That's not actually a problem, as I can use plughw:0,0. The issue
was the train of dropouts I was getting even if the CPU was reasonably
at ease.

I have been running other less heavy orchestras and the performance
seems OK, I was able to go down to -b8 -B80 running a phase vocoder
harmoniser with a GUI (I would get the occasional underrun if I start
clicking too much on other desktop items). There is no generally noticeable
latency.

I'm surprised with the demix, I thought I'd be able to run it in that machine.

Victor

At 16:20 13/01/2006, you wrote:
>If it's the sample format, it could be the low-level driver 
>(manufacturer's driver, sound card). Try a different format. Are you 
>trying float? Then try 16 bit int. Some cards only do certain sample rates 
>also.
>
>Regards,
>Mike
>
>-----Original Message-----
> >From: Victor Lazzarini 
> >Sent: Jan 13, 2006 11:02 AM
> >To: csound-devel@lists.sourceforge.net
> >Subject: Re: [Cs-dev] Linux performance
> >
> >I tried -o dac:hw:0 and I could not run it, as csound
> >says "unable to set sample format requested",
> >whereas with -o dac:plughw:0 it runs as before.
> >
> >The same orc runs on my Windows machine OK (with
> >-b1024).
> >
> >What kernel are you running? Do you have a low-latency patch?
> >
> >Victor
> >
> >At 15:50 13/01/2006, you wrote:
> >>On Friday 13 January 2006 16:40, Victor Lazzarini wrote:
> >>
> >> > I used the default outdevice (-odac) and anything from
> >> > -b64 -B512 up to -b2048 -B4096
> >>
> >>Try using -o dac:hw:0 with -+rtaudio=alsa (-o dac:hw:1 for
> >>the second card, etc.). Also, PVOC with a large FFT size may
> >>not be friendly to low latency real time audio, due to the
> >>uneven CPU usage.
> >>
> >>
> >>-------------------------------------------------------
> >>This SF.net email is sponsored by: Splunk Inc. Do you grep through log 
> files
> >>for problems?  Stop!  Download the new AJAX search engine that makes
> >>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> >>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> >>_______________________________________________
> >>Csound-devel mailing list
> >>Csound-devel@lists.sourceforge.net
> >>https://lists.sourceforge.net/lists/listinfo/csound-devel
> >
> >Victor Lazzarini
> >Music Technology Laboratory
> >Music Department
> >National University of Ireland, Maynooth
> >
> >
> >
> >-------------------------------------------------------
> >This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
> >for problems?  Stop!  Download the new AJAX search engine that makes
> >searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> >http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> >_______________________________________________
> >Csound-devel mailing list
> >Csound-devel@lists.sourceforge.net
> >https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Csound-devel mailing list
>Csound-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

In fact, with regards to that orc,  OSX 10.4 machines (comparable spec)
run it OK as well. So I don't really know. I expected Linux to beat
them all.

