[Scilab-users] Scilab stops calculus

classic Classic list List threaded Threaded
14 messages Options
jbaudais jbaudais
Reply | Threaded
Open this post in threaded view
|

[Scilab-users] Scilab stops calculus

Hello,

  Two times Scilab was stuck after 20 hours of simulation (I used to do simulations that need more than one week with Scilab 5). It was big simulation with many levels of iterations (to evaluate probabilities), with intermediate results written in files (different files with save command and timestamp), and with the iteration numbers written in the Scilab window (so many output lines). All worked fine still Scilab locks with:
- no CPU activity (Scilab really stops all calculus, but not exits),
- low memory use as expected (no mem leakage),
- the Scilab window blocked, showing the intermediate iteration numbers (Scilab is blocked within an iteration, there is no input-output issues in this part of the code), but nothing can be done (CTRL-C, CTRL-D or any control command).
I must kill the Scilab process within a Shell terminal. I also executed the "faulty" iteration alone (with 1e4 sub-iterations in more or less 2 hours), but there is no problem. I obtained the result. So I can get around the problem, but not solved it.

Is there some known limitations? limited prompt lines for example...

Thanks

--
Jean-Yves Baudais
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
Antoine Monmayrant Antoine Monmayrant
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

Hello Jean-Yves,

Are you running scilab under Linux?
It might not be related, but I noticed that when my computer resumes
from sleep, all the scilab GUI is unresponsive.
I don't have this issue when running scilab-cli.
Can you run your long simulation without gui?

Antoine

Le 10/03/2021 à 22:30, Jean-Yves Baudais a écrit :

> Hello,
>
>    Two times Scilab was stuck after 20 hours of simulation (I used to do simulations that need more than one week with Scilab 5). It was big simulation with many levels of iterations (to evaluate probabilities), with intermediate results written in files (different files with save command and timestamp), and with the iteration numbers written in the Scilab window (so many output lines). All worked fine still Scilab locks with:
> - no CPU activity (Scilab really stops all calculus, but not exits),
> - low memory use as expected (no mem leakage),
> - the Scilab window blocked, showing the intermediate iteration numbers (Scilab is blocked within an iteration, there is no input-output issues in this part of the code), but nothing can be done (CTRL-C, CTRL-D or any control command).
> I must kill the Scilab process within a Shell terminal. I also executed the "faulty" iteration alone (with 1e4 sub-iterations in more or less 2 hours), but there is no problem. I obtained the result. So I can get around the problem, but not solved it.
>
> Is there some known limitations? limited prompt lines for example...
>
> Thanks
>
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
jbaudais jbaudais
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

Hi Antoine,

> Are you running scilab under Linux?

Yes.

> It might not be related, but I noticed that when my computer resumes
> from sleep, all the scilab GUI is unresponsive.

I removed all sleep modes on my computer (to maintain VPN access and get around some other problems) and I never use them.

> Can you run your long simulation without gui?

It was. I tested the following

i=0;while 1, i=i+1;mprintf("%d\n",i); end;

and there is no problem up to i=2e8 with -nwni option, and up to i=5e7 with gui (I locked my screen during the simulation and unlocked it to stop the simulation, no problem to resume). So, my first feeling was wrong, it's not a prompt limitation.

Strange behaviour...

-- Jean-Yves
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
Antoine Monmayrant Antoine Monmayrant
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus


On 11/03/2021 11:08, Jean-Yves Baudais wrote:

> Hi Antoine,
>
>> Are you running scilab under Linux?
> Yes.
>
>> It might not be related, but I noticed that when my computer resumes
>> from sleep, all the scilab GUI is unresponsive.
> I removed all sleep modes on my computer (to maintain VPN access and get around some other problems) and I never use them.
>
>> Can you run your long simulation without gui?
> It was. I tested the following
>
> i=0;while 1, i=i+1;mprintf("%d\n",i); end;
>
> and there is no problem up to i=2e8 with -nwni option, and up to i=5e7 with gui (I locked my screen during the simulation and unlocked it to stop the simulation, no problem to resume). So, my first feeling was wrong, it's not a prompt limitation.
It might be java-related as -nwni is shutting off all the java-based
elements of Scilab.

