Horton

Re: BitmapHeapScan streaming read user and prelim refactoring

Tomas Vondra <tomas@vondra.me>

View thread View on postgr.es
From: Tomas Vondra <tomas@vondra.me>
To: Melanie Plageman <melanieplageman@gmail.com>
Cc: Nazir Bilal Yavuz <byavuz81@gmail.com>, Dilip Kumar <dilipbalaut@gmail.com>, Heikki Linnakangas <hlinnaka@iki.fi>, Thomas Munro <thomas.munro@gmail.com>, Andres Freund <andres@anarazel.de>, Pg Hackers <pgsql-hackers@postgresql.org>
Date: 2025-02-14T17:18:47Z
Lists: pgsql-hackers

Attachments

Hi,

Based on off-list discussion with Melanie, I ran a modified version of
the benchmark, with these changes:

- only: uncached runs on the raid-SATA device (4x SSD)
- only: work_mem=4MB
- only: pached-0004 branch
- all datasets, ndistinct/nmatches combinations, 1M and 10M rows, ...

and then two new parameters:

- io_combine_limit 1, 16, 32
- read-ahead, 16, 256, 4096

The complete logs are pushed to github, as usual:

  https://github.com/tvondra/bitmapscan-tests/tree/main/20250212-221123

There's also a .sh with the benchmarking script.

I'm attaching a PDF with the comparison, the columns are for

  - io_combine_limit (16, 256, 4096)
  - effective_io_concurrency (0, 1, 16)

And as usual, the colored columns are comparing patched-0004
(green=good, red=bad).

One way to look at the PDF is that there are 9 squares (in 3x3 layout),
for combinations of io_combine_limit and read-ahead values. The central
square is for the default values (ioc=16, ra=256), and you can compare
this to the surrounding squares.

Of course, this shows "behavior relative to master", a square being
"more green" does not mean the absolute performance improves, just that
it's "more better" relative to master.

I hope this explanation makes sense ...


FWIW this does not change anything in the detection of sequential access
patterns, discussed nearby, because the benchmarks started before Andres
looked into that. If needed, I can easily rerun these tests, I just need
a patch to apply.

But if there really is some sort of issue, it'd make sense why it's much
worse on the older SATA SSDs, while NVMe devices perform somewhat
better. Because AFAICS the NVMe devices are better at handling random
I/O with shorter queues.


regards

-- 
Tomas Vondra

Commits

  1. Fix bitmapheapscan incorrect recheck of NULL tuples

  2. Increase default maintenance_io_concurrency to 16

  3. Separate TBM[Shared|Private]Iterator and TBMIterateResult

  4. Improve read_stream.c advice for dense streams.

  5. Increase default effective_io_concurrency to 16

  6. Delay extraction of TIDBitmap per page offsets

  7. Add lossy indicator to TBMIterateResult

  8. Move BitmapTableScan per-scan setup into a helper

  9. Add and use BitmapHeapScanDescData struct

  10. Fix bitmap table scan crash on iterator release

  11. Bitmap Table Scans use unified TBMIterator

  12. Add common interface for TBMIterators

  13. Make table_scan_bitmap_next_block() async-friendly

  14. Move EXPLAIN counter increment to heapam_scan_bitmap_next_block

  15. Refactor tidstore.c iterator buffering.

  16. BitmapHeapScan: Remove incorrect assert and reset field

  17. Change BitmapAdjustPrefetchIterator to accept BlockNumber

  18. BitmapHeapScan: Use correct recheck flag for skip_fetch

  19. BitmapHeapScan: Push skip_fetch optimization into table AM

  20. BitmapHeapScan: postpone setting can_skip_fetch

  21. BitmapHeapScan: begin scan after bitmap creation

  22. Fix EXPLAIN Bitmap heap scan to count pages with no visible tuples

  23. Remove redundant snapshot copying from parallel leader to workers

  24. Remove some obsolete smgrcloseall() calls.

  25. Remove the "snapshot too old" feature.

  26. Compute XID horizon for page level index vacuum on primary.