Victor

At 16:32 13/01/2006, you wrote:
>That's not actually a problem, as I can use plughw:0,0. The issue
>was the train of dropouts I was getting even if the CPU was reasonably
>at ease.
>
>I have been running other less heavy orchestras and the performance
>seems OK, I was able to go down to -b8 -B80 running a phase vocoder
>harmoniser with a GUI (I would get the occasional underrun if I start
>clicking too much on other desktop items). There is no generally noticeable
>latency.
>
>I'm surprised with the demix, I thought I'd be able to run it in that machine.
>
>Victor
>
>At 16:20 13/01/2006, you wrote:
>>If it's the sample format, it could be the low-level driver 
>>(manufacturer's driver, sound card). Try a different format. Are you 
>>trying float? Then try 16 bit int. Some cards only do certain sample 
>>rates also.
>>
>>Regards,
>>Mike
>>
>>-----Original Message-----
>> >From: Victor Lazzarini 
>> >Sent: Jan 13, 2006 11:02 AM
>> >To: csound-devel@lists.sourceforge.net
>> >Subject: Re: [Cs-dev] Linux performance
>> >
>> >I tried -o dac:hw:0 and I could not run it, as csound
>> >says "unable to set sample format requested",
>> >whereas with -o dac:plughw:0 it runs as before.
>> >
>> >The same orc runs on my Windows machine OK (with
>> >-b1024).
>> >
>> >What kernel are you running? Do you have a low-latency patch?
>> >
>> >Victor
>> >
>> >At 15:50 13/01/2006, you wrote:
>> >>On Friday 13 January 2006 16:40, Victor Lazzarini wrote:
>> >>
>> >> > I used the default outdevice (-odac) and anything from
>> >> > -b64 -B512 up to -b2048 -B4096
>> >>
>> >>Try using -o dac:hw:0 with -+rtaudio=alsa (-o dac:hw:1 for
>> >>the second card, etc.). Also, PVOC with a large FFT size may
>> >>not be friendly to low latency real time audio, due to the
>> >>uneven CPU usage.
>> >>
>> >>
>> >>-------------------------------------------------------
>> >>This SF.net email is sponsored by: Splunk Inc. Do you grep through log 
>> files
>> >>for problems?  Stop!  Download the new AJAX search engine that makes
>> >>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>> >>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>> >>_______________________________________________
>> >>Csound-devel mailing list
>> >>Csound-devel@lists.sourceforge.net
>> >>https://lists.sourceforge.net/lists/listinfo/csound-devel
>> >
>> >Victor Lazzarini
>> >Music Technology Laboratory
>> >Music Department
>> >National University of Ireland, Maynooth
>> >
>> >
>> >
>> >-------------------------------------------------------
>> >This SF.net email is sponsored by: Splunk Inc. Do you grep through log 
>> files
>> >for problems?  Stop!  Download the new AJAX search engine that makes
>> >searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>> >http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>> >_______________________________________________
>> >Csound-devel mailing list
>> >Csound-devel@lists.sourceforge.net
>> >https://lists.sourceforge.net/lists/listinfo/csound-devel
>>
>>
>>
>>
>>
>>-------------------------------------------------------
>>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>>for problems?  Stop!  Download the new AJAX search engine that makes
>>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>>_______________________________________________
>>Csound-devel mailing list
>>Csound-devel@lists.sourceforge.net
>>https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>Victor Lazzarini
>Music Technology Laboratory
>Music Department
>National University of Ireland, Maynooth
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Csound-devel mailing list
>Csound-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

> I have been running other less heavy orchestras and the performance
> seems OK, I was able to go down to -b8 -B80 running a phase vocoder
> harmoniser with a GUI (I would get the occasional underrun if I start
> clicking too much on other desktop items). There is no generally noticeable
> latency.

You should be able to fix that with --sched too. I can pretty much do 
anything without clicks if I use flux box, and I have reniced X before 
too but with the Molnar patch that's not even necessary. I put Istvan's 
scheduling code into my host app and used varying pthread priorities and 
it works like a hot damn. I can send you code if you would like to see 
how the threading and scheduling was handled. I have my fltk gui moving 
128 knobs and updating 128 graphic windows with latency still dead low, 
no clicks. Thanks Istvan! Thanks Ingo!

Iain




-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I could not find Ingo's patch on linux-kernels.org. I put the latest kernel
and patched it with Andrew Morton's. Not really any noticeable difference
with the Core 4-shipped kernel.

Is there any specific Kernel configuration you might suggest? I would prefer
to have a low latency kernel than to have to nice X etc. What's a flux box?