I don't know whether this bug has been reported yet...

Antoine

>
> Strange behaviour...
>
> -- Jean-Yves
> _______________________________________________
> users mailing list
> [hidden email]
> http://lists.scilab.org/mailman/listinfo/users
>
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
mottelet mottelet
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

Hi,

Le 11/03/2021 à 15:56, Antoine Monmayrant a écrit :

>
> On 11/03/2021 11:08, Jean-Yves Baudais wrote:
>> Hi Antoine,
>>
>>> Are you running scilab under Linux?
>> Yes.
>>
>>> It might not be related, but I noticed that when my computer resumes
>>> from sleep, all the scilab GUI is unresponsive.
>> I removed all sleep modes on my computer (to maintain VPN access and
>> get around some other problems) and I never use them.
>>
>>> Can you run your long simulation without gui?
>> It was. I tested the following
>>
>> i=0;while 1, i=i+1;mprintf("%d\n",i); end;
>>
>> and there is no problem up to i=2e8 with -nwni option, and up to
>> i=5e7 with gui (I locked my screen during the simulation and unlocked
>> it to stop the simulation, no problem to resume). So, my first
>> feeling was wrong, it's not a prompt limitation.
> It might be java-related as -nwni is shutting off all the java-based
> elements of Scilab.
>
> I don't know whether this bug has been reported yet...

Did you try without the mprintf ?

S;


>
> Antoine
>
>>
>> Strange behaviour...
>>
>> -- Jean-Yves
>> _______________________________________________
>> users mailing list
>> [hidden email]
>> https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users 
>>
>>
> _______________________________________________
> users mailing list
> [hidden email]
> https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users 
>

--
Stéphane Mottelet
Ingénieur de recherche
EA 4297 Transformations Intégrées de la Matière Renouvelable
Département Génie des Procédés Industriels
Sorbonne Universités - Université de Technologie de Compiègne
CS 60319, 60203 Compiègne cedex
Tel : +33(0)344234688
http://www.utc.fr/~mottelet

_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
mottelet mottelet
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

In reply to this post by jbaudais
Hello Jean-Yves,

I am affraid that the problem could be due to the output, because I had
no problem to run the loop for almost one day under Linux (with latests 
branch-6.1 build):

--> unix("date");
jeudi 11 mars 2021, 17:11:48 (UTC+0100)

--> i=0;while 1, i=i+1;end;

-1-> unix("date");
vendredi 12 mars 2021, 08:13:08 (UTC+0100)

-1-> i
  i  =

    1.667D+11

-1->

S.


Le 11/03/2021 à 11:08, Jean-Yves Baudais a écrit :

> Hi Antoine,
>
>> Are you running scilab under Linux?
> Yes.
>
>> It might not be related, but I noticed that when my computer resumes
>> from sleep, all the scilab GUI is unresponsive.
> I removed all sleep modes on my computer (to maintain VPN access and get around some other problems) and I never use them.
>
>> Can you run your long simulation without gui?
> It was. I tested the following
>
> i=0;while 1, i=i+1;mprintf("%d\n",i); end;
>
> and there is no problem up to i=2e8 with -nwni option, and up to i=5e7 with gui (I locked my screen during the simulation and unlocked it to stop the simulation, no problem to resume). So, my first feeling was wrong, it's not a prompt limitation.
>
> Strange behaviour...
>
> -- Jean-Yves
> _______________________________________________
> users mailing list
> [hidden email]
> https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users

--
Stéphane Mottelet
Ingénieur de recherche
EA 4297 Transformations Intégrées de la Matière Renouvelable
Département Génie des Procédés Industriels
Sorbonne Universités - Université de Technologie de Compiègne
CS 60319, 60203 Compiègne cedex
Tel : +33(0)344234688
http://www.utc.fr/~mottelet

