Group 1 is not created when buffer size < 256.

classic Classic list List threaded Threaded
25 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Group 1 is not created when buffer size < 256.

Martin Marier-2
Hi users and devs,

I just moved to SC 3.9 (thanks to all devs for the great work!) and
bumped into an issue with Groups.  Here is a way to reproduce the
problem (I'm on Arch Linux).

First, start Jack with a buffer size of 64 samples, then fire up SC.

s.boot;
s.queryAllNodes;
// NODE TREE Group 0

g = Group.head(s);
// -> Group(1001)
// FAILURE IN SERVER /g_new Group 1 not found

When Jack is running with a buffer size of 256, I do not have this
issue.

s.boot;
s.queryAllNodes;
// NODE TREE Group 0
//    1 group

g = Group.head(s);
// -> Group(1001)

FAILURE IN SERVER /g_new Group 1 not found
FAILURE IN SERVER /s_new Group 1000 not found

So it looks like Group 1 is not created whenever I try to use a small
buffer size.  I never had this issue with 3.8.

Any hints?

Thank you all very much!

Martin
--
Martin Marier
http://www.martinmarier.com
http://www.prisedesoncreative.com
Vous pouvez utiliser ma clef publique pour m'envoyer des courriels chiffrés.
You can use my public key to send me encrypted emails.
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xFB5BDFBE93D31280

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

5tefan


Am 2018-02-04 um 21:51 schrieb [hidden email]:
> So it looks like Group 1 is not created whenever I try to use a small
> buffer size.  I never had this issue with 3.8.
>
> Any hints?

I'm running SC 3.9 on Ubuntustudio. I start Jack with a buffer size of
128 and I don't have any problem with that when creating a new group
like you did. Don't know... sorry

Stefan

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

Martin Marier-2
Hi,

I just noticed I needed a fairly big numOutputBusChannels to reproduce
the issue.  Here is what I have in my startup.scd:

s.options.sampleRate_(48000);
s.options.numAudioBusChannels_(2.pow(11));
s.options.numOutputBusChannels_(30);
s.options.numInputBusChannels_(2);
s.options.numBuffers_(2048);
s.options.memSize_(2.pow(21));
s.options.numWireBufs_(2.pow(16));

Could you try again with these ServerOptions?

Thanks!

Martin

2018-02-04T21:06:46+0000  [hidden email]:

> Am 2018-02-04 um 21:51 schrieb [hidden email]:
>> So it looks like Group 1 is not created whenever I try to use a small
>> buffer size.  I never had this issue with 3.8.
>>
>> Any hints?
>
> I'm running SC 3.9 on Ubuntustudio. I start Jack with a buffer size of
> 128 and I don't have any problem with that when creating a new group
> like you did. Don't know... sorry
>
> Stefan
>
> _______________________________________________
> sc-users mailing list
>
> info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
> archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
> search: http://www.listarc.bham.ac.uk/lists/sc-users/search/


--
Martin Marier
http://www.martinmarier.com
http://www.prisedesoncreative.com
Vous pouvez utiliser ma clef publique pour m'envoyer des courriels chiffrés.
You can use my public key to send me encrypted emails.
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xFB5BDFBE93D31280

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

5tefan


Am 2018-02-04 um 22:15 schrieb [hidden email]:
> s.options.sampleRate_(48000);
> s.options.numAudioBusChannels_(2.pow(11));
> s.options.numOutputBusChannels_(30);
> s.options.numInputBusChannels_(2);
> s.options.numBuffers_(2048);
> s.options.memSize_(2.pow(21));
> s.options.numWireBufs_(2.pow(16));

hmm...

... tried your settings (except that I'm running at a sample rate of
88200) but couldn't reproduce what you described.

Running

g = Group.head(s);

returns

-> Group(1001)

... I'm clueless.

Stefan

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

5tefan
In reply to this post by Martin Marier-2
This is my configuration in .jackrc:

/usr/bin/jackd -P75 -dfirewire -r88200 -p128 -n3 -i10 -o10

Maybe the problem (or solution) lies here somewhere?

