Skip to content

feat(transport/cubic): multiplicative decrease updates#2970

Merged
mxinden merged 5 commits intomozilla:mainfrom
omansfeld:cubic_decrease_updates
Sep 23, 2025
Merged

feat(transport/cubic): multiplicative decrease updates#2970
mxinden merged 5 commits intomozilla:mainfrom
omansfeld:cubic_decrease_updates

Conversation

@omansfeld
Copy link
Collaborator

@omansfeld omansfeld commented Sep 10, 2025 •
edited
Loading

Changes the following:

  • new fast convergence algorithm works without an extra variable for saving state (see RFC)
    • thus we get rid of last_max_cwnd
  • add cwnd_prior for logging, this will also be used in a later PR for dynamically changing the value of CUBIC_ALPHA
  • adding comments
  • renaming to match RFC 9438

Split off of #2535.
Part of #2967.
Depends on #2968.

@omansfeld omansfeld mentioned this pull request Sep 10, 2025
Closed
6 tasks
@codecov
Copy link

codecov bot commented Sep 10, 2025 •
edited
Loading

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.36%. Comparing base (8b7d440) to head (1720e3e).
⚠️ Report is 5 commits behind head on main.

Additional details and impacted files 
@@            Coverage Diff             @@
##             main    #2970      +/-   ##
==========================================
- Coverage   95.67%   93.36%   -2.31%     
==========================================
  Files         123      123              
  Lines       35706    35703       -3     
  Branches    35706    35703       -3     
==========================================
- Hits        34160    33333     -827     
- Misses       1506     1528      +22     
- Partials       40      842     +802
Components Coverage Δ
neqo-common 97.30% <ø> (-0.88%) ⬇️
neqo-crypto 83.31% <ø> (-7.17%) ⬇️
neqo-http3 93.32% <ø> (-1.81%) ⬇️
neqo-qpack 94.14% <ø> (-2.09%) ⬇️
neqo-transport 94.43% <100.00%> (-2.14%) ⬇️
neqo-udp 80.48% <ø> (-10.74%) ⬇️
mtu 85.76% <ø> (-1.74%) ⬇️

@mxinden
Copy link
Member

mxinden commented Sep 15, 2025

Keeping as a draft until #2968 is merged.

With #2968 merged, is this ready for review?

@omansfeld omansfeld marked this pull request as ready for review September 15, 2025 12:44
@omansfeld
Copy link
Collaborator Author

omansfeld commented Sep 15, 2025

Keeping as a draft until #2968 is merged.

With #2968 merged, is this ready for review?

It is, pending Lars's comment about general CI coverage

Before we start reviewing/merging PRs that change behavior, I'd like to discuss better (CI) testing for CC, and getting that merged first.

#2967 (comment)

@omansfeld
Copy link
Collaborator Author

omansfeld commented Sep 16, 2025

@mxinden This is just a refactor and the behavior is already tested here:

neqo/neqo-transport/src/cc/tests/cubic.rs

Lines 310 to 335 in 89cb6ea

fn congestion_event_congestion_avoidance_2() {
let mut cubic = ClassicCongestionControl::new(Cubic::default(), Pmtud::new(IP_ADDR, MTU));
// Set ssthresh to something small to make sure that cc is in the congection avoidance phase.
cubic.set_ssthresh(1);
// Set last_max_cwnd to something higher than cwnd so that the fast convergence is triggered.
let cwnd_initial_f64 = convert_to_f64(cubic.cwnd_initial());
cubic.set_last_max_cwnd(cwnd_initial_f64 * 10.0);
_ = fill_cwnd(&mut cubic, 0, now());
ack_packet(&mut cubic, 0, now());
assert_within(cubic.last_max_cwnd(), cwnd_initial_f64 * 10.0, f64::EPSILON);
assert_eq!(cubic.cwnd(), cubic.cwnd_initial());
// Trigger a congestion_event.
packet_lost(&mut cubic, 1);
assert_within(
cubic.last_max_cwnd(),
cwnd_initial_f64 * CUBIC_FAST_CONVERGENCE,
f64::EPSILON,
);
assert_eq!(cubic.cwnd(), cwnd_after_loss(cubic.cwnd_initial()));
}

So this PR should be ready for review/merge right now.

The refactor is mainly that fast convergence has a slightly different algorithm now that uses one less variable to keep state. Functionally it does the same thing.