_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
jbaudais jbaudais
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

Hello Stephan,

To be honest, I don't know where my problem comes from! If you said it cannot be an output problem, I follow you. So, I don't know why I couldn't wake it up and why it has fallen asleep. I need first to reproduce the problem with simple tests... (to rule out wrong code issue)

Thanks,

Jean-Yves

----- Original Message -----
> From: "Stéphane Mottelet"
> To: "Users mailing list for Scilab" scilab.org>
> Sent: Friday, 12 March, 2021 08:16:26
> Subject: Re: [Scilab-users] Scilab stops calculus

> Hello Jean-Yves,
>
> I am affraid that the problem could be due to the output, because I had
> no problem to run the loop for almost one day under Linux (with latests
> branch-6.1 build):
>
> --> unix("date");
> jeudi 11 mars 2021, 17:11:48 (UTC+0100)
>
> --> i=0;while 1, i=i+1;end;
>
> -1-> unix("date");
> vendredi 12 mars 2021, 08:13:08 (UTC+0100)
>
> -1-> i
>  i  =
>
>    1.667D+11
>
> -1->
>
> S.
>
>
> Le 11/03/2021 à 11:08, Jean-Yves Baudais a écrit :
>> Hi Antoine,
>>
>>> Are you running scilab under Linux?
>> Yes.
>>
>>> It might not be related, but I noticed that when my computer resumes
>>> from sleep, all the scilab GUI is unresponsive.
>> I removed all sleep modes on my computer (to maintain VPN access and get around
>> some other problems) and I never use them.
>>
>>> Can you run your long simulation without gui?
>> It was. I tested the following
>>
>> i=0;while 1, i=i+1;mprintf("%d\n",i); end;
>>
>> and there is no problem up to i=2e8 with -nwni option, and up to i=5e7 with gui
>> (I locked my screen during the simulation and unlocked it to stop the
>> simulation, no problem to resume). So, my first feeling was wrong, it's not a
>> prompt limitation.
>>
>> Strange behaviour...
>>
>> -- Jean-Yves
>> _______________________________________________
>> users mailing list
>> [hidden email]
>> https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users
>
> --
> Stéphane Mottelet
> Ingénieur de recherche
> EA 4297 Transformations Intégrées de la Matière Renouvelable
> Département Génie des Procédés Industriels
> Sorbonne Universités - Université de Technologie de Compiègne
> CS 60319, 60203 Compiègne cedex
> Tel : +33(0)344234688
> http://www.utc.fr/~mottelet
>
> _______________________________________________
> users mailing list
> [hidden email]
> http://lists.scilab.org/mailman/listinfo/users

--
Jean-Yves Baudais
Pour un bon usage de la messagerie https://intranet.univ-rennes2.fr/system/files/UHB/DRH/ACTION-SOCIALE/plaqu_dubonusage.b-1.pdf
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
mottelet mottelet
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

Stay tuned, I have relaunched the loop but this time including the
mprintf and will keep you informed of the result (freeze or no freeze)

S.

Le 12/03/2021 à 11:09, Jean-Yves Baudais a écrit :