Victor

At 19:44 13/01/2006, you wrote:

>>I have been running other less heavy orchestras and the performance
>>seems OK, I was able to go down to -b8 -B80 running a phase vocoder
>>harmoniser with a GUI (I would get the occasional underrun if I start
>>clicking too much on other desktop items). There is no generally noticeable
>>latency.
>
>You should be able to fix that with --sched too. I can pretty much do 
>anything without clicks if I use flux box, and I have reniced X before too 
>but with the Molnar patch that's not even necessary. I put Istvan's 
>scheduling code into my host app and used varying pthread priorities and 
>it works like a hot damn. I can send you code if you would like to see how 
>the threading and scheduling was handled. I have my fltk gui moving 128 
>knobs and updating 128 graphic windows with latency still dead low, no 
>clicks. Thanks Istvan! Thanks Ingo!
>
>Iain
>
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Csound-devel mailing list
>Csound-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Can you try this CSD in your system? Just to see if it runs with no
dropouts. Change "jarret1.wav" to any stereo file you have at hand.

Here it goes:


-d



sr=44100
kr=441
ksmps=100
nchnls=2

gip init 100

FLpanel "demix", 400,220, 400,100
iv1 FLvalue  " ", 40,20,340 ,10
iv2 FLvalue  " ", 40,20,340 ,60
iv3 FLvalue  " ", 40,20,340 ,110
gk1, ih1 FLslider  "azimuth discrimination",-1,1,0,5,iv1,300,20,10,10
gk2, ih2 FLslider  "subspace width",1,gip,0,5,iv2,300,20,10,60
gk3, ih3 FLslider   "threshold",0,90,0,5,iv3,300,20,10,110
gk4, ih4 FLslider   "gain",0,2,0,5,-1,300,20,10,160
gk5, ih5 FLbutton   "bypass",0,1, 2, 64, 28, 330,160,-1

FLsetVal_i -0.75, ih1
FLsetVal_i 25, ih2
FLsetVal_i 0, ih3
FLsetVal_i 1.5, ih4
FLpanelEnd

FLrun

instr 1

ifftsize = 2048
iwtype = 1    /* cleaner with hanning window */
kpos = gk1
kwidth = gk2

al,ar  diskin "jarrett1.wav", 1, 0, 1
flc  pvsanal   al, ifftsize, ifftsize/4, ifftsize, iwtype
frc  pvsanal   ar, ifftsize, ifftsize/4, ifftsize, iwtype
fdm  pvsdemix  flc, frc, kpos, kwidth,gip /* mystery opcode */
fst  pvstencil fdm, 0, ampdb(gk3), 1
adm  pvsynth   fst
aout = adm*gk4*(gk5) + (al+ar)*0.5*(1-gk5)
        outs    aout,aout

endin



f1 0 1025 7  0.01 1024 1
i1 0 360








At 19:44 13/01/2006, you wrote:

>>I have been running other less heavy orchestras and the performance
>>seems OK, I was able to go down to -b8 -B80 running a phase vocoder
>>harmoniser with a GUI (I would get the occasional underrun if I start
>>clicking too much on other desktop items). There is no generally noticeable
>>latency.
>
>You should be able to fix that with --sched too. I can pretty much do 
>anything without clicks if I use flux box, and I have reniced X before too 
>but with the Molnar patch that's not even necessary. I put Istvan's 
>scheduling code into my host app and used varying pthread priorities and 
>it works like a hot damn. I can send you code if you would like to see how 
>the threading and scheduling was handled. I have my fltk gui moving 128 
>knobs and updating 128 graphic windows with latency still dead low, no 
>clicks. Thanks Istvan! Thanks Ingo!
>
>Iain
>
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Csound-devel mailing list
>Csound-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I got

...
ftable 1:
new alloc for instr 1:
Csound tidy up: Segmentation fault
inactive allocs returned to freespace
end of score.              overall amps:      0.0      0.0
           overall samples out of range:        0        0
0 errors in performance
Elapsed time at end of performance: real: 0.563s, CPU: 0.010s
0 1024-byte soundblks of shorts written to dac
end of widget thread
*** glibc detected *** free(): invalid pointer: 0x08122628 ***
Csound tidy up: Aborted

It flashed up a dialog window before crashing

==John ffitch


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

It turns out to be deep in libsndfile -- on the 32bit machine.  The
crash was in diskin.

