JakeDern commented on PR #10128:
URL: https://github.com/apache/arrow-rs/pull/10128#issuecomment-5398730983

   I spun up a little arm64 machine on gcp and it seems to be an arm64 specific 
regression. The good news is it's at least reproducible, I will continue to 
poke at this:
   
   Before:
   
   ```
   jakedern@jakedern-arm64-test:~/repos/arrow-rs$ cargo bench -p arrow-ipc 
StreamWriter/write_10 --features zstd
       Finished `bench` profile [optimized] target(s) in 0.10s
        Running benches/ipc_reader.rs 
(target/release/deps/ipc_reader-cb539e9f107fc994)
        Running benches/ipc_writer.rs 
(target/release/deps/ipc_writer-d05199adc498efdb)
   Benchmarking arrow_ipc_stream_writer/StreamWriter/write_10: Collecting 100 
samples in estimated 5.2416 s (45k 
iteratarrow_ipc_stream_writer/StreamWriter/write_10
                           time:   [115.50 µs 116.01 µs 116.59 µs]
                           change: [−1.0167% −0.7073% −0.3216%] (p = 0.00 < 
0.05)
                           Change within noise threshold.
   Found 7 outliers among 100 measurements (7.00%)
     1 (1.00%) low severe
     3 (3.00%) high mild
     3 (3.00%) high severe
   Benchmarking arrow_ipc_stream_writer/StreamWriter/write_10/zstd: Collecting 
100 samples in estimated 5.6701 s (800 
iarrow_ipc_stream_writer/StreamWriter/write_10/zstd
                           time:   [7.0890 ms 7.0949 ms 7.1014 ms]
                           change: [−2.5044% −2.4145% −2.3232%] (p = 0.00 < 
0.05)
                           Performance has improved.
   Found 8 outliers among 100 measurements (8.00%)
     6 (6.00%) high mild
     2 (2.00%) high severe
   Benchmarking arrow_ipc_stream_writer/StreamWriter/write_10/dict: Collecting 
100 samples in estimated 5.2918 s (45k 
iarrow_ipc_stream_writer/StreamWriter/write_10/dict
                           time:   [103.47 µs 104.85 µs 106.60 µs]
                           change: [−1.5856% +0.5960% +2.7682%] (p = 0.60 > 
0.05)
                           No change in performance detected.
   Found 19 outliers among 100 measurements (19.00%)
     4 (4.00%) low severe
     15 (15.00%) high severe
   Benchmarking arrow_ipc_stream_writer/StreamWriter/write_10/dict/delta: 
Collecting 100 samples in estimated 5.4591 s 
arrow_ipc_stream_writer/StreamWriter/write_10/dict/delta
                           time:   [221.04 µs 223.23 µs 225.30 µs]
                           change: [−0.0930% +0.9083% +1.9294%] (p = 0.08 > 
0.05)
                           No change in performance detected. 
   ```
   
   After:
   ```
   jakedern@jakedern-arm64-test:~/repos/arrow-rs$ cargo bench -p arrow-ipc 
StreamWriter/write_10 --features zstd
       Updating crates.io index
      Compiling arrow-buffer v59.2.0 
(/home/jakedern/repos/arrow-rs/arrow-buffer)
      Compiling arrow-schema v59.2.0 
(/home/jakedern/repos/arrow-rs/arrow-schema)
      Compiling arrow-data v59.2.0 (/home/jakedern/repos/arrow-rs/arrow-data)
      Compiling arrow-array v59.2.0 (/home/jakedern/repos/arrow-rs/arrow-array)
      Compiling arrow-cmp v59.2.0 (/home/jakedern/repos/arrow-rs/arrow-cmp)
      Compiling arrow-select v59.2.0 
(/home/jakedern/repos/arrow-rs/arrow-select)
      Compiling arrow-ipc v59.2.0 (/home/jakedern/repos/arrow-rs/arrow-ipc)
       Finished `bench` profile [optimized] target(s) in 42.35s
        Running benches/ipc_reader.rs 
(target/release/deps/ipc_reader-cb539e9f107fc994)
        Running benches/ipc_writer.rs 
(target/release/deps/ipc_writer-d05199adc498efdb)
   Benchmarking arrow_ipc_stream_writer/StreamWriter/write_10: Collecting 100 
samples in estimated 5.1168 s (40k 
iteratarrow_ipc_stream_writer/StreamWriter/write_10
                           time:   [123.74 µs 124.03 µs 124.33 µs]
                           change: [+6.8925% +7.3561% +7.7905%] (p = 0.00 < 
0.05)
                           Performance has regressed.
   Found 1 outliers among 100 measurements (1.00%)
     1 (1.00%) low severe
   Benchmarking arrow_ipc_stream_writer/StreamWriter/write_10/zstd: Collecting 
100 samples in estimated 5.5733 s (800 
iarrow_ipc_stream_writer/StreamWriter/write_10/zstd
                           time:   [6.9616 ms 6.9628 ms 6.9641 ms]
                           change: [−1.9536% −1.8617% −1.7776%] (p = 0.00 < 
0.05)
                           Performance has improved.
   Found 1 outliers among 100 measurements (1.00%)
     1 (1.00%) high mild
   Benchmarking arrow_ipc_stream_writer/StreamWriter/write_10/dict: Collecting 
100 samples in estimated 5.1232 s (81k 
iarrow_ipc_stream_writer/StreamWriter/write_10/dict
                           time:   [63.543 µs 63.748 µs 63.967 µs]
                           change: [−40.448% −39.494% −38.577%] (p = 0.00 < 
0.05)
                           Performance has improved.
   Found 1 outliers among 100 measurements (1.00%)
     1 (1.00%) high mild
   Benchmarking arrow_ipc_stream_writer/StreamWriter/write_10/dict/delta: 
Collecting 100 samples in estimated 5.9280 s 
arrow_ipc_stream_writer/StreamWriter/write_10/dict/delta
                           time:   [194.84 µs 196.05 µs 197.27 µs]
                           change: [−12.844% −11.875% −10.895%] (p = 0.00 < 
0.05)
                           Performance has improved.
   Found 2 outliers among 100 measurements (2.00%)
     1 (1.00%) high mild
     1 (1.00%) high severe
   ```


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to