> Hello Stephan,
>
> To be honest, I don't know where my problem comes from! If you said it cannot be an output problem, I follow you. So, I don't know why I couldn't wake it up and why it has fallen asleep. I need first to reproduce the problem with simple tests... (to rule out wrong code issue)
>
> Thanks,
>
> Jean-Yves
>
> ----- Original Message -----
>> From: "Stéphane Mottelet"
>> To: "Users mailing list for Scilab" scilab.org>
>> Sent: Friday, 12 March, 2021 08:16:26
>> Subject: Re: [Scilab-users] Scilab stops calculus
>> Hello Jean-Yves,
>>
>> I am affraid that the problem could be due to the output, because I had
>> no problem to run the loop for almost one day under Linux (with latests
>> branch-6.1 build):
>>
>> --> unix("date");
>> jeudi 11 mars 2021, 17:11:48 (UTC+0100)
>>
>> --> i=0;while 1, i=i+1;end;
>>
>> -1-> unix("date");
>> vendredi 12 mars 2021, 08:13:08 (UTC+0100)
>>
>> -1-> i
>>   i  =
>>
>>     1.667D+11
>>
>> -1->
>>
>> S.
>>
>>
>> Le 11/03/2021 à 11:08, Jean-Yves Baudais a écrit :
>>> Hi Antoine,
>>>
>>>> Are you running scilab under Linux?
>>> Yes.
>>>
>>>> It might not be related, but I noticed that when my computer resumes
>>>> from sleep, all the scilab GUI is unresponsive.
>>> I removed all sleep modes on my computer (to maintain VPN access and get around
>>> some other problems) and I never use them.
>>>
>>>> Can you run your long simulation without gui?
>>> It was. I tested the following
>>>
>>> i=0;while 1, i=i+1;mprintf("%d\n",i); end;
>>>
>>> and there is no problem up to i=2e8 with -nwni option, and up to i=5e7 with gui
>>> (I locked my screen during the simulation and unlocked it to stop the
>>> simulation, no problem to resume). So, my first feeling was wrong, it's not a
>>> prompt limitation.
>>>
>>> Strange behaviour...
>>>
>>> -- Jean-Yves
>>> _______________________________________________
>>> users mailing list
>>> [hidden email]
>>> https://antispam.utc.fr/proxy/2/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users
>> --
>> Stéphane Mottelet
>> Ingénieur de recherche
>> EA 4297 Transformations Intégrées de la Matière Renouvelable
>> Département Génie des Procédés Industriels
>> Sorbonne Universités - Université de Technologie de Compiègne
>> CS 60319, 60203 Compiègne cedex
>> Tel : +33(0)344234688
>> https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/www.utc.fr/~mottelet
>>
>> _______________________________________________
>> users mailing list
>> [hidden email]
>> https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users

--
Stéphane Mottelet
Ingénieur de recherche
EA 4297 Transformations Intégrées de la Matière Renouvelable
Département Génie des Procédés Industriels
Sorbonne Universités - Université de Technologie de Compiègne
CS 60319, 60203 Compiègne cedex
Tel : +33(0)344234688
http://www.utc.fr/~mottelet

_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
Antoine Monmayrant Antoine Monmayrant
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

In reply to this post by jbaudais
Hello Jean-Yves,

I also ran your test under linux (Ubuntu 18.04, Xorg) and did have any
problem (I went beyond i=10^8).
What is your specific distro and window manager?

Antoine

Le 12/03/2021 à 11:09, Jean-Yves Baudais a écrit :