==John ffitch


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

jpff@codemist.co.uk wrote:

> It turns out to be deep in libsndfile -- on the 32bit machine.  The
> crash was in diskin.

Do you have a test case for that?

Erik
-- 
+-----------------------------------------------------------+
  Erik de Castro Lopo
+-----------------------------------------------------------+
"The earth is degenerating these days. Bribery and corruption abound.
Children no longer mind parents ...and it is evident that the end of
the world is approaching fast." -- Assyrian Tablet Engraved in 2800 B.C.


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

With a recompiled libsndfile I am getting a few dropouts, but not
much.  This is a 1.4GHz Centrino on a 32bit laptop -- so the sound is
not very good and I may be not hearing the dropouts....
==John ffitch


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Sorry Erik; I recompiled 1.0.13pre6 and all was OK.  I suspect that
there was a grade-A foul-up of headers versus libraries.....all my
fault and no reflection on libsndfile.
==John ffitch


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

If you click on bypass you will get the original audio
without the de-mixing, but with the computation, that'll
probably help to check for droupouts. Also, Istvan's alsa
module reports overruns to the console; Do you get many
of these? What about buffer settings?

Victor

At 12:19 16/01/2006, you wrote:
>With a recompiled libsndfile I am getting a few dropouts, but not
>much.  This is a 1.4GHz Centrino on a 32bit laptop -- so the sound is
>not very good and I may be not hearing the dropouts....
>==John ffitch
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Csound-devel mailing list
>Csound-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I was using alsa outofthebox, no fiddling with B or b; no overrun
messages, just a few suspect sounds.
  Should different settings of the sliders make a difference to
computation?
  I will try on the AMD3200+ when I am next home.  This machine is
1GHz desktop but the sound plain does not work (Hamerfall Lite
problems) and no speakers.  Seems to have the characteristic sound of
failure to keep up, even with bypass set.

==John ffitch


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Using gnome; I can't see that it would make that much
difference, but I'll try fluxbox.

Victor
At 15:56 16/01/2006, you wrote:
>On Mon, Jan 16, 2006 at 09:54:50AM +0000, Victor Lazzarini wrote:
> > Is there any specific Kernel configuration you might suggest? I would 
> prefer
> > to have a low latency kernel than to have to nice X etc. What's a flux box?
>
>What wm (window manager) are you using? Fluxbox is a light weight wm. see 
>-> http://fluxbox.sourceforge.net/
>You are not using kde or gnome are you. If you are that is probably your 
>problem. Those two desktops are nice for doing normal computer and office
>type work but for realtime work they can be you enemy. What other system 
>services are currently running? Are they all necessary?
>
>Jeremiah
>
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Csound-devel mailing list
>Csound-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Tried fluxbox and found no significant difference.

