Let me know if there are any specific tests to the variable journal you think should be added.

Both fixed and variable journals are still using a Read wrapper for replay, I wonder if this is still worth it or whether just using the page cache from Append is now good enough.

Both fixed and variable journals are still using a Read wrapper for replay, I wonder if this is still worth it or whether just using the page cache from Append is now good enough.

I think using Read should still outperform (and will avoid clobbering the cache). We should reserve it for random read workloads.

We can port the any::Fixed benchmark to any::Variable; this PR should result in a very nice performance gain with it.

We can port the any::Fixed benchmark to any::Variable; this PR should result in a very nice performance gain with it.

#1388

I think all issues have been addressed now.

this looks good to me. let's get the benchmark merged then I'll approve.

10x faster!

adb::variable_init/elements=100000 operations=1000000
                        time:   [333.55 ms 338.91 ms 345.30 ms]
                        change: [-89.380% -89.187% -88.948%] (p = 0.00 < 0.05)
                        Performance has improved.
adb::variable_init/elements=100000 operations=2000000
                        time:   [367.78 ms 373.06 ms 378.61 ms]
                        change: [-89.794% -89.615% -89.428%] (p = 0.00 < 0.05)
                        Performance has improved.
adb::variable_init/elements=200000 operations=1000000
                        time:   [614.35 ms 622.39 ms 633.07 ms]
                        change: [-89.272% -89.063% -88.841%] (p = 0.00 < 0.05)
                        Performance has improved.
adb::variable_init/elements=200000 operations=2000000
                        time:   [700.03 ms 722.48 ms 751.12 ms]
                        change: [-89.628% -89.303% -88.869%] (p = 0.00 < 0.05)
                        Performance has improved.

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.51%. Comparing base (5ea06c0) to head (b89a71d).
⚠️ Report is 5 commits behind head on main.

@@            Coverage Diff             @@
##             main    #1384      +/-   ##
==========================================
+ Coverage   91.34%   91.51%   +0.17%     
==========================================
  Files         265      265              
  Lines       66726    67134     +408     
==========================================
+ Hits        60950    61441     +491     
+ Misses       5776     5693      -83     

... and 5 files with indirect coverage changes

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 5ea06c0...b89a71d. Read the comment docs.