P281424/AMR
1
0
Fork 0
mirror of https://github.com/msberends/AMR.git synced 2026-05-31 21:41:54 +02:00
Code Releases Activity

Optimise parallel as.sir(): row-batch mode when n_cols < n_cores

Browse Source 
Previously parallel dispatch only parallelised by column, so a 6-column
dataset on a 16-core machine used at most 6 cores with the other 10 idle.
For large n this also caused memory-bandwidth saturation (each worker did
a full n-row scan of clinical_breakpoints simultaneously).

New row-batch mode (fork path, R >= 4.0, non-Windows):
  pieces_per_col = ceil(n_cores / n_cols)
  Jobs = n_cols × pieces_per_col  (≈ n_cores jobs total)
  Each job: one column × one row slice

Benefits:
  - All cores stay busy regardless of column count
  - Per-worker memory footprint shrinks by pieces_per_col ×
  - Breakpoints lookup cache pressure reduced per worker

PSOCK path (Windows / R < 4.0) is unchanged: per-job serialisation
overhead makes row batching unprofitable there.

run_as_sir_column() gains an optional `rows` parameter (NULL = all rows,
backward-compatible). Results are reassembled via as.sir(c(as.character(.)))
which is safe for already-clean SIR values.

https://claude.ai/code/session_012DXCXbZUC54Zij1z9bFiHR
This commit is contained in:
Claude
2026-04-24 22:01:09 +00:00
parent d770469a63
commit 060449e234
3 changed files with 78 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
NEWS.md
View File

@@ -37,6 +37,7 @@
* Fixed BRMO classification by including bacterial complexes (#275)
* Fixed `as.sir()` for data frames silently deleting columns whose AB class was already `<sir>` when called a second time (re-running on already-converted data) (#278)
* Fixed `as.sir()` for data frames incorrectly treating metadata columns (e.g. `patient`, `ward`) as antibiotic columns when their names coincidentally matched an antibiotic code; column content is now validated against AMR data patterns before inclusion
* Improved parallel computing in `as.sir()`: when the number of AB columns is smaller than the number of available cores, rows are now split into batches so all cores stay active (row-batch mode). Previously, a 6-column dataset on a 16-core machine would only use 6 cores; now all 16 are used, with each worker processing a smaller row slice (lower per-worker memory pressure)
### Updates
* Extensive `cli` integration for better message handling and clickable links in messages and warnings (#191, #265)