Victor

At 16:39 16/01/2006, you wrote:
>Using gnome; I can't see that it would make that much
>difference, but I'll try fluxbox.
>
>Victor
>At 15:56 16/01/2006, you wrote:
>>On Mon, Jan 16, 2006 at 09:54:50AM +0000, Victor Lazzarini wrote:
>> > Is there any specific Kernel configuration you might suggest? I would 
>> prefer
>> > to have a low latency kernel than to have to nice X etc. What's a flux 
>> box?
>>
>>What wm (window manager) are you using? Fluxbox is a light weight wm. see 
>>-> http://fluxbox.sourceforge.net/
>>You are not using kde or gnome are you. If you are that is probably your 
>>problem. Those two desktops are nice for doing normal computer and office
>>type work but for realtime work they can be you enemy. What other system 
>>services are currently running? Are they all necessary?
>>
>>Jeremiah
>>
>>
>>
>>-------------------------------------------------------
>>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>>for problems?  Stop!  Download the new AJAX search engine that makes
>>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>>_______________________________________________
>>Csound-devel mailing list
>>Csound-devel@lists.sourceforge.net
>>https://lists.sourceforge.net/lists/listinfo/csound-devel
>
>Victor Lazzarini
>Music Technology Laboratory
>Music Department
>National University of Ireland, Maynooth
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>_______________________________________________
>Csound-devel mailing list
>Csound-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Istvan Varga wrote:
> On Monday 16 January 2006 11:28, Istvan Varga wrote:
> 
> 
>>On Monday 16 January 2006 11:05, Victor Lazzarini wrote:
>>
>>
>>>Can you try this CSD in your system? Just to see if it runs with no
>>>dropouts. Change "jarret1.wav" to any stereo file you have at hand.
>>
>>I cannot run it in real time on my 1300 MHZ CPU (it would need more
>>than 100% CPU time), but after setting sr = 32000, it works fine
>>with -b 64 -B 512. This is with the RC2 RPM build from Dec 30 2005;
>>I do not recall any changes since then that would have resulted in
>>a significant slowdown.
> 
> 
> Compiled with icc instead of gcc, it now runs at the original sample
> and control rates using --sched -b 64 -B 512 with occasional dropouts,
> With -B 1024 it is OK even without --sched, and with --sched I can
> increase the sample rate to 48000.

What is icc? Is it free? Should I recompile csound5 differently to get 
best real time performance?

Thanks
Iain


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

Hi Victor, I can get it down to -b256 -B512 but no lower, with ksmp=64, 
100, or 128. It works ok lower for a bit, but moving the controls causes 
underruns. This uses about 75% cpu of an AMD 3000, so it's a hungry one.

I am using 2.6.13 with Ingo Desktop Preempt patch. I have been told that 
2.6.14 is a big improvement, and an Ingo patched 2.6.14 even better. 
Fluxbox mostly helps because the footprint of opening new windows and 
stuff is lower, so it's nice if you are going to be using fltk windows 
or doing other stuff.

Also, I noticed that on the build I am using, --sched does NOT work when 
called from the csd options, and there is no error msg to indicate so. 
It just doesn't do anything, I can't watch top or switch windows without 
massive underruns.I broke your code into an orc/sco pair and then was 
able to get better performance by using --sched and --experession-opt. 
Anyone know is this issue has been fixed? I tried a cvs update today and 
can't log in for some reason, so I do not know how out of date my build is.

