Closing Notes #57
Description
I profiled the key generating program with gprof to find out why it was so slow, it seems like the bulk of the time is spent in function calls in the secp256k1 library, which I think is part of the bitcoin library I included. I also tested the cache performance with perf, and it seems alright in that regard, although there's a few things I would change.
I considered using pthreads for the section that heavily uses secp256k1 functions, since we can't do much else about that, however the block of code is really big and it uses a lot of variables defined in greater scopes, so it seems like a hassle. Also, the way the program works is generally not that great because we don't even have a fast "weak random seed" generator.
Also, at some point a few months back I looked at libwebsockets since there was a persistent loss of data on the socket stream with blockchain.com. I'm not sure what the issue is, but it seems to be in the library code and not my own, and it's not worth figuring out for something like this.
Anyways, bitcoin is dope.