Detection flag as the output of the decoder

[pejmanmemar]

Hi,

First, I wanted to thank you for this great work. I have a question regarding the possible outputs of the provided decoders. Is it possible to receive a flag that the detection has happened besides the decoded code word? In other words, a flag that shows the decoder could detect the error, and the syndrome was nonzero.

Regards

[rtajan]

Hi,

Thank you very much for your interest in our job.

If I understand well your request, what you would like is a socket containing the syndrome after decoding.

I think that this is the purpose of the socket named CWD of decode_siho task for example.

My feeling comes after looking at the code of the decode_siho task for the Decoder_LDPC_BP_flooding decoder here.
CWD is filled with the complement of status and so it is the syndrome computed there.

Maybe @kouchy may have more information about the CWD socket.

Regards

[kouchy]

Hi,

I confirm what @rtajan said.

I would advice to use the CWD socket more that the status. It is true that the information is also contained in the status however the CWD socket is more explicit.

Regards

[pejmanmemar]

Hi

Thank you for your swift reply!
I tried the CWD socket in both decode_hiho and decode_siho. For Reed-Solomon Code, even in case of a corrupted code word, the CWD socket (also status socket) always returns 0. Is there any solution for that?

[Sciroccogti]

If you are using custom module like the example below:

class py_modulator(Py_Module):

    # Step 2
    def bpsk(self, b, x): # Your original code (independently from aff3ct)
        x[:] = 1.0 - 2.0*b[:]
        return 0
    
    # Step 3
    def __init__(self, N):
        # __init__ (step 3.1)
        Py_Module.__init__(self) # Call the aff3ct Py_Module __init__
        self.name = "py_BPSK"   # Set your module's name

        # __init__ (step 3.2)
        t_mod = self.create_task("modulate") # create a task for your module
        
        # __init__ (step 3.3)
        sb = self.create_socket_in (t_mod, "b", N, np.int32  ) # create an input socket for the task t_mod
        sx = self.create_socket_out(t_mod, "x", N, np.float32) # create an output socket for the task t_mod
    
        # __init__ (step 3.4)
        self.create_codelet(t_mod, lambda slf, lsk, fid: slf.bpsk(lsk[sb], lsk[sx])) # create codelet

When you change x[:] = 1.0 - 2.0*b[:] into x = 1.0 - 2.0*b[:], x will stay 0 forever and 0 is always returned.

Not sure if this can answer your question, but me myself had stuck in this for hours.