Skip to content

fix: Improve mark_as_acked #3056

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
larseggert wants to merge 3 commits into from
Closed

fix: Improve mark_as_acked #3056

larseggert wants to merge 3 commits into from
+6 −4

Conversation

@larseggert
Copy link
Collaborator

@larseggert larseggert commented Oct 16, 2025

Or at least try.

@larseggert
fix: Improve mark_as_acked
4d3f902 
Or at least try.
@codecov
Copy link

codecov bot commented Oct 16, 2025 •
edited
Loading

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.40%. Comparing base (b9c32c7) to head (eef6523).
⚠️ Report is 93 commits behind head on main.

Additional details and impacted files 
@@            Coverage Diff             @@
##             main    #3056      +/-   ##
==========================================
- Coverage   93.41%   93.40%   -0.01%     
==========================================
  Files         124      124              
  Lines       36234    36233       -1     
  Branches    36234    36233       -1     
==========================================
- Hits        33847    33844       -3     
  Misses       1540     1540              
- Partials      847      849       +2
Components Coverage Δ
neqo-common 97.32% <ø> (ø)
neqo-crypto 83.25% <ø> (-0.48%) ⬇️
neqo-http3 93.29% <ø> (ø)
neqo-qpack 94.18% <ø> (ø)
neqo-transport 94.53% <100.00%> (+0.05%) ⬆️
neqo-udp 78.94% <ø> (-0.48%) ⬇️
mtu 85.76% <ø> (ø)

@github-actions
Copy link
Contributor

github-actions bot commented Oct 16, 2025 •
edited
Loading

🐰 Bencher Report

Branchfix-mark_as_acked-opt
TestbedOn-prem
Click to view all benchmark results
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
google vs. neqo (cubic, paced)📈 view plot
🚷 view threshold		279.02 ms
(+0.26%)Baseline: 278.31 ms
282.67 ms
(98.71%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
msquic vs. neqo (cubic, paced)📈 view plot
🚷 view threshold		216.98 ms
(+9.06%)Baseline: 198.95 ms
236.31 ms
(91.82%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
neqo vs. google (cubic, paced)📈 view plot
🚷 view threshold		758.15 ms
(-0.19%)Baseline: 759.61 ms
774.57 ms
(97.88%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
neqo vs. msquic (cubic, paced)📈 view plot
🚷 view threshold		157.38 ms
(-0.25%)Baseline: 157.78 ms
160.62 ms
(97.99%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
neqo vs. neqo (cubic)📈 view plot
🚷 view threshold		93.45 ms
(+2.11%)Baseline: 91.53 ms
96.85 ms
(96.50%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
neqo vs. neqo (cubic, paced)📈 view plot
🚷 view threshold		94.05 ms
(+1.26%)Baseline: 92.88 ms
98.08 ms
(95.89%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
neqo vs. neqo (reno)📈 view plot
🚷 view threshold		91.39 ms
(-0.12%)Baseline: 91.50 ms
96.65 ms
(94.56%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
neqo vs. neqo (reno, paced)📈 view plot
🚷 view threshold		92.77 ms
(+0.01%)Baseline: 92.76 ms
97.76 ms
(94.90%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
neqo vs. quiche (cubic, paced)📈 view plot
🚷 view threshold		195.11 ms
(+0.76%)Baseline: 193.63 ms
196.96 ms
(99.06%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
neqo vs. s2n (cubic, paced)📈 view plot
🚷 view threshold		220.50 ms
(-0.28%)Baseline: 221.12 ms
224.08 ms
(98.40%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
quiche vs. neqo (cubic, paced)📈 view plot
🚷 view threshold		154.38 ms
(+0.85%)Baseline: 153.07 ms
158.35 ms
(97.50%)
BenchmarkLatencyBenchmark Result
milliseconds (ms)
(Result Δ%)		Upper Boundary
milliseconds (ms)
(Limit %)
s2n vs. neqo (cubic, paced)📈 view plot
🚷 view threshold		170.73 ms
(-1.76%)Baseline: 173.79 ms
178.01 ms
(95.91%)
🐰 View full continuous benchmarking report in Bencher

