scicos problems (I cannot drag&drop the "from workspace" component)

[Mariusz Grecki]

this is unfortunately present either in the newest Windows version or in older Linux versions... The small dialog says: Failed to drop "FROMWSB" block!

[Jean-Philippe Chancelier]

Dear Mariusz,
Thanks for reporting, you are right. Edit the file macros/blocks/Sources/FROMWSB.sci and replace
scs_m=FROM_WSB_define()
by
scs_m=FROMWSB_define()

To recompile the sci file type the followin code while in scicos-4.4
nsp->pwd
/usr/local/src/nsp2-trunk-gtk3/toolboxes/scicos-4.4
nsp->add_lib('macros/blocks/Sources/',compile=%t)

Regards,
jpc

On 16/04/2019 00:34, 'Mariusz Grecki' via tumbi wrote:

this is unfortunately present either in the newest Windows version or in older Linux versions... The small dialog says: Failed to drop "FROMWSB" block!

--
You received this message because you are subscribed to the Google Groups "tumbi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tumbi+un...@googlegroups.com.
To post to this group, send email to tu...@googlegroups.com.
Visit this group at https://groups.google.com/group/tumbi.
For more options, visit https://groups.google.com/d/optout.

[Mariusz Grecki]

Dear Jean-Philippe,

thanks a lot for your support. Your advice helped, this problem is solved. However, I still have problems with scicos (in particular with "from workspace" block). In the old version, which I have installed on my laptop (scicoslab-gtk_4.4.1-1_i386.trusty.deb) the following example works as expected, but in NSP not.

I set up the p variable according to the help description, call the scicos and put in the diagram only 2 blocks: "from workspace" (sources) and "cscope" (sinks). At "from workspace" block I put the variable name p in "Variable name" and I connected both blocks. Then I tried to run the simulation. A small dialog box appeared saying "Not enough information to find port size. I try to find the problem" (but I think it is not the real problem since in scicoslab version it complains the same) and then another (bigger dialog) appeared (please look at the enclosed picture). The same behavior I have observed on my parallella with compiled sources from https://cermics.enpc.fr/~jpc/nsp-tiddly/files/nsp2_trunk_2019-01-23.tgz.

As I wrote already - the old scicos version delivered with scicoslab works OK - the scope recorded waveform as expected. 

The nsp version I used is nsp-gtk3_1.1-2_amd64.jessie.deb.

Greetings

Mariusz

W dniu środa, 17 kwietnia 2019 08:27:40 UTC+2 użytkownik Jean-Philippe Chancelier napisał:

Dear Mariusz,
Thanks for reporting, you are right. Edit the file macros/blocks/Sources/FROMWSB.sci and replace
scs_m=FROM_WSB_define()
by
scs_m=FROMWSB_define()

To recompile the sci file type the followin code while in scicos-4.4
nsp->pwd
/usr/local/src/nsp2-trunk-gtk3/toolboxes/scicos-4.4
nsp->add_lib('macros/blocks/Sources/',compile=%t)

Regards,
jpc

On 16/04/2019 00:34, 'Mariusz Grecki' via tumbi wrote:

this is unfortunately present either in the newest Windows version or in older Linux versions... The small dialog says: Failed to drop "FROMWSB" block!

--
You received this message because you are subscribed to the Google Groups "tumbi" group.

To unsubscribe from this group and stop receiving emails from it, send an email to tu...@googlegroups.com.

[Jean-Philippe Chancelier]

Dear Mariusz,

The way you provide values in the workspace is not the same in Nsp than in Scicoslab. You have an example in demos/Tests

the nsp/scicos model is called fromws.cos and the script fromws.sce can be used to create workspace data for the model fromws.cos.

In Nsp, the workspace values must be global values which are hash tables with a field time and a field values (which must be a cell).

First check that the demos works.

Regards

jpc

To unsubscribe from this group and stop receiving emails from it, send an email to tumbi+un...@googlegroups.com.

[Mariusz Grecki]

Dear Jean-Philippe,

in fact, I did not expect such changes between scicoslab and nsp, in particular, that I had impression scicos is the same in both.... But anyway, now it is clear that it must be done differently, but it still does not work... I tried to run the test you have suggested and the result is essentially exactly the same :(

Please look at the enclosed picture. I run the fromwsb.sce script, then scicos and scicos "Simulate" -> "Run" brought exactly the same error message. Now I am really puzzled. Do I make something wrong? The platform is VirtualBox with Ubuntu Xenial Xerus 16.04 64b and nsp https://cermics.enpc.fr/~jpc/nsp-tiddly/files/ubuntu/nsp-gtk3_1.3-1_amd64.xenial.deb

Regards

Mariusz

[Jean-Philippe Chancelier]

Dear  Mariusz,

I haven't tried to rerun the test. I'll make a try this week-end

Regards

jpc

To unsubscribe from this group and stop receiving emails from it, send an email to tumbi+un...@googlegroups.com.

[Jean-Philippe Chancelier]

Dear Mariusz,

I just make a try and it works for me. When compiling you have to say that the port size are [21,21] as commented in the schema. Then it work,

at least on my machine (ubuntu). The simulation is too fast and I have to slow down the graphics by setting the realtime scaling to 1 in the Simulate -> Setup menu.

Regards

jpc

To unsubscribe from this group and stop receiving emails from it, send an email to tumbi+un...@googlegroups.com.

[Mariusz Grecki]

Uff... it works. Sorry for taking your time. I promise for the future to read descriptions more carefully.

Merci beaucoup!

Mariusz

[Jean-Philippe Chancelier]

No worry, documentation is missing in nsp.
What do you intend to do with parallella ? do you think that we can accelerated functions in nsp with the help of the parallella chip ?
Regards
jpc

To unsubscribe from this group and stop receiving emails from it, send an email to tumbi+un...@googlegroups.com.

[Mariusz Grecki]

Dear Jean-Philippe,

concerning the parallella platform, I do not think it may compete now with high-performance solutions, like multi-core CPUs or modern GPUs. What parallela is still good is the performance/power factor. My interest in parallella is rather to learn how parallel architecture may be used in "everyday programming". I have some experience with GPUs but epiphany chip (this is the co-processing chip in parallella computer) has much different architecture than other designs. It is simple, focusing on power consumption and has numerous limitations that does not allow to treat the computing cores of epiphany chip like typical GPU or CPU cores. The main limitation there is the 32kB limit for local memory (for both data and code) that challenges the code and data partitioning. That's why I do not think about parallella as a high-performance platform for running nsp programs, rather I can see nsp as nice software platform for parallella helping to manage and run user algorithms, that are partially implemented on parallel epiphany cores and partially implemented with usage of reach nsp numerical libraries and running on parallella ARM CPU.

Regards.

Mariusz

[Mariusz Grecki]

One more problem related to FROM WORKSPACE block. The help says 'the "values" field can be Real,Complex, int8,int16,int32,uint8,uint16, and uint32'. I did not check all of them but certainly, complex type gives an error (enclosed figure). Is the complex type really allowed there? If it is allowed I would also expect blocks like 'real' and 'imag' and/or 'amp' and 'phase'. None of these blocks I could find in the palettes.