Skip to content

Use userspace page cache for datalake benchmarks#818

Draft
alexey-milovidov wants to merge 5 commits intomainfrom
use-page-cache-for-datalake
Draft

Use userspace page cache for datalake benchmarks#818
alexey-milovidov wants to merge 5 commits intomainfrom
use-page-cache-for-datalake

Conversation

@alexey-milovidov
Copy link
Member

Summary

  • Switch clickhouse-datalake and clickhouse-datalake-partitioned from filesystem cache (/dev/shm/) to the userspace page cache
  • Replace filesystem_caches config with page_cache_size: auto in clickhouse-local.yaml
  • Replace --filesystem_cache_name cache with --use_page_cache_for_object_storage 1 in query invocations

Test plan

  • Run clickhouse-datalake benchmark and verify hot runs use the page cache
  • Run clickhouse-datalake-partitioned benchmark and verify hot runs use the page cache
  • Compare results against previous filesystem cache numbers

🤖 Generated with Claude Code

…chmarks

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@alexey-milovidov alexey-milovidov marked this pull request as draft March 13, 2026 21:54
alexey-milovidov and others added 4 commits March 13, 2026 23:12
This ensures the userspace page cache persists across tries.
A fresh process per query group means try 1 is naturally cold
(empty page cache) and tries 2-3 are hot, without needing drop_caches.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant