SOSP Experiment Design & Audit
Designs and audits rigorous experimental evaluations for SOSP and top-tier systems research papers.
SOSP Experiments helps systems researchers design, structure, and audit the evaluation section of SOSP papers before submission. Because SOSP author responses forbid new experiments, this skill ensures every performance and correctness claim is tied to sound evidence, robust state-of-the-art baselines, and properly tuned workloads. It guides the setup of ablation studies, tail latency distributions (p99/p999), trace-driven microbenchmarks, failure testing, and transparent reporting of overheads and adverse results.
Key Features
01Audits experimental sections for blind spots, overhead hiding, and unfair configurations
02Establishes fair, documented baselines against published state-of-the-art systems
03Maps every paper claim directly to reproducible experimental evidence and ablation studies
041,090 GitHub stars
05Enforces rigorous tail latency and variance reporting standards (p50/p99 spreads and offered load)
06Designs multi-layer workloads across microbenchmarks, end-to-end applications, and traces
Use Cases
01Structuring the experimental matrix and evaluation section before writing a systems conference submission
02Formulating rigorous ablation and fault-injection studies to isolate core system mechanisms
03Conducting a pre-submission review of benchmark fairness and baseline tuning to survive PC review