Hope that helps,
Iain

Victor Lazzarini wrote:
> Can you try this CSD in your system? Just to see if it runs with no
> dropouts. Change "jarret1.wav" to any stereo file you have at hand.
> 
> Here it goes:
> 
> 
> -d
> 
> 
> 
> sr=44100
> kr=441
> ksmps=100
> nchnls=2
> 
> gip init 100
> 
> FLpanel "demix", 400,220, 400,100
> iv1 FLvalue  " ", 40,20,340 ,10
> iv2 FLvalue  " ", 40,20,340 ,60
> iv3 FLvalue  " ", 40,20,340 ,110
> gk1, ih1 FLslider  "azimuth discrimination",-1,1,0,5,iv1,300,20,10,10
> gk2, ih2 FLslider  "subspace width",1,gip,0,5,iv2,300,20,10,60
> gk3, ih3 FLslider   "threshold",0,90,0,5,iv3,300,20,10,110
> gk4, ih4 FLslider   "gain",0,2,0,5,-1,300,20,10,160
> gk5, ih5 FLbutton   "bypass",0,1, 2, 64, 28, 330,160,-1
> 
> FLsetVal_i -0.75, ih1
> FLsetVal_i 25, ih2
> FLsetVal_i 0, ih3
> FLsetVal_i 1.5, ih4
> FLpanelEnd
> 
> FLrun
> 
> instr 1
> 
> ifftsize = 2048
> iwtype = 1    /* cleaner with hanning window */
> kpos = gk1
> kwidth = gk2
> 
> al,ar  diskin "jarrett1.wav", 1, 0, 1
> flc  pvsanal   al, ifftsize, ifftsize/4, ifftsize, iwtype
> frc  pvsanal   ar, ifftsize, ifftsize/4, ifftsize, iwtype
> fdm  pvsdemix  flc, frc, kpos, kwidth,gip /* mystery opcode */
> fst  pvstencil fdm, 0, ampdb(gk3), 1
> adm  pvsynth   fst
> aout = adm*gk4*(gk5) + (al+ar)*0.5*(1-gk5)
>        outs    aout,aout
> 
> endin
> 
> 
> 
> f1 0 1025 7  0.01 1024 1
> i1 0 360
> 
> 
> 
> 
> 
> 
> 
> 
> At 19:44 13/01/2006, you wrote:
> 
>>> I have been running other less heavy orchestras and the performance
>>> seems OK, I was able to go down to -b8 -B80 running a phase vocoder
>>> harmoniser with a GUI (I would get the occasional underrun if I start
>>> clicking too much on other desktop items). There is no generally 
>>> noticeable
>>> latency.
>>
>>
>> You should be able to fix that with --sched too. I can pretty much do 
>> anything without clicks if I use flux box, and I have reniced X before 
>> too but with the Molnar patch that's not even necessary. I put 
>> Istvan's scheduling code into my host app and used varying pthread 
>> priorities and it works like a hot damn. I can send you code if you 
>> would like to see how the threading and scheduling was handled. I have 
>> my fltk gui moving 128 knobs and updating 128 graphic windows with 
>> latency still dead low, no clicks. Thanks Istvan! Thanks Ingo!
>>
>> Iain
>>
>>
>>
>>
>> -------------------------------------------------------
>> This SF.net email is sponsored by: Splunk Inc. Do you grep through log 
>> files
>> for problems?  Stop!  Download the new AJAX search engine that makes
>> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
>> _______________________________________________
>> Csound-devel mailing list
>> Csound-devel@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 
> 
> Victor Lazzarini
> Music Technology Laboratory
> Music Department
> National University of Ireland, Maynooth
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log 
> files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Csound-devel mailing list
> Csound-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/csound-devel
> 


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

What does the Intel compiler do for AMD processors?
==John ffitch


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

I am setting up a new linux laptop for my own use and
I am wondering what you would suggest as a *reasonably priced*
soundcard for it, which is well supported by Alsa.

I have some m-audio USB cards around here, but they don't
really work with Alsa. So I'm wondering about other USB cards
or perhaps FireWire (although last time I looked these weren't
yet supported).

Any suggestions welcome.


Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net

In fact, I have to go back on what I have just said. Plugged in
the m-audio audiophile USB, booted up Fedora Core 4 and
it's working.

I could not get this to work on a desktop machine by fiddling around
with loading the snd-usb-audio module, and here it just works
straight out of the box. So there you go...

Victor
At 19:05 17/01/2006, you wrote:
>I am setting up a new linux laptop for my own use and
>I am wondering what you would suggest as a *reasonably priced*
>soundcard for it, which is well supported by Alsa.
>
>I have some m-audio USB cards around here, but they don't
>really work with Alsa. So I'm wondering about other USB cards
>or perhaps FireWire (although last time I looked these weren't
>yet supported).
>
>Any suggestions welcome.
>
>
>Victor Lazzarini
>Music Technology Laboratory
>Music Department
>National University of Ireland, Maynooth
>
>
>-------------------------------------------------------
>This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
>for problems?  Stop!  Download the new AJAX search engine that makes
>searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
>http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
>_______________________________________________
>Csound-devel mailing list
>Csound-devel@lists.sourceforge.net
>https://lists.sourceforge.net/lists/listinfo/csound-devel

Victor Lazzarini
Music Technology Laboratory
Music Department
National University of Ireland, Maynooth 



-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net