@github-actions
Copy link
Contributor

github-actions bot commented Oct 16, 2025 •
edited
Loading

🐰 Bencher Report

Branchfix-mark_as_acked-opt
TestbedOn-prem
Click to view all benchmark results
BenchmarkLatencyBenchmark Result
nanoseconds (ns)
(Result Δ%)		Upper Boundary
nanoseconds (ns)
(Limit %)
1-conn/1-100mb-req/mtu-1504 (aka. Upload)/client📈 view plot
🚷 view threshold		206,680,000.00 ns
(-0.03%)Baseline: 206,752,325.58 ns
216,458,370.24 ns
(95.48%)
1-conn/1-100mb-resp/mtu-1504 (aka. Download)/client📈 view plot
🚷 view threshold		199,000,000.00 ns
(-0.94%)Baseline: 200,895,174.42 ns
211,171,780.28 ns
(94.24%)
1-conn/1-1b-resp/mtu-1504 (aka. HPS)/client📈 view plot
🚷 view threshold		38,694,000.00 ns
(+22.29%)Baseline: 31,641,276.16 ns
42,885,724.77 ns
(90.23%)
1-conn/10_000-parallel-1b-resp/mtu-1504 (aka. RPS)/client📈 view plot
🚷 view threshold		289,460,000.00 ns
(-0.66%)Baseline: 291,387,209.30 ns
303,642,017.47 ns
(95.33%)
1-streams/each-1000-bytes/simulated-time📈 view plot
🚷 view threshold		119,230,000.00 ns
(+0.47%)Baseline: 118,673,575.58 ns
120,713,746.14 ns
(98.77%)
1-streams/each-1000-bytes/wallclock-time📈 view plot
🚷 view threshold		580,530.00 ns
(-2.04%)Baseline: 592,644.62 ns
615,533.84 ns
(94.31%)
1000-streams/each-1-bytes/simulated-time📈 view plot
🚷 view threshold		2,330,900,000.00 ns
(-82.76%)Baseline: 13,523,404,069.77 ns
23,011,158,648.63 ns
(10.13%)
1000-streams/each-1-bytes/wallclock-time📈 view plot
🚷 view threshold		12,555,000.00 ns
(-9.32%)Baseline: 13,844,950.58 ns
15,082,092.84 ns
(83.24%)
1000-streams/each-1000-bytes/simulated-time📈 view plot
🚷 view threshold		16,354,000,000.00 ns
(-12.39%)Baseline: 18,665,799,418.60 ns
20,609,155,466.95 ns
(79.35%)
1000-streams/each-1000-bytes/wallclock-time📈 view plot
🚷 view threshold		49,630,000.00 ns
(-2.05%)Baseline: 50,669,459.30 ns
57,081,615.58 ns
(86.95%)
RxStreamOrderer::inbound_frame()📈 view plot
🚷 view threshold		108,150,000.00 ns
(-1.41%)Baseline: 109,691,191.86 ns
111,596,019.96 ns
(96.91%)
coalesce_acked_from_zero 1+1 entries📈 view plot
🚷 view threshold		89.52 ns
(+0.73%)Baseline: 88.87 ns
90.13 ns
(99.32%)
coalesce_acked_from_zero 10+1 entries📈 view plot
🚷 view threshold		106.17 ns
(+0.09%)Baseline: 106.08 ns
107.22 ns
(99.02%)
coalesce_acked_from_zero 1000+1 entries📈 view plot
🚷 view threshold		91.50 ns
(+1.31%)Baseline: 90.32 ns
94.94 ns
(96.38%)
coalesce_acked_from_zero 3+1 entries📈 view plot
🚷 view threshold		106.21 ns
(-0.36%)Baseline: 106.59 ns
107.68 ns
(98.64%)
decode 1048576 bytes, mask 3f📈 view plot
🚷 view threshold		1,761,000.00 ns
(+7.74%)Baseline: 1,634,541.86 ns
1,808,886.75 ns
(97.35%)
decode 1048576 bytes, mask 7f📈 view plot
🚷 view threshold		5,045,400.00 ns
(-0.42%)Baseline: 5,066,546.80 ns
5,113,075.77 ns
(98.68%)
decode 1048576 bytes, mask ff📈 view plot
🚷 view threshold		3,006,100.00 ns
(-0.77%)Baseline: 3,029,460.17 ns
3,053,652.70 ns
(98.44%)
decode 4096 bytes, mask 3f📈 view plot
🚷 view threshold		6,233.30 ns
(-15.31%)Baseline: 7,360.25 ns
10,378.43 ns
(60.06%)
decode 4096 bytes, mask 7f📈 view plot
🚷 view threshold		19,593.00 ns
(-1.06%)Baseline: 19,802.90 ns
20,466.00 ns
(95.73%)
decode 4096 bytes, mask ff📈 view plot
🚷 view threshold		11,392.00 ns
(+0.23%)Baseline: 11,365.60 ns
12,523.87 ns
(90.96%)
sent::Packets::take_ranges📈 view plot
🚷 view threshold		4,634.10 ns
(-1.84%)Baseline: 4,720.99 ns
4,959.49 ns
(93.44%)
transfer/pacing-false/same-seed/simulated-time/run📈 view plot
🚷 view threshold		25,234,000,000.00 ns
(-0.70%)Baseline: 25,410,716,374.27 ns
26,030,182,060.52 ns
(96.94%)
transfer/pacing-false/same-seed/wallclock-time/run📈 view plot
🚷 view threshold		24,373,000.00 ns
(-5.12%)Baseline: 25,687,356.73 ns
27,106,753.12 ns
(89.91%)
transfer/pacing-false/varying-seeds/simulated-time/run📈 view plot
🚷 view threshold		25,200,000,000.00 ns
(+0.10%)Baseline: 25,175,423,976.61 ns
25,224,686,559.53 ns
(99.90%)
transfer/pacing-false/varying-seeds/wallclock-time/run📈 view plot
🚷 view threshold		24,704,000.00 ns
(-4.14%)Baseline: 25,770,286.55 ns
27,415,048.33 ns
(90.11%)
transfer/pacing-true/same-seed/simulated-time/run📈 view plot
🚷 view threshold		25,301,000,000.00 ns
(-1.08%)Baseline: 25,578,292,397.66 ns
25,883,203,543.21 ns
(97.75%)
transfer/pacing-true/same-seed/wallclock-time/run📈 view plot
🚷 view threshold		26,101,000.00 ns
(-3.44%)Baseline: 27,031,976.61 ns
28,606,041.22 ns
(91.24%)
transfer/pacing-true/varying-seeds/simulated-time/run📈 view plot
🚷 view threshold		24,997,000,000.00 ns
(+0.01%)Baseline: 24,995,336,257.31 ns
25,043,987,836.67 ns
(99.81%)
transfer/pacing-true/varying-seeds/wallclock-time/run📈 view plot
🚷 view threshold		24,932,000.00 ns
(-5.10%)Baseline: 26,270,652.05 ns
27,999,795.18 ns
(89.04%)
🐰 View full continuous benchmarking report in Bencher

