neqo icon indicating copy to clipboard operation
neqo copied to clipboard
verified publisher icon mozilla

chore: Bump quinn-udp to 0.5.5

Open larseggert opened this issue 1 year ago • 6 comments

larseggert avatar Sep 13 '24 13:09 larseggert

I would like to hold back here for now. Larger context posted on Matrix:

Upgrading mozilla-central to windows-sys v0.59.0

phabricator#D220890 upgrades neqo_glue to use quinn-udp v0.5.5.

quinn-udp v0.5.5 uses windows-sys v0.59.0. All other crates in mozilla-central, 13 in total, use windows-sys v0.52.0. See each linked in this SearchFox search.

If I understand the Mozilla policy on duplicate crate versions correctly, one should strive for a single version of each crate in mozilla-central only. This makes sense to me.

Upgrading each of these 13 crates to windows-sys v0.59.0 is a significant undertaking which I assume will take weeks.

Before I start this large effort, posting here to make sure I am not misunderstanding the policy.

https://matrix.to/#/!OYZMPVBAZnIpKdnqHG:mozilla.org/$dytLNfz94wPzYCs__CnIjD3bXsyd4u-FfeZm1lE2gtg?via=mozilla.org&via=matrix.org&via=seirdy.one

The only relevant fix of v0.5.5 is https://github.com/quinn-rs/quinn/issues/1947. Given the low popularity of Android x86, I don't think this is a blocker for Neqo, right?

(Note, given that it is a patch release, I can update quinn-udp in mozilla-central without updating mozilla/neqo first.)

mxinden avatar Sep 13 '24 13:09 mxinden

We can definitely hold back here. Thanks for the background!

larseggert avatar Sep 13 '24 13:09 larseggert

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 95.35%. Comparing base (2ca29c1) to head (bf82a5a). Report is 73 commits behind head on main.

Additional details and impacted files 
@@            Coverage Diff             @@
##             main    #2107      +/-   ##
==========================================
- Coverage   95.35%   95.35%   -0.01%     
==========================================
  Files         112      112              
  Lines       36325    36325              
==========================================
- Hits        34638    34637       -1     
- Misses       1687     1688       +1

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

codecov[bot] avatar Sep 13 '24 13:09 codecov[bot]

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[bot] avatar Sep 13 '24 14:09 github-actions[bot]

Benchmark results

Performance differences relative to 0b106c286f8f9c879f2122daab9e8bd12aa9a703.

coalesce_acked_from_zero 1+1 entries: No change in performance detected.
       time:   [99.262 ns 99.600 ns 99.967 ns]
       change: [-0.4834% +0.1321% +0.6536%] (p = 0.67 > 0.05)

Found 10 outliers among 100 measurements (10.00%)
6 (6.00%) high mild
4 (4.00%) high severe
coalesce_acked_from_zero 3+1 entries: No change in performance detected.
       time:   [117.37 ns 117.77 ns 118.19 ns]
       change: [-0.0167% +0.4792% +1.0002%] (p = 0.06 > 0.05)

Found 18 outliers among 100 measurements (18.00%)
2 (2.00%) low mild
16 (16.00%) high severe
coalesce_acked_from_zero 10+1 entries: No change in performance detected.
       time:   [116.78 ns 117.19 ns 117.70 ns]
       change: [-0.1247% +0.6974% +1.7956%] (p = 0.15 > 0.05)

Found 14 outliers among 100 measurements (14.00%)
3 (3.00%) low severe
1 (1.00%) high mild
10 (10.00%) high severe
coalesce_acked_from_zero 1000+1 entries: No change in performance detected.
       time:   [97.327 ns 97.445 ns 97.582 ns]
       change: [-1.3123% -0.5102% +0.2133%] (p = 0.21 > 0.05)