For reference:
Old RFC: https://datatracker.ietf.org/doc/html/rfc8312#section-4.6
New RFC: https://datatracker.ietf.org/doc/html/rfc9438#name-fast-convergence

@mxinden mxinden changed the title cubic: multiplicative decrease updates refactor(transport/cubic): multiplicative decrease updates Sep 17, 2025
mxinden
mxinden reviewed Sep 17, 2025
View reviewed changes
@omansfeld omansfeld mentioned this pull request Sep 17, 2025
Merged
@github-actions
Copy link
Contributor

github-actions bot commented Sep 17, 2025 •
edited
Loading

🐰 Bencher Report

Branchcubic_decrease_updates
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		207,290,000.00 ns
(-1.83%)Baseline: 211,157,640.45 ns
218,765,548.43 ns
(94.75%)
1-conn/1-100mb-resp/mtu-1504 (aka. Download)/client📈 view plot
🚷 view threshold		199,440,000.00 ns
(-2.98%)Baseline: 205,556,629.21 ns
214,279,795.21 ns
(93.07%)
1-conn/1-1b-resp/mtu-1504 (aka. HPS)/client📈 view plot
🚷 view threshold		28,491,000.00 ns
(+0.65%)Baseline: 28,305,719.10 ns
28,818,517.10 ns
(98.86%)
1-conn/10_000-parallel-1b-resp/mtu-1504 (aka. RPS)/client📈 view plot
🚷 view threshold		295,140,000.00 ns
(+0.37%)Baseline: 294,063,483.15 ns
301,012,816.27 ns
(98.05%)
1-streams/each-1000-bytes/simulated-time📈 view plot
🚷 view threshold		119,230,000.00 ns
(+1.40%)Baseline: 117,587,191.01 ns
120,307,629.74 ns
(99.10%)
1-streams/each-1000-bytes/wallclock-time📈 view plot
🚷 view threshold		593,520.00 ns
(-1.93%)Baseline: 605,230.22 ns
629,624.70 ns
(94.27%)
1000-streams/each-1-bytes/simulated-time📈 view plot
🚷 view threshold		14,997,000,000.00 ns
(+0.05%)Baseline: 14,989,528,089.89 ns
15,008,117,892.06 ns
(99.93%)
1000-streams/each-1-bytes/wallclock-time📈 view plot
🚷 view threshold		14,098,000.00 ns
(-2.45%)Baseline: 14,451,617.98 ns
15,057,961.61 ns
(93.62%)
1000-streams/each-1000-bytes/simulated-time📈 view plot
🚷 view threshold		18,910,000,000.00 ns
(+0.09%)Baseline: 18,893,078,651.69 ns
19,109,024,465.38 ns
(98.96%)
1000-streams/each-1000-bytes/wallclock-time📈 view plot
🚷 view threshold		50,599,000.00 ns
(-6.68%)Baseline: 54,219,505.62 ns
60,428,590.95 ns
(83.73%)
RxStreamOrderer::inbound_frame()📈 view plot
🚷 view threshold		109,530,000.00 ns
(-0.07%)Baseline: 109,604,044.94 ns
111,581,896.08 ns
(98.16%)
coalesce_acked_from_zero 1+1 entries📈 view plot
🚷 view threshold		88.92 ns
(+0.38%)Baseline: 88.58 ns
89.23 ns
(99.66%)
coalesce_acked_from_zero 10+1 entries📈 view plot
🚷 view threshold		106.60 ns
(+0.45%)Baseline: 106.12 ns
107.01 ns
(99.62%)
coalesce_acked_from_zero 1000+1 entries📈 view plot
🚷 view threshold		89.55 ns
(-0.33%)Baseline: 89.84 ns
93.74 ns
(95.53%)
coalesce_acked_from_zero 3+1 entries📈 view plot
🚷 view threshold		107.47 ns
(+0.78%)Baseline: 106.64 ns
107.59 ns
(99.89%)
decode 1048576 bytes, mask 3f📈 view plot
🚷 view threshold		1,589,200.00 ns
(-0.21%)Baseline: 1,592,615.73 ns
1,600,018.62 ns
(99.32%)
decode 1048576 bytes, mask 7f📈 view plot
🚷 view threshold		5,043,600.00 ns
(-0.33%)Baseline: 5,060,098.88 ns
5,078,765.16 ns
(99.31%)
decode 1048576 bytes, mask ff📈 view plot
🚷 view threshold		3,029,100.00 ns
(-0.13%)Baseline: 3,033,119.10 ns
3,045,489.35 ns
(99.46%)
decode 4096 bytes, mask 3f📈 view plot
🚷 view threshold		8,276.30 ns
(-0.21%)Baseline: 8,293.92 ns
8,334.79 ns
(99.30%)
decode 4096 bytes, mask 7f📈 view plot
🚷 view threshold		19,997.00 ns
(-0.04%)Baseline: 20,005.52 ns
20,090.21 ns
(99.54%)
decode 4096 bytes, mask ff📈 view plot
🚷 view threshold		11,662.00 ns
(-1.31%)Baseline: 11,817.29 ns
12,009.37 ns
(97.11%)
sent::Packets::take_ranges📈 view plot
🚷 view threshold		4,569.60 ns
(-4.76%)Baseline: 4,797.89 ns
5,032.89 ns
(90.79%)
transfer/pacing-false/same-seed/simulated-time/run📈 view plot
🚷 view threshold		25,152,000,000.00 ns
transfer/pacing-false/same-seed/wallclock-time/run📈 view plot
🚷 view threshold		26,525,000.00 ns
(+1.91%)Baseline: 26,026,758.62 ns
26,823,767.14 ns
(98.89%)
transfer/pacing-false/varying-seeds/simulated-time/run📈 view plot
🚷 view threshold		25,157,000,000.00 ns
(-0.03%)Baseline: 25,164,091,954.02 ns
25,209,371,794.20 ns
(99.79%)
transfer/pacing-false/varying-seeds/wallclock-time/run📈 view plot
🚷 view threshold		26,935,000.00 ns
(+2.08%)Baseline: 26,387,103.45 ns
27,149,800.23 ns
(99.21%)
transfer/pacing-true/same-seed/simulated-time/run📈 view plot
🚷 view threshold		25,588,000,000.00 ns
transfer/pacing-true/same-seed/wallclock-time/run📈 view plot
🚷 view threshold		27,876,000.00 ns
(+1.21%)Baseline: 27,543,632.18 ns
28,353,524.57 ns
(98.32%)
transfer/pacing-true/varying-seeds/simulated-time/run📈 view plot
🚷 view threshold		24,991,000,000.00 ns
(-0.01%)Baseline: 24,993,954,022.99 ns
25,043,989,857.86 ns
(99.79%)
transfer/pacing-true/varying-seeds/wallclock-time/run📈 view plot
🚷 view threshold		27,542,000.00 ns
(+2.22%)Baseline: 26,944,229.89 ns
27,699,879.03 ns
(99.43%)
🐰 View full continuous benchmarking report in Bencher

@omansfeld
cubic: multiplicative decrease updates
9cda419 
- fast convergence algorithm
- renaming
- updating docs
- adding cwnd_prior
- fixing tests
@omansfeld
- rename fast convergence test
b5dde4f
@omansfeld
addressing review comments (doc wording)
66aa0af
@omansfeld omansfeld force-pushed the cubic_decrease_updates branch from 3ea26cb to 66aa0af Compare September 17, 2025 15:09
@omansfeld omansfeld changed the title refactor(transport/cubic): multiplicative decrease updates feature(transport/cubic): multiplicative decrease updates Sep 22, 2025
@github-actions
Copy link
Contributor

github-actions bot commented Sep 22, 2025

Failed Interop Tests

QUIC Interop Runner, client vs. server

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 Sep 22, 2025

Client/server transfer results

Performance differences relative to 8b7d440.

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 457.6 ± 4.7 449.5 468.2 69.9 ± 6.8
google vs. neqo (cubic, paced) 276.6 ± 4.2 270.6 286.6 115.7 ± 7.6 💚 -1.3 -0.5%
msquic vs. msquic 162.7 ± 43.2 135.9 428.5 196.7 ± 0.7
msquic vs. neqo (cubic, paced) 181.6 ± 24.1 150.3 278.0 176.2 ± 1.3 -7.7 -4.1%
neqo vs. google (cubic, paced) 760.0 ± 4.0 753.8 773.6 42.1 ± 8.0 -0.6 -0.1%
neqo vs. msquic (cubic, paced) 155.8 ± 4.8 149.0 182.9 205.5 ± 6.7 💚 -2.4 -1.5%
neqo vs. neqo (cubic) 90.0 ± 4.9 78.5 108.1 355.4 ± 6.5 0.4 0.5%
neqo vs. neqo (cubic, paced) 90.6 ± 4.8 80.9 105.8 353.3 ± 6.7 💚 -1.4 -1.5%
neqo vs. neqo (reno) 90.6 ± 4.9 83.7 107.4 353.4 ± 6.5 0.7 0.7%
neqo vs. neqo (reno, paced) 91.6 ± 4.6 84.5 100.0 349.4 ± 7.0 -0.8 -0.9%
neqo vs. quiche (cubic, paced) 193.0 ± 3.9 186.3 204.1 165.8 ± 8.2 💚 -1.6 -0.8%
neqo vs. s2n (cubic, paced) 218.8 ± 3.9 212.6 226.4 146.2 ± 8.2 💚 -2.2 -1.0%
quiche vs. neqo (cubic, paced) 152.5 ± 4.4 141.6 163.5 209.9 ± 7.3 1.2 0.8%
quiche vs. quiche 145.3 ± 5.6 136.7 167.9 220.2 ± 5.7
s2n vs. neqo (cubic, paced) 173.4 ± 4.8 161.2 187.6 184.5 ± 6.7 0.4 0.2%
s2n vs. s2n 255.8 ± 30.5 234.6 349.5 125.1 ± 1.0

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

@github-actions
Copy link
Contributor

github-actions bot commented Sep 22, 2025

Benchmark results

Performance differences relative to 8b7d440.

1-conn/1-100mb-resp/mtu-1504 (aka. Download)/client: Change within noise threshold.
       time:   [199.02 ms 199.44 ms 199.95 ms]
       thrpt:  [500.12 MiB/s 501.41 MiB/s 502.46 MiB/s]
change:
       time:   [−1.2092% −0.9454% −0.6403%] (p = 0.00 < 0.05)
       thrpt:  [+0.6444% +0.9545% +1.2240%]

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high severe

1-conn/10_000-parallel-1b-resp/mtu-1504 (aka. RPS)/client: No change in performance detected.
       time:   [293.26 ms 295.14 ms 297.07 ms]
       thrpt:  [33.662 Kelem/s 33.883 Kelem/s 34.100 Kelem/s]
change:
       time:   [−0.1720% +0.6582% +1.4494%] (p = 0.12 > 0.05)
       thrpt:  [−1.4287% −0.6539% +0.1723%]

Found 2 outliers among 100 measurements (2.00%)

1 (1.00%) low mild

1 (1.00%) high mild

1-conn/1-1b-resp/mtu-1504 (aka. HPS)/client: No change in performance detected.
       time:   [28.415 ms 28.491 ms 28.576 ms]
       thrpt:  [34.995   B/s 35.099   B/s 35.192   B/s]
change:
       time:   [−0.4699% −0.0447% +0.3760%] (p = 0.85 > 0.05)
       thrpt:  [−0.3746% +0.0447% +0.4721%]

Found 20 outliers among 100 measurements (20.00%)

4 (4.00%) low severe

1 (1.00%) high mild

15 (15.00%) high severe

1-conn/1-100mb-req/mtu-1504 (aka. Upload)/client: Change within noise threshold.
       time:   [207.01 ms 207.29 ms 207.60 ms]
       thrpt:  [481.70 MiB/s 482.42 MiB/s 483.07 MiB/s]
change:
       time:   [−0.6017% −0.4128% −0.2161%] (p = 0.00 < 0.05)
       thrpt:  [+0.2166% +0.4145% +0.6053%]

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:   [11.622 µs 11.662 µs 11.708 µs]
       change: [−0.4545% +0.1013% +0.6223%] (p = 0.72 > 0.05)

Found 17 outliers among 100 measurements (17.00%)

1 (1.00%) low severe

5 (5.00%) low mild

1 (1.00%) high mild

10 (10.00%) high severe

decode 1048576 bytes, mask ff: No change in performance detected.
       time:   [3.0196 ms 3.0291 ms 3.0403 ms]
       change: [−0.2729% +0.1636% +0.6066%] (p = 0.49 > 0.05)

Found 9 outliers among 100 measurements (9.00%)

9 (9.00%) high severe

decode 4096 bytes, mask 7f: No change in performance detected.
       time:   [19.941 µs 19.997 µs 20.065 µs]
       change: [−0.4638% −0.0131% +0.4022%] (p = 0.95 > 0.05)

