ci: Disable some caching

[larseggert]

The docker and python caches put quite a bit of stress on the GHA cache, causing it - for example - to time out the benchmark results too soon. See https://github.com/mozilla/neqo/actions/caches

(Need to check how much slower this makes the build though, hence draft PR.)

[larseggert]

Closed in favor of #1770

[github-actions]

Benchmark results

Performance differences relative to bb3ab60.

• coalesce_acked_from_zero 1+1 entries
  time: [195.09 ns 195.49 ns 195.94 ns]
  change: [-0.7990% -0.3807% +0.0404%] (p = 0.07 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 3+1 entries
  time: [235.31 ns 235.93 ns 236.62 ns]
  change: [-0.4294% -0.1159% +0.2111%] (p = 0.47 > 0.05)
  No change in performance detected.

• coalesce_acked_from_zero 10+1 entries
  time: [234.06 ns 234.62 ns 235.32 ns]
  change: [-1.0856% -0.6206% -0.1851%] (p = 0.01 < 0.05)
  Change within noise threshold.

• coalesce_acked_from_zero 1000+1 entries
  time: [215.19 ns 215.49 ns 215.88 ns]
  change: [-0.1484% +0.5619% +1.3377%] (p = 0.12 > 0.05)
  No change in performance detected.

• RxStreamOrderer::inbound_frame()
  time: [117.99 ms 118.07 ms 118.16 ms]
  change: [-0.4741% -0.3742% -0.2798%] (p = 0.00 < 0.05)
  Change within noise threshold.

• transfer/Run multiple transfers with varying seeds
  time: [120.90 ms 121.20 ms 121.48 ms]
  thrpt: [32.926 MiB/s 33.004 MiB/s 33.084 MiB/s]
  change:
  time: [+2.0893% +2.4389% +2.7740%] (p = 0.00 < 0.05)
  thrpt: [-2.6991% -2.3809% -2.0465%]
  💔 Performance has regressed.

• transfer/Run multiple transfers with the same seed
  time: [121.94 ms 122.09 ms 122.24 ms]
  thrpt: [32.722 MiB/s 32.763 MiB/s 32.803 MiB/s]
  change:
  time: [+2.5979% +2.7916% +2.9953%] (p = 0.00 < 0.05)
  thrpt: [-2.9082% -2.7158% -2.5321%]
  💔 Performance has regressed.

⬇️ Download logs