Stefan

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

Martin Marier-2
Hi Stefan,

Thanks for you help.  Right now, I am using jack_control (jack_dbu,
here).  Here is (the important part of) my script:

jack_control eps driver alsa eps realtime True eps realtime-priority 60
jack_control ds alsa dps device hw:PCH dps capture hw:PCH,0 dps playback hw:PCH,0
jack_control dps period 64 dps nperiods 2 dps rate 48000

I noticed that if I use 3 periods and 128 samples, everything works
smoothly.  But at 2 and 64, Group 1 is not created.

Martin


2018-02-04T21:29:38+0000  [hidden email]:

> This is my configuration in .jackrc:
>
> /usr/bin/jackd -P75 -dfirewire -r88200 -p128 -n3 -i10 -o10
>
> Maybe the problem (or solution) lies here somewhere?
>
> Stefan
>
> _______________________________________________
> sc-users mailing list
>
> info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
> archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
> search: http://www.listarc.bham.ac.uk/lists/sc-users/search/


--
Martin Marier
http://www.martinmarier.com
http://www.prisedesoncreative.com
Vous pouvez utiliser ma clef publique pour m'envoyer des courriels chiffrés.
You can use my public key to send me encrypted emails.
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xFB5BDFBE93D31280

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

Martin Marier-2

No issues with nperiods 3 and period 64, but the problem is present with
nperiods 2 and period 128.


2018-02-04T21:44:51+0000  Martin Marier <[hidden email]>:

> Hi Stefan,
>
> Thanks for you help.  Right now, I am using jack_control (jack_dbu,
> here).  Here is (the important part of) my script:
>
> jack_control eps driver alsa eps realtime True eps realtime-priority 60
> jack_control ds alsa dps device hw:PCH dps capture hw:PCH,0 dps playback hw:PCH,0
> jack_control dps period 64 dps nperiods 2 dps rate 48000
>
> I noticed that if I use 3 periods and 128 samples, everything works
> smoothly.  But at 2 and 64, Group 1 is not created.
>
> Martin
>
>
> 2018-02-04T21:29:38+0000  [hidden email]:
>> This is my configuration in .jackrc:
>>
>> /usr/bin/jackd -P75 -dfirewire -r88200 -p128 -n3 -i10 -o10
>>
>> Maybe the problem (or solution) lies here somewhere?
>>
>> Stefan
>>
>> _______________________________________________
>> sc-users mailing list
>>
>> info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
>> archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
>> search: http://www.listarc.bham.ac.uk/lists/sc-users/search/


--
Martin Marier
http://www.martinmarier.com
http://www.prisedesoncreative.com
Vous pouvez utiliser ma clef publique pour m'envoyer des courriels chiffrés.
You can use my public key to send me encrypted emails.
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xFB5BDFBE93D31280

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

5tefan
In reply to this post by Martin Marier-2


Am 2018-02-04 um 22:44 schrieb [hidden email]:
> I noticed that if I use 3 periods and 128 samples, everything works
> smoothly.  But at 2 and 64, Group 1 is not created.

I'm not using jack_control - .jackrc is created automatically by
QJackCtl. Anyway, 3 periods and 128 samples is exactly what I'm using.
I'm gonna give your settings a try (2 periods, 64 samples).

Stefan

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

5tefan
In reply to this post by Martin Marier-2


Am 2018-02-04 um 22:51 schrieb [hidden email]:
> No issues with nperiods 3 and period 64, but the problem is present with
> nperiods 2 and period 128.

just tried - no problems for me :\

... I can't tell for the rest of your setup - it differs quite a bit
from mine. I'm afraid I can't give any hint where to look for the source
of the problem.

Stefan

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

patrick
In reply to this post by Martin Marier-2
Hi Martin,

I just tried SC with jack (qjackctl) at 44100, period/buffer 2, frames/period 64. I get Group (1) after booting the server.