Found 10 outliers among 100 measurements (10.00%)
4 (4.00%) high mild
6 (6.00%) high severe
RxStreamOrderer::inbound_frame(): Change within noise threshold.
       time:   [111.76 ms 111.81 ms 111.86 ms]
       change: [+0.3179% +0.3878% +0.4561%] (p = 0.00 Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) low mild
transfer/pacing-false/varying-seeds: No change in performance detected.
       time:   [25.794 ms 26.883 ms 27.961 ms]
       change: [-9.0898% -4.1188% +1.1104%] (p = 0.13 > 0.05)
transfer/pacing-true/varying-seeds: No change in performance detected.
       time:   [34.263 ms 36.011 ms 37.816 ms]
       change: [-7.1292% -0.4220% +6.3358%] (p = 0.89 > 0.05)

Found 2 outliers among 100 measurements (2.00%)
2 (2.00%) high mild
transfer/pacing-false/same-seed: No change in performance detected.
       time:   [30.910 ms 31.775 ms 32.603 ms]
       change: [-6.3582% -2.9567% +0.3754%] (p = 0.09 > 0.05)

Found 1 outliers among 100 measurements (1.00%)
1 (1.00%) low mild
transfer/pacing-true/same-seed: No change in performance detected.
       time:   [38.850 ms 41.687 ms 44.546 ms]
       change: [-17.188% -8.3531% +1.0388%] (p = 0.08 > 0.05)
1-conn/1-100mb-resp (aka. Download)/client: Change within noise threshold.
       time:   [113.39 ms 113.83 ms 114.27 ms]
       thrpt:  [875.14 MiB/s 878.49 MiB/s 881.91 MiB/s]
change:
       time:   [-1.5799% -1.0621% -0.5212%] (p = 0.00 +1.0735% +1.6052%]

Found 4 outliers among 100 measurements (4.00%)
1 (1.00%) low severe
1 (1.00%) low mild
2 (2.00%) high mild
1-conn/10_000-parallel-1b-resp (aka. RPS)/client: :broken_heart: Performance has regressed.
       time:   [323.81 ms 326.90 ms 330.00 ms]
       thrpt:  [30.303 Kelem/s 30.590 Kelem/s 30.882 Kelem/s]
change:
       time:   [+1.1660% +2.6251% +4.2072%] (p = 0.00 -2.5580% -1.1526%]
1-conn/1-1b-resp (aka. HPS)/client: Change within noise threshold.
       time:   [33.491 ms 33.656 ms 33.830 ms]
       thrpt:  [29.560  elem/s 29.712  elem/s 29.859  elem/s]
change:
       time:   [-2.2615% -1.4308% -0.6364%] (p = 0.00 +1.4515% +2.3138%]

Found 3 outliers among 100 measurements (3.00%)
2 (2.00%) high mild
1 (1.00%) high severe

Client/server transfer results

Transfer of 33554432 bytes over loopback.

Client Server CC Pacing Mean [ms] Min [ms] Max [ms] Relative
msquic msquic 192.6 ± 94.5 94.0 355.6 1.00
neqo msquic reno on 278.1 ± 89.1 217.5 462.3 1.00
neqo msquic reno 268.9 ± 86.9 214.3 460.4 1.00
neqo msquic cubic on 222.9 ± 24.9 201.1 288.7 1.00
neqo msquic cubic 267.7 ± 93.0 203.6 442.0 1.00
msquic neqo reno on 181.8 ± 127.3 83.2 581.9 1.00
msquic neqo reno 155.0 ± 99.7 83.1 364.2 1.00
msquic neqo cubic on 127.8 ± 91.1 80.9 432.6 1.00
msquic neqo cubic 154.9 ± 100.1 81.5 388.5 1.00
neqo neqo reno on 249.8 ± 124.5 140.3 580.2 1.00
neqo neqo reno 251.6 ± 126.0 133.8 542.3 1.00
neqo neqo cubic on 217.9 ± 95.1 136.8 422.8 1.00
neqo neqo cubic 258.4 ± 126.4 141.9 582.9 1.00

:arrow_down: Download logs

github-actions[bot] avatar Sep 13 '24 15:09 github-actions[bot]

Firefox builds for this PR

The following builds are available for testing. Crossed-out builds did not succeed.

github-actions[bot] avatar Sep 13 '24 15:09 github-actions[bot]

Closing here in favor of https://github.com/mozilla/neqo/pull/2209.

mxinden avatar Oct 30 '24 10:10 mxinden