Linux, Tixati 3.44 shows Remote download speed: Rem B/s=6.80 MiB/s.
Copied from peer's log window with piece status enabled: ...
...220 seconds, remote completed 99 pieces, 128M piece size
12672 MB total in 220s calculates to 57.6 MB/s. It doesn't add up.
It's funny because Tixati itself uploads at over 40-45 MiB/s bandwidth to the remote peer. Either the pieces enter a blackhole on the other end or Tixati is wrong about either of the numbers :-) Maybe the big piece size results in a wrong calculation.
Thank you for creating Tixati!
by Guest on 2026/07/06 12:11:11 PM
IIRC, the client calculates that based on a list of finished pieces peer reports, there is no dedicated field in protocols for that. Peers might not update that list often enough or in expected fashion, or disconnect prematurely.
A real controlled test is probably required.
Your protocol description is correct. Practically, it was close to a controlled test.
The piece completion messages came from Tixati peer log. This means Tixati can't correctly calculate remote speed based on received HAVE messages from remote peer.
Peer's HAVE messages amounted to more than my upload speed. That's expected in presence of other seeds. But Tixati shouldn't show me a slower remote speed than: remote finished piece count * piece size / time elapsed.
PS: I set output formatting to show MiB units. The factor of approximately 8 is suspicious... bit to byte conversion?