@larseggert larseggert marked this pull request as ready for review October 16, 2025 14:31
Copilot AI review requested due to automatic review settings October 16, 2025 14:31
Copilot AI reviewed Oct 16, 2025
View reviewed changes
Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

This PR improves the mark_as_acked method in the TxBuffer implementation by replacing an inefficient rotate_left + truncate operation with a more efficient drain operation for removing acknowledged bytes from the front of a VecDeque buffer.

  • Replaces rotate_left() + truncate() with drain() for better performance
  • Adds detailed comments explaining the performance benefits of the new approach
  • Maintains the same functionality while improving cache efficiency

Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.

@larseggert
Update neqo-transport/src/send_stream.rs
2b39584 
Co-authored-by: Copilot <[email protected]>
Signed-off-by: Lars Eggert <[email protected]>
Copilot AI review requested due to automatic review settings October 16, 2025 14:34
Copilot AI reviewed Oct 16, 2025
View reviewed changes
Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.

mxinden
mxinden reviewed Oct 17, 2025
View reviewed changes
Copy link
Member

@mxinden mxinden left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I tried this out in #2680. I was not able to get a clear performance improvement signal.

I don't think rotate_left "always performs a full rotation". Depending on previous operations on the VecDeque, it might simply update a pointer. See this #2680 (comment) for a detailed description.

@codspeed-hq
Copy link

codspeed-hq bot commented Nov 12, 2025

CodSpeed Performance Report

Merging #3056 will improve performances by 22.36%

Comparing larseggert:fix-mark_as_acked-opt (eef6523) with main (b9c32c7)

Summary

⚡ 2 improvements
✅ 21 untouched

Benchmarks breakdown

Mode Benchmark BASE HEAD Change
⚡ Simulation client 845.7 ms 691.2 ms +22.36%
⚡ Simulation client 852.3 ms 766.3 ms +11.22%

@github-actions
Copy link
Contributor

github-actions bot commented Nov 12, 2025

Failed Interop Tests

QUIC Interop Runner, client vs. server, differences relative to b9c32c7.

neqo-latest as client

neqo-latest as server

All results

Succeeded Interop Tests

QUIC Interop Runner, client vs. server

neqo-latest as client

neqo-latest as server

Unsupported Interop Tests

QUIC Interop Runner, client vs. server

neqo-latest as client

neqo-latest as server

@github-actions
Copy link
Contributor

github-actions bot commented Nov 12, 2025

Benchmark results

Performance differences relative to b9c32c7.

1-conn/1-100mb-resp/mtu-1504 (aka. Download)/client: Change within noise threshold.
       time:   [198.49 ms 199.00 ms 199.52 ms]
       thrpt:  [501.21 MiB/s 502.52 MiB/s 503.81 MiB/s]
change:
       time:   [-1.5209% -1.1745% -0.8448%] (p = 0.00 < 0.05)
       thrpt:  [+0.8520% +1.1885% +1.5444%]

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high mild

1-conn/10_000-parallel-1b-resp/mtu-1504 (aka. RPS)/client: Change within noise threshold.
       time:   [287.76 ms 289.46 ms 291.18 ms]
       thrpt:  [34.343 Kelem/s 34.547 Kelem/s 34.751 Kelem/s]
change:
       time:   [+0.6651% +1.4903% +2.3611%] (p = 0.00 < 0.05)
       thrpt:  [-2.3066% -1.4684% -0.6607%]

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high mild

1-conn/1-1b-resp/mtu-1504 (aka. HPS)/client: No change in performance detected.
       time:   [38.564 ms 38.694 ms 38.842 ms]
       thrpt:  [25.746   B/s 25.844   B/s 25.931   B/s]
change:
       time:   [-1.0880% -0.4442% +0.1598%] (p = 0.17 > 0.05)
       thrpt:  [-0.1596% +0.4462% +1.1000%]

Found 4 outliers among 100 measurements (4.00%)

4 (4.00%) high severe

1-conn/1-100mb-req/mtu-1504 (aka. Upload)/client: No change in performance detected.
       time:   [206.39 ms 206.68 ms 206.99 ms]
       thrpt:  [483.12 MiB/s 483.85 MiB/s 484.52 MiB/s]
change:
       time:   [-0.2502% -0.0529% +0.1373%] (p = 0.60 > 0.05)
       thrpt:  [-0.1371% +0.0530% +0.2508%]

Found 4 outliers among 100 measurements (4.00%)

3 (3.00%) high mild

1 (1.00%) high severe

decode 4096 bytes, mask ff: No change in performance detected.
       time:   [11.345 µs 11.392 µs 11.443 µs]
       change: [-0.0593% +0.5699% +1.2459%] (p = 0.10 > 0.05)

Found 22 outliers among 100 measurements (22.00%)