My initial feeling about this is that it is a emacs (scel?) issue. Do you get the same issue when using scide?

Cheers,
Patrick


From: [hidden email] <[hidden email]> on behalf of [hidden email] <[hidden email]>
Sent: February 4, 2018 4:57 PM
To: [hidden email]
Subject: Re: [sc-users] Group 1 is not created when buffer size < 256.
 


Am 2018-02-04 um 22:51 schrieb [hidden email]:
> No issues with nperiods 3 and period 64, but the problem is present with
> nperiods 2 and period 128.

just tried - no problems for me :\

... I can't tell for the rest of your setup - it differs quite a bit
from mine. I'm afraid I can't give any hint where to look for the source
of the problem.

Stefan

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

patrick
I just read the issue report on Github. So this is related to the number of output channels the server is configured to have? The issue is reproducible when configured to 30 output channels, but not with the default 2 channels?

Patrick

From: [hidden email] <[hidden email]> on behalf of [hidden email] <[hidden email]>
Sent: February 4, 2018 5:48 PM
To: [hidden email]
Subject: Re: [sc-users] Group 1 is not created when buffer size < 256.
 
Hi Martin,

I just tried SC with jack (qjackctl) at 44100, period/buffer 2, frames/period 64. I get Group (1) after booting the server.

My initial feeling about this is that it is a emacs (scel?) issue. Do you get the same issue when using scide?

Cheers,
Patrick


From: [hidden email] <[hidden email]> on behalf of [hidden email] <[hidden email]>
Sent: February 4, 2018 4:57 PM
To: [hidden email]
Subject: Re: [sc-users] Group 1 is not created when buffer size < 256.
 


Am 2018-02-04 um 22:51 schrieb [hidden email]:
> No issues with nperiods 3 and period 64, but the problem is present with
> nperiods 2 and period 128.

just tried - no problems for me :\

... I can't tell for the rest of your setup - it differs quite a bit
from mine. I'm afraid I can't give any hint where to look for the source
of the problem.

Stefan

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

patrick
Sorry Martin... I'm not seeing this with 30 channel SC and 2/64 in JACK.

Please test scide vs scel. Also, make sure you test system is as clean as possible by ignoring and extensions folders and undoing any hacks/customizations you may have. You may also want to try and build the 3.9 branch on Github. That will effectively give you not-yet-released 3.9.1 (as of yesterday).

Patrick

From: [hidden email] <[hidden email]> on behalf of [hidden email] <[hidden email]>
Sent: February 4, 2018 5:54 PM
To: [hidden email]
Subject: Re: [sc-users] Group 1 is not created when buffer size < 256.
 
I just read the issue report on Github. So this is related to the number of output channels the server is configured to have? The issue is reproducible when configured to 30 output channels, but not with the default 2 channels?

Patrick

From: [hidden email] <[hidden email]> on behalf of [hidden email] <[hidden email]>
Sent: February 4, 2018 5:48 PM
To: [hidden email]
Subject: Re: [sc-users] Group 1 is not created when buffer size < 256.
 
Hi Martin,

I just tried SC with jack (qjackctl) at 44100, period/buffer 2, frames/period 64. I get Group (1) after booting the server.

My initial feeling about this is that it is a emacs (scel?) issue. Do you get the same issue when using scide?

Cheers,
Patrick


From: [hidden email] <[hidden email]> on behalf of [hidden email] <[hidden email]>
Sent: February 4, 2018 4:57 PM
To: [hidden email]
Subject: Re: [sc-users] Group 1 is not created when buffer size < 256.
 


Am 2018-02-04 um 22:51 schrieb [hidden email]:
> No issues with nperiods 3 and period 64, but the problem is present with
> nperiods 2 and period 128.

just tried - no problems for me :\

... I can't tell for the rest of your setup - it differs quite a bit
from mine. I'm afraid I can't give any hint where to look for the source
of the problem.

