Revert "perf(udp): multi-message receive on apple (#2302)" #2352
Conversation
This reverts commit 3c48567.
mxinden
requested review from
KershawChang,
martinthomson and
larseggert
as code owners
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #2352 +/- ##
==========================================
+ Coverage 95.29% 95.30% +0.01%
==========================================
Files 114 114
Lines 36844 36858 +14
Branches 36844 36858 +14
==========================================
+ Hits 35110 35128 +18
+ Misses 1728 1724 -4
Partials 6 6 ☔ View full report in Codecov by Sentry. |
Failed Interop TestsQUIC Interop Runner, client vs. server, differences relative to 085fa62. neqo-latest as client
neqo-latest as server
All resultsSucceeded Interop TestsQUIC Interop Runner, client vs. server neqo-latest as client
neqo-latest as server
Unsupported Interop TestsQUIC Interop Runner, client vs. server neqo-latest as client
neqo-latest as server
|
Benchmark resultsPerformance differences relative to 085fa62. decode 4096 bytes, mask ff: No change in performance detected. time: [11.176 µs 11.223 µs 11.282 µs]
change: [-0.0450% +0.6787% +1.4128%] (p = 0.08 > 0.05)
decode 1048576 bytes, mask ff: No change in performance detected. time: [3.0161 ms 3.0258 ms 3.0369 ms]
change: [-1.0595% -0.3860% +0.2125%] (p = 0.26 > 0.05)
decode 4096 bytes, mask 7f: No change in performance detected. time: [19.534 µs 19.587 µs 19.644 µs]
change: [-0.3124% +0.0163% +0.3351%] (p = 0.93 > 0.05)
decode 1048576 bytes, mask 7f: No change in performance detected. time: [5.1607 ms 5.1723 ms 5.1854 ms]
change: [-1.3430% -0.4242% +0.2013%] (p = 0.35 > 0.05)
decode 4096 bytes, mask 3f: No change in performance detected. time: [5.5291 µs 5.5529 µs 5.5830 µs]
change: [-0.5999% +1.5931% +5.3857%] (p = 0.59 > 0.05)
decode 1048576 bytes, mask 3f: No change in performance detected. time: [1.7580 ms 1.7596 ms 1.7625 ms]
change: [-0.0153% +0.0744% +0.2459%] (p = 0.52 > 0.05)
coalesce_acked_from_zero 1+1 entries: No change in performance detected. time: [99.329 ns 99.656 ns 99.989 ns]
change: [-0.0878% +0.3087% +0.7104%] (p = 0.13 > 0.05)
coalesce_acked_from_zero 3+1 entries: No change in performance detected. time: [117.27 ns 117.69 ns 118.17 ns]
change: [-0.1934% +0.3516% +0.8787%] (p = 0.22 > 0.05)
coalesce_acked_from_zero 10+1 entries: No change in performance detected. time: [116.58 ns 116.91 ns 117.34 ns]
change: [-0.2761% +0.2147% +0.7272%] (p = 0.42 > 0.05)
coalesce_acked_from_zero 1000+1 entries: No change in performance detected. time: [97.317 ns 97.433 ns 97.564 ns]
change: [-1.0142% -0.1978% +0.5591%] (p = 0.65 > 0.05)
RxStreamOrderer::inbound_frame(): Change within noise threshold. time: [111.04 ms 111.10 ms 111.16 ms]
change: [-0.3343% -0.1769% -0.0504%] (p = 0.01 < 0.05)
SentPackets::take_ranges: No change in performance detected. time: [5.5160 µs 5.6975 µs 5.8932 µs]
change: [-1.9785% +0.7002% +3.5085%] (p = 0.64 > 0.05)
transfer/pacing-false/varying-seeds: Change within noise threshold. time: [41.575 ms 41.668 ms 41.772 ms]
change: [-0.7369% -0.4437% -0.1270%] (p = 0.00 < 0.05)
transfer/pacing-true/varying-seeds: No change in performance detected. time: [42.042 ms 42.123 ms 42.212 ms]
change: [-0.1417% +0.1567% +0.4511%] (p = 0.30 > 0.05)
transfer/pacing-false/same-seed: Change within noise threshold. time: [41.552 ms 41.628 ms 41.712 ms]
change: [-1.1799% -0.9313% -0.6763%] (p = 0.00 < 0.05)
transfer/pacing-true/same-seed: Change within noise threshold. time: [41.659 ms 41.726 ms 41.801 ms]
change: [+0.3295% +0.5793% +0.8316%] (p = 0.00 < 0.05)
1-conn/1-100mb-resp/mtu-1504 (aka. Download)/client: 💚 Performance has improved. time: [891.10 ms 899.58 ms 908.39 ms]
thrpt: [110.08 MiB/s 111.16 MiB/s 112.22 MiB/s]
change:
time: [-4.2354% -2.8237% -1.2988%] (p = 0.00 < 0.05)
thrpt: [+1.3159% +2.9057% +4.4227%]
1-conn/10_000-parallel-1b-resp/mtu-1504 (aka. RPS)/client: No change in performance detected. time: [314.87 ms 316.53 ms 318.19 ms]
thrpt: [31.428 Kelem/s 31.593 Kelem/s 31.759 Kelem/s]
change:
time: [-1.5746% -0.7046% +0.1574%] (p = 0.11 > 0.05)
thrpt: [-0.1571% +0.7096% +1.5998%]
1-conn/1-1b-resp/mtu-1504 (aka. HPS)/client: No change in performance detected. time: [33.933 ms 34.142 ms 34.372 ms]
thrpt: [29.093 elem/s 29.289 elem/s 29.470 elem/s]
change:
time: [-1.3486% -0.4978% +0.2746%] (p = 0.24 > 0.05)
thrpt: [-0.2739% +0.5003% +1.3670%]
1-conn/1-100mb-resp/mtu-1504 (aka. Upload)/client: Change within noise threshold. time: [1.6633 s 1.6825 s 1.7014 s]
thrpt: [58.774 MiB/s 59.435 MiB/s 60.122 MiB/s]
change:
time: [-3.1501% -1.5471% -0.1051%] (p = 0.04 < 0.05)
thrpt: [+0.1052% +1.5714% +3.2526%]
Client/server transfer resultsTransfer of 33554432 bytes over loopback.
|
|
I wonder if I messed something up on the quinn-udp side. Off-by-one error on TX or RX when dealing with the buffer vector? Hm. |
|
@larseggert but that would not explain why it is failing for me on Linux, right? |
|
Debugging this further, the neqo/neqo-transport/src/streams.rs Lines 420 to 422 in 8bc926a |
|
Ah, I believe neqo/neqo-transport/src/streams.rs Lines 416 to 422 in 8bc926a Say that the stream is remote initiated, then I will create a pull request. Edit: #2358 |
|
Closing here in favor of #2358. |
This reverts commit 3c48567.
RPSbenchmark sometimes fails withTransportError(PeerError(5). Git bisect shows 3c48567 introduced the regression.Until I figure out the reason, let's revert.