Parvfect / HelixWorks

Code for the Channels and Decoding Methods
0 stars 1 forks source link

Extending to 8C4 Symbols to get FER curve as a function of Read Length #18

Closed Parvfect closed 11 months ago

Parvfect commented 11 months ago

Currently works for base case, but symbol uncertainty is limited, so decoder does not have to do much. How does it perform for a pick of 4?

Image

Parvfect commented 11 months ago

Works for larger read lengths, here is the profiling

image
Parvfect commented 11 months ago

can't randomly pop - ffield has finite elements that can be accessed in G - have to pop from end

Parvfect commented 11 months ago

Why is lower read length slowing down decoder? Shouldn't it stil work the same - and max iterations end it's loop? Well no - cause the possibilities of each symbol skyrockets - so the recursive method skyrockets.

Parvfect commented 11 months ago

Adapting code for FER - getting curve for fixed H, G, C with channel changing shit Reads below 5 take forever, so getting preliminary curve from 5 onwards.

Parvfect commented 11 months ago

FER curve not functional - something about the check if it's correctly decoded. Going to sleep now

Parvfect commented 11 months ago

Goal for the next hour is to get the FER curve for this case. Current problem is first resolving the errors for a larger read length (should not be too tough since decoding is working, just need to test why FER isn't). And then fixing the small read error case.

Parvfect commented 11 months ago

image Wrong ffdim for G and C, but it's a start

Parvfect commented 11 months ago

Let's actually see what happens with small R for right ffdim - should have done small iterations first "sigh"

Even 7 Reads are taking forever and it's doing my head in. I'm better than this and I'm going to fix it

Parvfect commented 11 months ago

Done

Image