Stefan

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

ddw_music
patrick wrote
> Sorry Martin... I'm not seeing this with 30 channel SC and 2/64 in JACK.

If Jack configuration settings change the behavior without any change in
SC's codebase, then the issue is related to Jack's realtime performance...
which means it's related to specific hardware (CPU, clock speed, CPU
performance settings, soundcard), driver configuration, and system
optimizations for realtime audio (interrupt priorities etc.). The odds of it
being reproducible on other systems are practically zero.

At the same time, there's a difference between 3.8 and 3.9, so it isn't all
Jack's fault.

The first thing I want to know is: Does SC send the message to create the
group but that fails silently, or does it not send the message at all?

Martin, could you copy/paste the complete post window output after "s.boot"?
I.e. "booting 57110..." and everything after that?


> Please test scide vs scel.

I don't understand the rationale here.


> Also, make sure you test system is as clean as possible by ignoring and
> extensions folders and undoing any hacks/customizations you may have.

That's reasonable... plain vanilla as possible.

hjh



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

patrick
>Please test scide vs scel.
>
>I don't understand the rationale here.

To eliminate a potential variable. I know Martin uses emacs, so I'm simply asking him to confirm that he gets the same behavior with scide. Nothing wrong with that.

Patrick

From: [hidden email] <[hidden email]> on behalf of [hidden email] <[hidden email]>
Sent: February 4, 2018 6:22 PM
To: [hidden email]
Subject: [sc-users] Re: Group 1 is not created when buffer size < 256.
 
patrick wrote
> Sorry Martin... I'm not seeing this with 30 channel SC and 2/64 in JACK.

If Jack configuration settings change the behavior without any change in
SC's codebase, then the issue is related to Jack's realtime performance...
which means it's related to specific hardware (CPU, clock speed, CPU
performance settings, soundcard), driver configuration, and system
optimizations for realtime audio (interrupt priorities etc.). The odds of it
being reproducible on other systems are practically zero.

At the same time, there's a difference between 3.8 and 3.9, so it isn't all
Jack's fault.

The first thing I want to know is: Does SC send the message to create the
group but that fails silently, or does it not send the message at all?

Martin, could you copy/paste the complete post window output after "s.boot"?
I.e. "booting 57110..." and everything after that?


> Please test scide vs scel.

I don't understand the rationale here.


> Also, make sure you test system is as clean as possible by ignoring and
> extensions folders and undoing any hacks/customizations you may have.

That's reasonable... plain vanilla as possible.

hjh



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

ddw_music
patrick wrote
>>Please test scide vs scel.
>>
>>I don't understand the rationale here.
>
> To eliminate a potential variable. I know Martin uses emacs,

Ah ok, the rationale is to remove emacs as a potential cause. Ok, I missed
that. I don't believe it will make a difference.

In that case, testing sclang from the command line (no editor) would be even
better.

hjh



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

Martin Marier-2

Thank you both for your support!

I did try to use scide and sclang from the command line, but I get the
same issue.  There is, however, something fishy about the output I get
when I boot the server.

The full output is pasted below, but the line "/notify: already
registered" is the most interesting one.

I'll have more time to troubleshoot this evening (UT -5:00).


-> localhost
booting server 'localhost' on address: 127.0.0.1:57110
Found 332 LADSPA plugins
JackDriver: client name is 'SuperCollider'
SC_AudioDriver: sample rate = 48000.000000, driver's block size = 64
JackDriver: connected  system:capture_1 to SuperCollider:in_1
JackDriver: connected  system:capture_2 to SuperCollider:in_2
JackDriver: connected  SuperCollider:out_1 to system:playback_1
JackDriver: connected  SuperCollider:out_2 to system:playback_2
JackDriver: connected  SuperCollider:out_3 to system:playback_3
JackDriver: connected  SuperCollider:out_4 to system:playback_4
SuperCollider 3 server ready.
JackDriver: max output latency 1.3 ms
Requested notification messages from server 'localhost'
localhost: server process's maxLogins (1) matches with my options.
localhost: keeping clientID (0) as confirmed by server process.
Requested notification messages from server 'localhost'
/notify : already registered
localhost - already registered with clientID 0.
Shared memory server interface initialized