Found 17 outliers among 100 measurements (17.00%)

1 (1.00%) low severe

1 (1.00%) low mild

3 (3.00%) high mild

12 (12.00%) high severe

decode 1048576 bytes, mask 7f: No change in performance detected.
       time:   [5.0327 ms 5.0436 ms 5.0551 ms]
       change: [−0.4346% −0.0755% +0.2538%] (p = 0.67 > 0.05)

Found 11 outliers among 100 measurements (11.00%)

11 (11.00%) high severe

decode 4096 bytes, mask 3f: No change in performance detected.
       time:   [8.2524 µs 8.2763 µs 8.3073 µs]
       change: [−1.2731% −0.4022% +0.2652%] (p = 0.35 > 0.05)

Found 19 outliers among 100 measurements (19.00%)

8 (8.00%) low mild

1 (1.00%) high mild

10 (10.00%) high severe

decode 1048576 bytes, mask 3f: No change in performance detected.
       time:   [1.5849 ms 1.5892 ms 1.5949 ms]
       change: [−0.6173% −0.0952% +0.4157%] (p = 0.71 > 0.05)

Found 8 outliers among 100 measurements (8.00%)

2 (2.00%) high mild

6 (6.00%) high severe

1-streams/each-1000-bytes/wallclock-time: No change in performance detected.
       time:   [589.88 µs 593.52 µs 597.62 µs]
       change: [−1.0576% +0.0602% +1.0451%] (p = 0.92 > 0.05)

Found 19 outliers among 100 measurements (19.00%)

19 (19.00%) high severe

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

time:   [119.01 ms 119.23 ms 119.44 ms]

thrpt:  [8.1763 KiB/s 8.1908 KiB/s 8.2054 KiB/s]

change:

time:   [−0.1549% +0.0958% +0.3603%] (p = 0.48 > 0.05)

thrpt:  [−0.3590% −0.0957% +0.1552%]

No change in performance detected.

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high mild

1000-streams/each-1-bytes/wallclock-time: No change in performance detected.
       time:   [14.042 ms 14.098 ms 14.182 ms]
       change: [−0.3112% +0.1620% +0.8293%] (p = 0.62 > 0.05)

Found 1 outliers among 100 measurements (1.00%)

1 (1.00%) high severe

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

time:   [14.984 s 14.997 s 15.010 s]

thrpt:  [66.622   B/s 66.679   B/s 66.737   B/s]

change:

time:   [−0.1512% −0.0208% +0.1040%] (p = 0.74 > 0.05)

thrpt:  [−0.1039% +0.0208% +0.1515%]

No change in performance detected.

1000-streams/each-1000-bytes/wallclock-time: Change within noise threshold.
       time:   [50.409 ms 50.599 ms 50.795 ms]
       change: [−1.7733% −1.2668% −0.7724%] (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:   [18.736 s 18.910 s 19.086 s]

thrpt:  [51.168 KiB/s 51.642 KiB/s 52.123 KiB/s]

change:

time:   [−1.4732% −0.1734% +1.0811%] (p = 0.80 > 0.05)

thrpt:  [−1.0695% +0.1737% +1.4952%]

No change in performance detected.

coalesce_acked_from_zero 1+1 entries: No change in performance detected.
       time:   [88.583 ns 88.924 ns 89.263 ns]
       change: [−0.7653% −0.1736% +0.3470%] (p = 0.55 > 0.05)

Found 10 outliers among 100 measurements (10.00%)

9 (9.00%) high mild

1 (1.00%) high severe

coalesce_acked_from_zero 3+1 entries: No change in performance detected.
       time:   [107.08 ns 107.47 ns 107.88 ns]
       change: [−0.5460% +0.0183% +0.5348%] (p = 0.96 > 0.05)

Found 17 outliers among 100 measurements (17.00%)

8 (8.00%) high mild

9 (9.00%) high severe

coalesce_acked_from_zero 10+1 entries: No change in performance detected.
       time:   [106.20 ns 106.60 ns 107.09 ns]
       change: [−0.2235% +0.4779% +1.5881%] (p = 0.34 > 0.05)

Found 15 outliers among 100 measurements (15.00%)

