neqo v0.7.5

[KershawChang]

V0.7.5

[github-actions]

Benchmark results

Performance differences relative to c0ca26d.

• drain a timer quickly time: [419.22 ns 425.58 ns 431.52 ns]
  change: [-3.7620% -2.5749% -1.3033%] (p = 0.00 < 0.05)
  💚 Performance has improved.

• coalesce_acked_from_zero 1+1 entries
  time: [201.06 ns 201.49 ns 201.95 ns]
  change: [+2.3619% +2.7362% +3.0902%] (p = 0.00 < 0.05)
  💔 Performance has regressed.

• coalesce_acked_from_zero 3+1 entries
  time: [242.08 ns 242.71 ns 243.37 ns]
  change: [+1.2115% +1.7296% +2.2844%] (p = 0.00 < 0.05)
  💔 Performance has regressed.

• coalesce_acked_from_zero 10+1 entries
  time: [240.47 ns 241.17 ns 242.03 ns]
  change: [+0.9507% +1.4612% +1.9720%] (p = 0.00 < 0.05)
  Change within noise threshold.

• coalesce_acked_from_zero 1000+1 entries
  time: [219.94 ns 228.73 ns 248.12 ns]
  change: [+1.7692% +3.9592% +7.2683%] (p = 0.01 < 0.05)
  💔 Performance has regressed.

• RxStreamOrderer::inbound_frame()
  time: [119.02 ms 119.09 ms 119.17 ms]
  change: [+0.2199% +0.3425% +0.4638%] (p = 0.00 < 0.05)
  Change within noise threshold.

• transfer/Run multiple transfers with varying seeds
  time: [118.64 ms 118.93 ms 119.23 ms]
  thrpt: [33.548 MiB/s 33.632 MiB/s 33.715 MiB/s]
  change:
  time: [-1.5188% -1.1585% -0.8041%] (p = 0.00 < 0.05)
  thrpt: [+0.8106% +1.1721% +1.5422%]
  Change within noise threshold.

• transfer/Run multiple transfers with the same seed
  time: [119.56 ms 119.73 ms 119.92 ms]
  thrpt: [33.357 MiB/s 33.409 MiB/s 33.456 MiB/s]
  change:
  time: [-1.2507% -1.0407% -0.8255%] (p = 0.00 < 0.05)
  thrpt: [+0.8324% +1.0516% +1.2666%]
  Change within noise threshold.

• 1-conn/1-100mb-resp (aka. Download)/client
  time: [1.0832 s 1.0886 s 1.0945 s]
  thrpt: [91.366 MiB/s 91.862 MiB/s 92.317 MiB/s]
  change:
  time: [-2.9778% -2.0892% -1.2621%] (p = 0.00 < 0.05)
  thrpt: [+1.2782% +2.1338% +3.0692%]
  💚 Performance has improved.

• 1-conn/10_000-parallel-1b-resp (aka. RPS)/client
  time: [421.01 ms 423.24 ms 425.50 ms]
  thrpt: [23.502 Kelem/s 23.628 Kelem/s 23.752 Kelem/s]
  change:
  time: [-1.0975% -0.3274% +0.4519%] (p = 0.39 > 0.05)
  thrpt: [-0.4498% +0.3285% +1.1097%]
  No change in performance detected.

• 1-conn/1-1b-resp (aka. HPS)/client
  time: [51.652 ms 51.918 ms 52.195 ms]
  thrpt: [19.159 elem/s 19.261 elem/s 19.360 elem/s]
  change:
  time: [-1.5010% -0.6299% +0.2880%] (p = 0.17 > 0.05)
  thrpt: [-0.2872% +0.6339% +1.5238%]
  No change in performance detected.

Client/server transfer results

Transfer of 134217728 bytes over loopback.

Client	Server	CC	Pacing	Mean [ms]	Min [ms]	Max [ms]	Relative
msquic	msquic			741.3 ± 218.9	435.3	1057.4	1.00
neqo	msquic	reno	on	987.2 ± 217.8	818.3	1399.8	1.00
neqo	msquic	reno		1105.6 ± 281.0	781.1	1529.9	1.00
neqo	msquic	cubic	on	1095.1 ± 191.0	845.7	1425.4	1.00
neqo	msquic	cubic		1092.6 ± 258.1	842.7	1617.3	1.00
msquic	neqo	reno	on	4429.6 ± 128.1	4250.5	4606.5	1.00
msquic	neqo	reno		4417.2 ± 170.1	4174.0	4648.2	1.00
msquic	neqo	cubic	on	4452.0 ± 249.2	4158.8	4975.1	1.00
msquic	neqo	cubic		4516.5 ± 295.8	4174.1	4949.7	1.00
neqo	neqo	reno	on	3641.6 ± 263.5	3175.1	4002.3	1.00
neqo	neqo	reno		3415.7 ± 228.0	3095.3	3917.4	1.00
neqo	neqo	cubic	on	4276.7 ± 478.3	3085.4	4988.8	1.00
neqo	neqo	cubic		4265.1 ± 154.0	4017.7	4447.1	1.00

⬇️ Download logs