2018-02-04T23:43:40+0000  [hidden email]:

> patrick wrote
>>>Please test scide vs scel.
>>>
>>>I don't understand the rationale here.
>>
>> To eliminate a potential variable. I know Martin uses emacs,
>
> Ah ok, the rationale is to remove emacs as a potential cause. Ok, I missed
> that. I don't believe it will make a difference.
>
> In that case, testing sclang from the command line (no editor) would be even
> better.
>
> hjh


--
Martin Marier
http://www.martinmarier.com
http://www.prisedesoncreative.com
Vous pouvez utiliser ma clef publique pour m'envoyer des courriels chiffrés.
You can use my public key to send me encrypted emails.
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xFB5BDFBE93D31280

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

Martin Marier-2
Hi all,

I had more time to troubleshoot this issue and found that it only
happens when either Ctk or SenseWorld are installed.

Can anyone reproduce the issue with this extra bit of information?

Thank you all!

Martin

2018-02-05T18:50:52+0000  [hidden email]:

> Thank you both for your support!
>
> I did try to use scide and sclang from the command line, but I get the
> same issue.  There is, however, something fishy about the output I get
> when I boot the server.
>
> The full output is pasted below, but the line "/notify: already
> registered" is the most interesting one.
>
> I'll have more time to troubleshoot this evening (UT -5:00).
>
>
> -> localhost
> booting server 'localhost' on address: 127.0.0.1:57110
> Found 332 LADSPA plugins
> JackDriver: client name is 'SuperCollider'
> SC_AudioDriver: sample rate = 48000.000000, driver's block size = 64
> JackDriver: connected  system:capture_1 to SuperCollider:in_1
> JackDriver: connected  system:capture_2 to SuperCollider:in_2
> JackDriver: connected  SuperCollider:out_1 to system:playback_1
> JackDriver: connected  SuperCollider:out_2 to system:playback_2
> JackDriver: connected  SuperCollider:out_3 to system:playback_3
> JackDriver: connected  SuperCollider:out_4 to system:playback_4
> SuperCollider 3 server ready.
> JackDriver: max output latency 1.3 ms
> Requested notification messages from server 'localhost'
> localhost: server process's maxLogins (1) matches with my options.
> localhost: keeping clientID (0) as confirmed by server process.
> Requested notification messages from server 'localhost'
> /notify : already registered
> localhost - already registered with clientID 0.
> Shared memory server interface initialized
>
>
> 2018-02-04T23:43:40+0000  [hidden email]:
>> patrick wrote
>>>>Please test scide vs scel.
>>>>
>>>>I don't understand the rationale here.
>>>
>>> To eliminate a potential variable. I know Martin uses emacs,
>>
>> Ah ok, the rationale is to remove emacs as a potential cause. Ok, I missed
>> that. I don't believe it will make a difference.
>>
>> In that case, testing sclang from the command line (no editor) would be even
>> better.
>>
>> hjh


--
Martin Marier
http://www.martinmarier.com
http://www.prisedesoncreative.com
Vous pouvez utiliser ma clef publique pour m'envoyer des courriels chiffrés.
You can use my public key to send me encrypted emails.
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xFB5BDFBE93D31280

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

ddw_music
Martin Marier-2 wrote
> I had more time to troubleshoot this issue and found that it only
> happens when either Ctk or SenseWorld are installed.
>
> Can anyone reproduce the issue with this extra bit of information?

Again, I don't think it will be reproducible without matching CPU speed,
power settings, interrupt priorities, etc.

