Wesley Lin

I found that WHAT I need is how to speed up checking, checking torrent of a big file(100G+) is very time-consuming.

I have looked into the code `torrent::start_checking()` I seems that `num_outstanding=max(aio_threads, checking_mem_usage something)` So I guess when I increasing the `aio_threads`, the checking file threads should be increased too. But...

@master255 @arvidn Thanks, but my files are always new for seeding, so I am looking for the way to speed up first-time file checking . I found that hashing_threads is...

@AllSeeingEyeTolledEweSew `flag_seed_mode` works, thanks. by the way `set_piece_hashes` also a time killer, maybe it is impossible to rewrite it in my code instead of calling it directly, I will have...

I am creating a torrent, now `set_piece_hashes` takes me nearly 20min with a 200G file.

I think that is limited by disk I/O. I copy the [set_piece_hashes](https://github.com/arvidn/libtorrent/blob/libtorrent-1_1_14/src/create_torrent.cpp#L260) function out, and modify the `disk_thread.set_num_threads(1);` to a bigger number, along with a bigger `aio_threads`, that can really...

TCP recv-Q is high, so I guess the increasing incoming/outgoing UDP maybe affect the TCP receiving performance. Maybe I can try to reduce the uploading by reducing the unchoke slots...

I think limit the connections from the same peer may help. But I only find `connections_limit` to controller the global connections, maybe I miss some options like `connections_limit_per_peer`?（by default I...

> finally I found the option `network_threads`, which can increase the working thread num from 1 to N. That solve my problem. oh no yet

the bottle neck is absolutely in recving&sending, only one thread run all the actions