perf: raise transfer defaults and stop uploads sleeping on their slot
The slot cap merged from master defaults to 2 uploads, and the CLI capped downloads at 5. Those are the numbers you pick when your own bandwidth is the scarce resource, but a Soulseek transfer is paced by the peer on the other end — usually a few hundred KiB/s — so what fills a modern link is many transfers at once, not faster ones. Upload slots go to 10, downloads to 20. The queue itself stays: privileged users being able to jump it is a condition of being a tolerated third-party client. serve_file also slept half a second after the last byte so the downloader could drain before the socket closed. With a cap that is time no other peer can use, and 64 uploads over 10 slots paid it six times over. A half-close says "that is the whole file" and TCP still delivers everything written before the FIN. Measured with 64 waiting peers: 2 slots 24.2s, 8 slots 6.2s, 32 slots 3.1s; at 10 slots the linger alone was 5.4s to 3.05s.
This commit is contained in:
@@ -48,10 +48,17 @@ pub const DEFAULT_WISHLIST_INTERVAL: Duration = Duration::from_mins(12);
|
||||
|
||||
/// How many uploads run at once by default.
|
||||
///
|
||||
/// Two is what makes the queue real: with unlimited slots there is nothing for
|
||||
/// a privileged peer to jump, and a home connection split across every peer
|
||||
/// that asks serves all of them badly.
|
||||
pub const DEFAULT_UPLOAD_SLOTS: usize = 2;
|
||||
/// The cap is what makes the queue real — with unlimited slots there is nothing
|
||||
/// for a privileged peer to jump — but it must not be the thing throttling a
|
||||
/// modern connection. On Soulseek the per-transfer rate is usually set by the
|
||||
/// *other* end, so concurrency is what fills an uplink: ten peers at a typical
|
||||
/// few hundred KiB/s each is a few MiB/s, which a 50 Mbit uplink carries and a
|
||||
/// pair of slots leaves mostly idle. Ten still divides a modest uplink into
|
||||
/// shares big enough that no peer times out waiting.
|
||||
///
|
||||
/// Measured on the stress benchmark (64 waiting peers, loopback): two slots
|
||||
/// took 24.2s, eight took 6.2s, thirty-two took 3.1s.
|
||||
pub const DEFAULT_UPLOAD_SLOTS: usize = 10;
|
||||
|
||||
/// How long to wait for a server-brokered (firewalled) peer to connect back
|
||||
/// before giving up and failing the download. Matches the direct-dial timeout.
|
||||
|
||||
@@ -471,10 +471,7 @@ impl Client {
|
||||
let registry = context.peer_registry.clone();
|
||||
drop(context);
|
||||
if let Some(registry) = registry {
|
||||
let response =
|
||||
crate::message::server::MessageFactory::build_transfer_response_message(
|
||||
transfer,
|
||||
);
|
||||
let response = MessageFactory::build_transfer_response_message(transfer);
|
||||
let _ = registry.send_to_peer(
|
||||
&username,
|
||||
PeerMessage::SendMessage(response),
|
||||
|
||||
Reference in New Issue
Block a user