I'd like to see what messages are going out during server boot. I'm not sure
I trust dumpOSC for this, so let's try this. First, put the attached file
into your Extensions folder, and recompile the classlib. Then:

s.addr = DebugNetAddr("127.0.0.1", 57110);
s.boot;

I get:

-> localhost
latency nil SysClock logical time 355.041886065 thisThread's logical time
355.041886065
        [ "/sync", -925223946 ]

-> localhost
booting server 'localhost' on address: 127.0.0.1:57110
Found 380 LADSPA plugins
JackDriver: client name is 'SuperCollider'
SC_AudioDriver: sample rate = 44100.000000, driver's block size = 1024
JackDriver: connected  system:capture_1 to SuperCollider:in_1
JackDriver: connected  system:capture_2 to SuperCollider:in_2
JackDriver: connected  SuperCollider:out_1 to system:playback_1
JackDriver: connected  SuperCollider:out_2 to system:playback_2
SuperCollider 3 server ready.
JackDriver: max output latency 46.4 ms
Requested notification messages from server 'localhost'
latency nil SysClock logical time 356.713374449 thisThread's logical time
356.713374449
        [ "/notify", 1, 0 ]

localhost: server process's maxLogins (1) matches with my options.
localhost: keeping clientID (0) as confirmed by server process.
latency nil SysClock logical time 356.742752383 thisThread's logical time
356.742752383
        [ "/d_recv", "data[ 1650 ]", nil ]

/* lots of /d_recv and /sync, and then:

latency nil SysClock logical time 357.011112984 thisThread's logical time
357.011112984
        [ "/g_new", 1, 0, 0 ]

latency nil SysClock logical time 357.011112984 thisThread's logical time
357.011112984
        [ "/sync", 1003 ]

latency nil SysClock logical time 357.038871856 thisThread's logical time
357.038871856
        [ "/sync", 1004 ]

latency nil SysClock logical time 357.173197203 thisThread's logical time
357.173197203
        [ "/sync", 1005 ]

Shared memory server interface initialized

You can ignore the /d_recv messages, but "[ "/g_new", 1, 0, 0 ]" is
important.

This will tell us if sclang is sending the message (I think it is). Also do
another test:

ServerBoot.add({ |server| server.dumpOSC(1) });
s.boot;

This will tell us what scsynth is receiving. I get fewer /d_recv messages
(because dumpOSC comes a bit late), but I do see "[ "/g_new", 1, 0, 0 ]" in
there.

hjh

DebugNetAddr.sc
<http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/file/t149410/DebugNetAddr.sc>  



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

Martin Marier-2
Hi James,

Here is the output I get using your DebugNetAddr.sc file.

--------------------------
-> localhost
server 'localhost' unresponsive, rebooting ...
'/quit' sent

latency nil SysClock logical time 9.813606074 thisThread's logical time 9.813606074
        [ 24, 0 ]

latency nil SysClock logical time 9.813606074 thisThread's logical time 9.813606074
        [ "/sync", 586759981 ]

-> localhost
booting server 'localhost' on address: 127.0.0.1:57110
Found 330 LADSPA plugins
JackDriver: client name is 'SuperCollider'
SC_AudioDriver: sample rate = 48000.000000, driver's block size = 128
JackDriver: connected  system:capture_1 to SuperCollider:in_1
JackDriver: connected  system:capture_2 to SuperCollider:in_2
JackDriver: connected  SuperCollider:out_1 to system:playback_1
JackDriver: connected  SuperCollider:out_2 to system:playback_2
JackDriver: connected  SuperCollider:out_3 to system:playback_3
JackDriver: connected  SuperCollider:out_4 to system:playback_4
SuperCollider 3 server ready.
JackDriver: max output latency 2.7 ms
Requested notification messages from server 'localhost'
latency nil SysClock logical time 11.471088962 thisThread's logical time 11.471088962
        [ "/notify", 1, 0 ]

localhost: server process's maxLogins (1) matches with my options.
localhost: keeping clientID (0) as confirmed by server process.
latency nil SysClock logical time 11.474373748 thisThread's logical time 11.474373748
        [ "/d_recv", "data[ 3037 ]", nil ]

... many more /d_recv ...

