- Link paper and update task category to other (c6fb61e342b787b0f2a63d11404953a68a461661)
- Update README.md (2291fdfc581485159069c56d380f5980a2ef6965)
Co-authored-by: Niels Rogge <nielsr@users.noreply.huggingface.co>
Update README with Git LFS fetch instructions
rajveerb+1: Claude Opus 4.6 (1M context)
2b537bc
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Disable HuggingFace dataset viewer
rajveerb+1: Claude Opus 4.6
6c5e149
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Update paper title and add submodule context with branch info
rajveerb+1: Claude Opus 4.6
ef79e56
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add HuggingFace dataset card metadata and fix LCAS description
rajveerb+1: Claude Opus 4.6
38d2e4a
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add dataset README with structure and schema documentation
rajveerb+1: Claude Opus 4.6
882c35d
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Remove run logs for QPS values outside paper's evaluated range
rajveerb+1: Claude Opus 4.6
d01aad2
Crawler: removed 16 run log dirs for arrival=0.0625 (16 QPS) and
arrival=0.125 (8 QPS). Paper evaluates QPS 0.5–4.
ANNS: removed 24 run log dirs for arrival=0.0625 (16 QPS),
arrival=0.125 (8 QPS), and arrival=0.25 (4 QPS). Paper evaluates
QPS 0.25–2.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Simplify scheduler naming in run_log directory structure and update
scheduler references in config YAML files.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Suppress console output from workload stats scripts and save to tables/ directory
rajveerb+1: Claude Haiku 4.5
7bd86c9
- anns/compute_workload_stats.py: Replace all print() statements with file writes
* Add --output-dir parameter (default: "tables")
* Save output to workload_stats_anns.txt
- crawl/compute_workload_stats.py: Replace all print() statements with file writes
* Add --output-dir parameter (default: "tables")
* Save output to workload_stats_crawler.txt
No console output - all statistics are written to files in the specified output directory.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Remove oeda_pbas crawler run logs from H200 and H100
rajveerb+1: Claude Haiku 4.5
299e4c2
This scheduler is not included in the paper evaluation.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Add collected_outputs_streaming.json for ANNS H200 experiments
rajveerb+1: Claude Opus 4.6
ee6a98b
28 JSON files containing per-request token invalidation data needed
by the plot_tokens_invalidated_aggregated.py script (fig:tokens-invalidated-ccdf).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add workload data, run logs, and perf model files