2 (2.00%) low severe

4 (4.00%) low mild

16 (16.00%) high severe

decode 1048576 bytes, mask ff: No change in performance detected.
       time:   [2.9933 ms 3.0061 ms 3.0223 ms]
       change: [-1.4035% -0.3551% +0.5513%] (p = 0.51 > 0.05)

Found 10 outliers among 100 measurements (10.00%)

1 (1.00%) high mild

9 (9.00%) high severe

decode 4096 bytes, mask 7f: No change in performance detected.
       time:   [19.551 µs 19.593 µs 19.642 µs]
       change: [-6.6155% -2.3772% +0.0824%] (p = 0.26 > 0.05)

Found 16 outliers among 100 measurements (16.00%)

2 (2.00%) low severe

5 (5.00%) low mild

1 (1.00%) high mild

8 (8.00%) high severe

decode 1048576 bytes, mask 7f: No change in performance detected.
       time:   [5.0302 ms 5.0454 ms 5.0638 ms]
       change: [-0.7014% -0.1538% +0.3940%] (p = 0.59 > 0.05)

Found 14 outliers among 100 measurements (14.00%)

1 (1.00%) high mild

13 (13.00%) high severe

decode 4096 bytes, mask 3f: No change in performance detected.
       time:   [6.2069 µs 6.2333 µs 6.2671 µs]
       change: [-1.3761% +0.7968% +4.4202%] (p = 0.70 > 0.05)

Found 11 outliers among 100 measurements (11.00%)

3 (3.00%) low mild

2 (2.00%) high mild

6 (6.00%) high severe

decode 1048576 bytes, mask 3f: No change in performance detected.
       time:   [1.7579 ms 1.7610 ms 1.7673 ms]
       change: [+0.0005% +0.1835% +0.5370%] (p = 0.24 > 0.05)

Found 2 outliers among 100 measurements (2.00%)

1 (1.00%) high mild

1 (1.00%) high severe

1-streams/each-1000-bytes/wallclock-time: Change within noise threshold.
       time:   [578.64 µs 580.53 µs 582.72 µs]
       change: [-2.0295% -1.4732% -0.9183%] (p = 0.00 < 0.05)

Found 6 outliers among 100 measurements (6.00%)

6 (6.00%) high severe

1-streams/each-1000-bytes/simulated-time

time:   [119.01 ms 119.23 ms 119.45 ms]

thrpt:  [8.1757 KiB/s 8.1907 KiB/s 8.2059 KiB/s]

change:

time:   [-0.1726% +0.0901% +0.3445%] (p = 0.51 > 0.05)

thrpt:  [-0.3433% -0.0900% +0.1729%]

No change in performance detected.

1000-streams/each-1-bytes/wallclock-time: 💚 Performance has improved.
       time:   [12.516 ms 12.555 ms 12.595 ms]
       change: [-1.8889% -1.4423% -1.0138%] (p = 0.00 < 0.05)
1000-streams/each-1-bytes/simulated-time: No change in performance detected.
       time:   [2.3274 s 2.3309 s 2.3344 s]
       thrpt:  [428.37   B/s 429.02   B/s 429.67   B/s]
change:
       time:   [-0.3595% -0.1247% +0.1066%] (p = 0.28 > 0.05)
       thrpt:  [-0.1065% +0.1248% +0.3608%]
1000-streams/each-1000-bytes/wallclock-time: 💚 Performance has improved.
       time:   [49.530 ms 49.630 ms 49.729 ms]
       change: [-1.8133% -1.5147% -1.2216%] (p = 0.00 < 0.05)

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high mild

1000-streams/each-1000-bytes/simulated-time

time:   [16.127 s 16.354 s 16.585 s]

thrpt:  [58.882 KiB/s 59.714 KiB/s 60.555 KiB/s]

change:

time:   [-2.8897% -0.5609% +1.7818%] (p = 0.64 > 0.05)

thrpt:  [-1.7506% +0.5640% +2.9757%]

No change in performance detected.

coalesce_acked_from_zero 1+1 entries: No change in performance detected.
       time:   [89.218 ns 89.518 ns 89.818 ns]
       change: [-0.5183% +0.0655% +0.6084%] (p = 0.83 > 0.05)

