linen.k:        62             0.000004 20.650515
>linseg.k:       266.971564     0.000559 29.994448
>octpch.1:       74
>oscil.ka:       69      17       46
>oscil.kk:       69      0        41
>oscil.kk:       69      12       47
>oscili.kk:      69      21       49
>outs:           0       9.563279 21.655083
>phasor.a:       38      15.979998 32.000007
>phasor.k:       38      -0.000798 27
>randh.k:        60      0.058034 34.004509
>randi.k:        76      0.015561 35.999951
>rassign:        2
>reson:          39      13.002546 99.545002
>reverb:         6963.481481       77.038957 157.999993
>subkk:          0       0        9
>table.a:        93      23.062626 43.997796
>table.k:        93      0        45
>tablei.k:       93      0        61
>upsamp:         0       4        12
>
>ags:            497     917.921440 79475.154651
>nlfilt:         1068.813496     52.985889 100.158886
>vdelay:         2073.886076     53.572573 69.049966
>butterlp:       19       29.005359 45.478359
>butterhi:       19       30.000387 35
>butterbp:       20       30     71
>zar:            14       1      41
>aasign:         0        1      15
>zacl:           14      51      54
>/.ak:           0       6       20
>zaw:            14      1.000018 38.998197
>integrate:      13      5       17
>zkr:            14      0       32
>zkw:            14      0       31.999964
>zkcl:           14      0       101
>reinit:         0       0       16335.400000
>rms:            32      10      21
>sqrt.i:         8
>sqrt.k:         0       0       8
>exp.i:          19.412868
>exp.k:          0       0       19.412868
>!=.i:           13
>!=.k:           0       0       13
>timout:         40     -0.002915 15.235836
>kcgoto:         0      0.000022 8.758767
>kgoto:          0       0       7
>bilbar:         371.5   1856.028046 85.999807
>
>
>And yes, Oevind is rioght that there is parameter dependency... which is
>why not all the counts are integer
>==John ff
>
>>> I think it would be possible to add to Csound to instrument the opcode
>>> routines with a cpu time log that would be sent to a socket or written
>>> to a file.
>>>
>>> It might also be possible to simply use the logs produced by a
>>> profiler to do pretty much exactly the same thing.
>>>
>>> Probably what John means by time consuming is simply setting up the
>>> opcodes to get called. That could be daunting, given the number of
>>> opcodes!  Perhaps it would be possible to scan the opcode type dump
>>> and programmatically generate a Csound orchestra that would call
>>> opcodes with dummy arguments and discard the results. That ought to be
>>> enough to get timings. But I suspect writing such a script might be
>>> harder than it seems.
>>
>> Apologies for short messages -- having a connectivity problem here at ICMC
>>
>> My numbers come from valgrind, and the time consuming stuff is simply that
>> the tool that decodes the counts does not allow for cut-n-paste so the
>> numbers have to be copied.  I run twice, with different k-rates so i can
>> solve the equatuons for k-rate and a-rate instructions.  Thtat part of the
>> program is easy.  It is decodeing the valgrand raw data that has flawed me
>> so far.  could not find a student willing to do this as a project.
>>
>> I ned hese numbers for load-balance in ParCS
>>
>> back to fighting networks & ssh
>>
>> ==John
>>
>>
>>
>> Send bugs reports to the Sourceforge bug tracker
>>             https://sourceforge.net/tracker/?group_id=81968&atid=564599
>> Discussions of bugs and features can be posted here
>> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
>> csound"
>>
>>
>>
>>
>
>
>
>
>Send bugs reports to the Sourceforge bug tracker
>            https://sourceforge.net/tracker/?group_id=81968&atid=564599
>Discussions of bugs and features can be posted here
>To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"
>

Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"


Date2011-08-03 11:38
FromPekonen Jussi
SubjectRe: [Csnd] Cpu usage of opcodes

On Aug 3, 2011, at 13:00 , Tarmo Johannes wrote:

One day would be really nice to have such info about all (or most) opcodes. Perhaps really some student or volunteer...

Great idea Tarmo. This could easily become a nice and interesting thesis work for someone (Victor? John ff? One of your students?). When these benchmarking results are out, folks can (more) easily figure out what the heck in their instruments is consuming all computing power.

Jussi Pekonen




Date2011-08-03 12:09
Fromjpff@cs.bath.ac.uk
SubjectRe: [Csnd] Cpu usage of opcodes
> Great! Thanks!
> One day would be really nice to have such info about all (or most)
> opcodes. Perhaps really some student or volunteer...
> What do the negative values mean?
> Tarmojpff@cs.bath.ac.uk kirjutas:>

Possibly a glitch in the formalism or small variation.  You will note that
all the negative values are small, and probably should be supressed to
zero.  Indeed that is one of the reasons i have not distributed the list
before -- but following the discussions.....

==John ff



Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"