I believe Victor is correct.

Here are my data for SR=44100, with different values for ksmps:

ksmps=10: csound CPU=98, total CPU=95
16: 65, 75
32: 27, 50
64: 24, 50
100: 21, 58
128: 21, 58
441: 20, 38

Powers of two for ksmps seem to be irrelevent. (Please note that all these 
data are highly approximate.)

Art Hunkins

----- Original Message ----- 
From: "victor" 
To: 
Sent: Tuesday, January 13, 2009 2:26 PM
Subject: [Csnd] Re: Re: Re: Re: SR and CPU Usage on the OLPC?


>I think the power-of-two issue is only to do with
> full-duplex on portaudio.
>
> Victor
> ----- Original Message ----- 
> From: "Steven Yi" 
> To: 
> Sent: Tuesday, January 13, 2009 6:31 PM
> Subject: [Csnd] Re: Re: Re: SR and CPU Usage on the OLPC?
>
>
>> Hi Art,
>>
>> Would you also try power-of-two for ksmps?  Maybe try 16, 32, and 64?
>> I don't remember exactly but there was a thread a long time ago that
>> suggested using power-of-two for ksmps yielding better results; I'm
>> not sure if I'm remembering that correctly or not, but I've been
>> following that for a while now.
>>
>> Thanks!
>> steven
>>
>> On Tue, Jan 13, 2009 at 9:51 AM, Art Hunkins  wrote:
>>> Victor,
>>>
>>> Thanks for this very helpful observation. It led to the solution to my
>>> problem. It indeed turns out that the control rate is considerably more
>>> important than SR in regard to CPU usage.
>>>
>>> Data:
>>> SR=44100, ksmps=10: csound CPU usage=98, overall CPU usage=95
>>> SR=36000, ksmps=10: csound CPU=61, overall CPU=71
>>> SR=36000, ksmps=100: csound CPU=4, overall CPU=5
>>>
>>> The latter is a *dramatic* difference - one I can live with. Finally,
>>> "sensekey tamed."
>>>
>>> (Any explanations greatly appreciated.)
>>>
>>> Thanks again, Victor.
>>>
>>> Art Hunkins
>>>
>>> ----- Original Message ----- From: "victor" 
>>> To: 
>>> Sent: Tuesday, January 13, 2009 3:57 AM
>>> Subject: [Csnd] Re: SR and CPU Usage on the OLPC?
>>>
>>>
>>>> I also noticed that changing the control rate also makes
>>>> a great deal of difference.
>>>>
>>>> Victor
>>>> ----- Original Message ----- From: "Art Hunkins" 
>>>> To: 
>>>> Sent: Tuesday, January 13, 2009 3:40 AM
>>>> Subject: [Csnd] SR and CPU Usage on the OLPC?
>>>>
>>>>
>>>>> Does anyone have an explanation for the following bizarre (but 
>>>>> verified)
>>>>> data regarding SR's and CPU usage on the OLPC? My test .csd is with
>>>>> ksmps=10, realtime, with MIDI input.
>>>>>
>>>>> SR=44100: Csound CPU usage = 26%, total CPU usage = 54%
>>>>> (SR=48000: 32% and 52%)
>>>>> SR=36000: 4% and 6%
>>>>> SR=32000: 3% and 5%
>>>>> SR=22050: <3% and 4%
>>>>>
>>>>> All conditions, to my knowledge, except for SR, were identical.
>>>>>
>>>>> It makes no sense to me, but the difference is real. I'll be using
>>>>> SR=32000 or 36000 instead of 44100.
>>>>>
>>>>> Whatever the reason, I'll now be able to port *some* of what I'd hoped 
>>>>> to
>>>>> the OLPC.
>>>>>
>>>>> Art Hunkins
>>>>>
>>>>>
>>>>> Send bugs reports to this list.
>>>>> To unsubscribe, send email sympa@lists.bath.ac.uk with body 
>>>>> "unsubscribe
>>>>> csound"
>>>>
>>>>
>>>>
>>>> Send bugs reports to this list.
>>>> To unsubscribe, send email sympa@lists.bath.ac.uk with body 
>>>> "unsubscribe
>>>> csound"
>>>
>>>
>>>
>>> Send bugs reports to this list.
>>> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
>>> csound"
>>>
>>
>>
>> Send bugs reports to this list.
>> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe 
>> csound"
>
>
>
> Send bugs reports to this list.
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe 
> csound"