latency nil SysClock logical time 11.474373748 thisThread's logical time 11.474373748
        [ "/sync", 1001 ]

latency nil SysClock logical time 11.474373748 thisThread's logical time 10.754259857
        [ "/d_recv", "data[ 1791 ]", nil ]

latency nil SysClock logical time 11.474373748 thisThread's logical time 10.754259857
        [ "/sync", 1002 ]

Requested notification messages from server 'localhost'
latency nil SysClock logical time 12.172731783 thisThread's logical time 12.172731783
        [ "/notify", 1, 0 ]

/notify : already registered
localhost - already registered with clientID 0.
latency nil SysClock logical time 12.254455678 thisThread's logical time 12.254455678
        [ "/sync", 1003 ]

Shared memory server interface initialized
--------------------------

As you can (not) see, there are no /g_new messages in there.  I do not
see the /g_new message either when using dumpOSC.

If I remove the Ctk quark, everything is fine and i do see the
[ "/g_new", 1, 0, 0 ] message.

Thank you very much for your help.

Cheers,
Martin


2018-02-07T16:46:22+0000  [hidden email]:

> Martin Marier-2 wrote
>> I had more time to troubleshoot this issue and found that it only
>> happens when either Ctk or SenseWorld are installed.
>>
>> Can anyone reproduce the issue with this extra bit of information?
>
> Again, I don't think it will be reproducible without matching CPU speed,
> power settings, interrupt priorities, etc.
>
> I'd like to see what messages are going out during server boot. I'm not sure
> I trust dumpOSC for this, so let's try this. First, put the attached file
> into your Extensions folder, and recompile the classlib. Then:
>
> s.addr = DebugNetAddr("127.0.0.1", 57110);
> s.boot;
>
> You can ignore the /d_recv messages, but "[ "/g_new", 1, 0, 0 ]" is
> important.
>
> This will tell us if sclang is sending the message (I think it is). Also do
> another test:
>
> ServerBoot.add({ |server| server.dumpOSC(1) });
> s.boot;
>
> This will tell us what scsynth is receiving. I get fewer /d_recv messages
> (because dumpOSC comes a bit late), but I do see "[ "/g_new", 1, 0, 0 ]" in
> there.
>
> hjh
>
> DebugNetAddr.sc
> <http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/file/t149410/DebugNetAddr.sc>

--
Martin Marier
http://www.martinmarier.com
http://www.prisedesoncreative.com
Vous pouvez utiliser ma clef publique pour m'envoyer des courriels chiffrés.
You can use my public key to send me encrypted emails.
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xFB5BDFBE93D31280

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
Reply | Threaded
Open this post in threaded view
|

Re: Group 1 is not created when buffer size < 256.

ddw_music
Martin Marier-2 wrote
> /notify : already registered
> localhost - already registered with clientID 0.

I've searched both Ctk and SenseWorld for "notify" and I can't find them. I
can think of two explanations: either you have a different version of the
quarks, or something else in your system is doing this.


> latency nil SysClock logical time 12.254455678 thisThread's logical time
> 12.254455678
> [ "/sync", 1003 ]

Hm. Possibly this sync is not receiving a reply, blocking the rest of the
init process. Just a guess.

I'm totally baffled, to be honest. Something happens when Ctk is installed,
which doesn't happen when Ctk is not installed, and there's no evidence of
extra "notify" calls in Ctk itself. I don't know how to proceed.

Except: There was also the suggestion to test with a minimal configuration
-- no extensions (except Ctk, because you need it to reproduce the issue). I
see that you tried from commandline but didn't confirm explicitly that Ctk
alone, on top of a plain vanilla environment, was enough to cause the
problem. (Maybe you did test that, but it isn't clear from the thread
history.)

I think it may be Ctk + something else that we haven't found yet.

hjh



--
Sent from: http://new-supercollider-mailing-lists-forums-use-these.2681727.n2.nabble.com/SuperCollider-Users-New-Use-this-f2676391.html

_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
12