3 (3.00%) low severe

2 (2.00%) low mild

2 (2.00%) high mild

8 (8.00%) high severe

coalesce_acked_from_zero 1000+1 entries: No change in performance detected.
       time:   [89.428 ns 89.551 ns 89.691 ns]
       change: [−0.9254% +0.1754% +1.2288%] (p = 0.76 > 0.05)

Found 12 outliers among 100 measurements (12.00%)

2 (2.00%) high mild

10 (10.00%) high severe

RxStreamOrderer::inbound_frame(): Change within noise threshold.
       time:   [109.37 ms 109.53 ms 109.79 ms]
       change: [−0.7895% −0.6148% −0.3293%] (p = 0.00 < 0.05)

Found 9 outliers among 100 measurements (9.00%)

7 (7.00%) low mild

1 (1.00%) high mild

1 (1.00%) high severe

sent::Packets::take_ranges: No change in performance detected.
       time:   [4.4720 µs 4.5696 µs 4.6742 µs]
       change: [−4.7448% −1.4587% +1.6001%] (p = 0.39 > 0.05)

Found 3 outliers among 100 measurements (3.00%)

3 (3.00%) high mild

transfer/pacing-false/varying-seeds/wallclock-time/run: Change within noise threshold.
       time:   [26.885 ms 26.935 ms 26.986 ms]
       change: [+0.8013% +1.0582% +1.3165%] (p = 0.00 < 0.05)

Found 2 outliers among 100 measurements (2.00%)

1 (1.00%) low mild

1 (1.00%) high mild

transfer/pacing-false/varying-seeds/simulated-time/run: No change in performance detected.
       time:   [25.119 s 25.157 s 25.194 s]
       thrpt:  [162.58 KiB/s 162.82 KiB/s 163.06 KiB/s]
change:
       time:   [−0.1834% +0.0272% +0.2308%] (p = 0.81 > 0.05)
       thrpt:  [−0.2302% −0.0272% +0.1837%]
transfer/pacing-true/varying-seeds/wallclock-time/run: Change within noise threshold.
       time:   [27.463 ms 27.542 ms 27.624 ms]
       change: [+0.2344% +0.6146% +0.9858%] (p = 0.00 < 0.05)
transfer/pacing-true/varying-seeds/simulated-time/run: No change in performance detected.
       time:   [24.949 s 24.991 s 25.033 s]
       thrpt:  [163.62 KiB/s 163.90 KiB/s 164.17 KiB/s]
change:
       time:   [−0.0959% +0.1462% +0.3802%] (p = 0.23 > 0.05)
       thrpt:  [−0.3787% −0.1459% +0.0960%]

Found 7 outliers among 100 measurements (7.00%)

4 (4.00%) low mild

3 (3.00%) high mild

transfer/pacing-false/same-seed/wallclock-time/run: Change within noise threshold.
       time:   [26.490 ms 26.525 ms 26.576 ms]
       change: [+1.6328% +1.8750% +2.1157%] (p = 0.00 < 0.05)

Found 3 outliers among 100 measurements (3.00%)

1 (1.00%) high mild

2 (2.00%) high severe

transfer/pacing-false/same-seed/simulated-time/run: No change in performance detected.
       time:   [25.152 s 25.152 s 25.152 s]
       thrpt:  [162.85 KiB/s 162.85 KiB/s 162.85 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: No change in performance detected.
       time:   [27.849 ms 27.876 ms 27.904 ms]
       change: [−0.1251% +0.0015% +0.1247%] (p = 0.98 > 0.05)

Found 4 outliers among 100 measurements (4.00%)

4 (4.00%) high mild

transfer/pacing-true/same-seed/simulated-time/run: No change in performance detected.
       time:   [25.588 s 25.588 s 25.588 s]
       thrpt:  [160.07 KiB/s 160.07 KiB/s 160.07 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.

@mxinden mxinden changed the title feature(transport/cubic): multiplicative decrease updates feat(transport/cubic): multiplicative decrease updates Sep 23, 2025
mxinden
mxinden approved these changes Sep 23, 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.

Thanks for answering my many questions.

@mxinden mxinden added this pull request to the merge queue Sep 23, 2025
Merged via the queue into mozilla:main with commit 836046b Sep 23, 2025
93 of 94 checks passed
@mxinden mxinden mentioned this pull request Oct 14, 2025
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@larseggert larseggert larseggert approved these changes

@mxinden mxinden mxinden approved these changes

@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.

3 participants

@omansfeld @mxinden @larseggert