Nothing
Blog-post-ready benchmark table templates for shard performance documentation.
### Scaling benchmark: {suite_name}
**Dataset:** {dataset_id}
**Dims:** T={T}, P={P}, V={V}
**Machine:** {cpu}, {ram}GB, {os}, R {r_ver}, BLAS: {blas}
| Backend | Workers | Time (s) ↓ | Speedup ↑ | Peak RSS (GB) ↓ | End RSS (GB) ↓ | RSS drift (GB) ↓ | Hidden copies (MB) ↓ | Recycles |
|---|---:|---:|---:|---:|---:|---:|---:|---:|
| serial | 1 | __ | 1.00× | __ | __ | __ | __ | 0 |
| foreach + doParallel | 2 | __ | __× | __ | __ | __ | __ | 0 |
| foreach + doParallel | 4 | __ | __× | __ | __ | __ | __ | 0 |
| foreach + doParallel | 8 | __ | __× | __ | __ | __ | __ | 0 |
| shard | 2 | __ | __× | __ | __ | __ | __ | __ |
| shard | 4 | __ | __× | __ | __ | __ | __ | __ |
| shard | 8 | __ | __× | __ | __ | __ | __ | __ |
### Memory stability benchmark: {suite_name}
**Workload:** {task_count} tasks, {workers} workers
**Goal:** demonstrate flat RSS + deterministic cleanup
| Backend | Workers | Tasks | Peak RSS (GB) ↓ | End RSS (GB) ↓ | Drift (GB) ↓ | Failures | Recycles |
|---|---:|---:|---:|---:|---:|---:|---:|
| foreach + doParallel | __ | __ | __ | __ | __ | __ | 0 |
| future.apply | __ | __ | __ | __ | __ | __ | 0 |
| shard | __ | __ | __ | __ | __ | __ | __ |
If you want the post to basically write itself from the tables:
# shard: Deterministic Parallelism in R Without Memory Surprises
## The problem (with one painful chart)
- peak RSS grows over time with long-running workers
- hidden copies multiply inputs
## shard's contract (3 bullets)
- shared immutable inputs
- explicit output buffers
- supervised workers with recycling
## Benchmarks
### 1) Many regressions (QR-once)
[Template A table]
- time scaling plot
- peak RSS plot
### 2) 10k-task drift test
[Template B table]
- RSS drift over task index plot
## Diagnostics: proving "no hidden copies"
- materialized_mb
- recycle events
## When to use shard (and when not to)
## Next: doShard + distributed v2
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.