Found 11 outliers among 100 measurements (11.00%)

9 (9.00%) high mild

2 (2.00%) high severe

coalesce_acked_from_zero 3+1 entries: No change in performance detected.
       time:   [105.88 ns 106.21 ns 106.55 ns]
       change: [-0.4634% -0.0494% +0.3944%] (p = 0.82 > 0.05)

Found 16 outliers among 100 measurements (16.00%)

1 (1.00%) low mild

4 (4.00%) high mild

11 (11.00%) high severe

coalesce_acked_from_zero 10+1 entries: No change in performance detected.
       time:   [105.68 ns 106.17 ns 106.76 ns]
       change: [-0.8417% +0.3794% +1.7740%] (p = 0.61 > 0.05)

Found 17 outliers among 100 measurements (17.00%)

1 (1.00%) low severe

4 (4.00%) low mild

1 (1.00%) high mild

11 (11.00%) high severe

coalesce_acked_from_zero 1000+1 entries: No change in performance detected.
       time:   [91.390 ns 91.502 ns 91.630 ns]
       change: [-0.5768% +0.0949% +0.7523%] (p = 0.78 > 0.05)

Found 13 outliers among 100 measurements (13.00%)

6 (6.00%) high mild

7 (7.00%) high severe

RxStreamOrderer::inbound_frame(): 💚 Performance has improved.
       time:   [108.09 ms 108.15 ms 108.22 ms]
       change: [-2.2157% -1.9440% -1.7724%] (p = 0.00 < 0.05)

Found 21 outliers among 100 measurements (21.00%)

4 (4.00%) low severe

7 (7.00%) low mild

8 (8.00%) high mild

2 (2.00%) high severe

sent::Packets::take_ranges: No change in performance detected.
       time:   [4.5195 µs 4.6341 µs 4.7421 µs]
       change: [-2.0541% +1.2917% +4.6908%] (p = 0.47 > 0.05)

Found 2 outliers among 100 measurements (2.00%)

2 (2.00%) high mild

transfer/pacing-false/varying-seeds/wallclock-time/run: Change within noise threshold.
       time:   [24.663 ms 24.704 ms 24.749 ms]
       change: [-1.0255% -0.7961% -0.5502%] (p = 0.00 < 0.05)

Found 2 outliers among 100 measurements (2.00%)

1 (1.00%) high mild

1 (1.00%) high severe

transfer/pacing-false/varying-seeds/simulated-time/run: No change in performance detected.
       time:   [25.164 s 25.200 s 25.237 s]
       thrpt:  [162.30 KiB/s 162.54 KiB/s 162.77 KiB/s]
change:
       time:   [-0.2619% -0.0528% +0.1555%] (p = 0.62 > 0.05)
       thrpt:  [-0.1552% +0.0528% +0.2625%]

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high mild

transfer/pacing-true/varying-seeds/wallclock-time/run: Change within noise threshold.
       time:   [24.863 ms 24.932 ms 25.010 ms]
       change: [-3.0101% -2.6726% -2.2948%] (p = 0.00 < 0.05)

Found 3 outliers among 100 measurements (3.00%)

2 (2.00%) high mild

1 (1.00%) high severe

transfer/pacing-true/varying-seeds/simulated-time/run: No change in performance detected.
       time:   [24.956 s 24.997 s 25.039 s]
       thrpt:  [163.58 KiB/s 163.86 KiB/s 164.13 KiB/s]
change:
       time:   [-0.2402% -0.0222% +0.2000%] (p = 0.85 > 0.05)
       thrpt:  [-0.1996% +0.0222% +0.2408%]
transfer/pacing-false/same-seed/wallclock-time/run: Change within noise threshold.
       time:   [24.354 ms 24.373 ms 24.394 ms]
       change: [-0.5640% -0.3759% -0.2191%] (p = 0.00 < 0.05)

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high severe

transfer/pacing-false/same-seed/simulated-time/run: No change in performance detected.
       time:   [25.234 s 25.234 s 25.234 s]
       thrpt:  [162.32 KiB/s 162.32 KiB/s 162.32 KiB/s]
change:
       time:   [+0.0000% +0.0000% +0.0000%] (p = NaN > 0.05)
       thrpt:  [+0.0000% +0.0000% +0.0000%]
transfer/pacing-true/same-seed/wallclock-time/run: Change within noise threshold.
       time:   [26.087 ms 26.101 ms 26.117 ms]
       change: [-1.3144% -1.1085% -0.9644%] (p = 0.00 < 0.05)

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high mild

transfer/pacing-true/same-seed/simulated-time/run: No change in performance detected.
       time:   [25.301 s 25.301 s 25.301 s]
       thrpt:  [161.89 KiB/s 161.89 KiB/s 161.89 KiB/s]
change:
       time:   [+0.0000% +0.0000% +0.0000%] (p = NaN > 0.05)
       thrpt:  [+0.0000% +0.0000% +0.0000%]

Download data for profiler.firefox.com or download performance comparison data.

@github-actions
Copy link
Contributor

github-actions bot commented Nov 12, 2025

Client/server transfer results

Performance differences relative to b9c32c7.

Transfer of 33554432 bytes over loopback, min. 100 runs. All unit-less numbers are in milliseconds.

Client vs. server (params) Mean ± σ Min Max MiB/s ± σ Δ main Δ main
google vs. google 455.1 ± 4.0 449.2 466.0 70.3 ± 8.0
google vs. neqo (cubic, paced) 279.0 ± 4.1 273.4 288.3 114.7 ± 7.8 -0.9 -0.3%
msquic vs. msquic 201.2 ± 83.5 142.0 652.7 159.0 ± 0.4
msquic vs. neqo (cubic, paced) 217.0 ± 65.1 147.5 420.9 147.5 ± 0.5 2.3 1.1%
neqo vs. google (cubic, paced) 758.2 ± 6.1 751.6 808.6 42.2 ± 5.2 1.1 0.1%
neqo vs. msquic (cubic, paced) 157.4 ± 4.1 151.1 165.6 203.3 ± 7.8 -0.1 -0.1%
neqo vs. neqo (cubic) 93.5 ± 4.6 84.5 105.3 342.4 ± 7.0 0.5 0.5%
neqo vs. neqo (cubic, paced) 94.1 ± 4.7 84.9 104.7 340.2 ± 6.8 💚 -1.8 -1.9%
neqo vs. neqo (reno) 91.4 ± 4.1 85.0 100.5 350.1 ± 7.8 💚 -2.5 -2.6%
neqo vs. neqo (reno, paced) 92.8 ± 4.5 86.8 107.2 344.9 ± 7.1 💚 -2.0 -2.1%
neqo vs. quiche (cubic, paced) 195.1 ± 4.6 187.6 203.1 164.0 ± 7.0 💔 2.2 1.1%
neqo vs. s2n (cubic, paced) 220.5 ± 3.4 213.5 228.5 145.1 ± 9.4 -0.4 -0.2%
quiche vs. neqo (cubic, paced) 154.4 ± 4.5 145.5 167.5 207.3 ± 7.1 💚 -3.8 -2.4%
quiche vs. quiche 146.5 ± 4.8 139.6 159.4 218.5 ± 6.7
s2n vs. neqo (cubic, paced) 170.7 ± 4.9 161.8 182.7 187.4 ± 6.5 -0.1 -0.1%
s2n vs. s2n 248.2 ± 26.3 230.6 349.6 128.9 ± 1.2

Download data for profiler.firefox.com or download performance comparison data.

@larseggert
Copy link
Collaborator Author

larseggert commented Dec 11, 2025

Closing here; we should revisit this once #2695 becomes actionable.

@larseggert larseggert closed this Dec 11, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@mxinden mxinden mxinden left review comments

Copilot code review Copilot Copilot left review comments

@KershawChang KershawChang Awaiting requested review from KershawChang KershawChang is a code owner

@martinthomson martinthomson Awaiting requested review from martinthomson martinthomson is a code owner

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@larseggert @mxinden