> Hello Stephan,
>
> To be honest, I don't know where my problem comes from! If you said it cannot be an output problem, I follow you. So, I don't know why I couldn't wake it up and why it has fallen asleep. I need first to reproduce the problem with simple tests... (to rule out wrong code issue)
>
> Thanks,
>
> Jean-Yves
>
> ----- Original Message -----
>> From: "Stéphane Mottelet"
>> To: "Users mailing list for Scilab" scilab.org>
>> Sent: Friday, 12 March, 2021 08:16:26
>> Subject: Re: [Scilab-users] Scilab stops calculus
>> Hello Jean-Yves,
>>
>> I am affraid that the problem could be due to the output, because I had
>> no problem to run the loop for almost one day under Linux (with latests
>> branch-6.1 build):
>>
>> --> unix("date");
>> jeudi 11 mars 2021, 17:11:48 (UTC+0100)
>>
>> --> i=0;while 1, i=i+1;end;
>>
>> -1-> unix("date");
>> vendredi 12 mars 2021, 08:13:08 (UTC+0100)
>>
>> -1-> i
>>   i  =
>>
>>     1.667D+11
>>
>> -1->
>>
>> S.
>>
>>
>> Le 11/03/2021 à 11:08, Jean-Yves Baudais a écrit :
>>> Hi Antoine,
>>>
>>>> Are you running scilab under Linux?
>>> Yes.
>>>
>>>> It might not be related, but I noticed that when my computer resumes
>>>> from sleep, all the scilab GUI is unresponsive.
>>> I removed all sleep modes on my computer (to maintain VPN access and get around
>>> some other problems) and I never use them.
>>>
>>>> Can you run your long simulation without gui?
>>> It was. I tested the following
>>>
>>> i=0;while 1, i=i+1;mprintf("%d\n",i); end;
>>>
>>> and there is no problem up to i=2e8 with -nwni option, and up to i=5e7 with gui
>>> (I locked my screen during the simulation and unlocked it to stop the
>>> simulation, no problem to resume). So, my first feeling was wrong, it's not a
>>> prompt limitation.
>>>
>>> Strange behaviour...
>>>
>>> -- Jean-Yves
>>> _______________________________________________
>>> users mailing list
>>> [hidden email]
>>> https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users
>> --
>> Stéphane Mottelet
>> Ingénieur de recherche
>> EA 4297 Transformations Intégrées de la Matière Renouvelable
>> Département Génie des Procédés Industriels
>> Sorbonne Universités - Université de Technologie de Compiègne
>> CS 60319, 60203 Compiègne cedex
>> Tel : +33(0)344234688
>> http://www.utc.fr/~mottelet
>>
>> _______________________________________________
>> users mailing list
>> [hidden email]
>> http://lists.scilab.org/mailman/listinfo/users
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
jbaudais jbaudais
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

> What is your specific distro and window manager?

Unbuntu 18.04 and 16.04 (two computers), Gnome on both.

-- Jean-Yves
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
jbaudais jbaudais
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

> OK, and you see this issue on both?

The fisrt time on 18.04 (laptop) and the second on 16.04 (desktop PC).

> Are you using a wayland session or an Xorg one?

XDG_SESSION_TYPE=x11 on both

-- Jean-Yves
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
Antoine Monmayrant Antoine Monmayrant
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus


On 12/03/2021 15:06, Jean-Yves Baudais wrote:
>> OK, and you see this issue on both?
> The fisrt time on 18.04 (laptop) and the second on 16.04 (desktop PC).
>
>> Are you using a wayland session or an Xorg one?
> XDG_SESSION_TYPE=x11 on both

OK, I'll have to check, I think I can get my hands on a 16.04 and 18.04
with X11...

Antoine

>
> -- Jean-Yves
>
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
Antoine Monmayrant Antoine Monmayrant
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

In reply to this post by jbaudais

Le 12/03/2021 à 15:06, Jean-Yves Baudais a écrit :
>> OK, and you see this issue on both?
> The fisrt time on 18.04 (laptop) and the second on 16.04 (desktop PC).
>
>> Are you using a wayland session or an Xorg one?
> XDG_SESSION_TYPE=x11 on both
OK, I let your test running for a long time on a 18.04 with X11: I
reached few 10⁹ without any issue.

Did you manage to reproduce this bug 100% of the time?

Antoine

>
> -- Jean-Yves
>
_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users
jbaudais jbaudais
Reply | Threaded
Open this post in threaded view
|

Re: Scilab stops calculus

Hi,

   Bad news for me, but good ones for Scilab. I checked the following code

i=0;
a=getdate();
while 1 then
   i=i+1;
   b=getdate()-a;
   mprintf("%d: %d/%d %d:%d:%d,%d\n",i,b([6,2,7:10]));
   sleep(1,"s");
end

on two computers, desktop Ubuntu 16.04 Gnome X11 and laptop Ubuntu 18.04
Gnome X11, with and without GUI during more than 38 hours, and in the
four cases Scilab didn't froze. So, no problem for Scilab. (But the
mystery remains, only for me.)

--Jean-Yves

_______________________________________________
users mailing list
[hidden email]
http://lists.scilab.org/mailman/listinfo/users