byarielm.fyi / remoras

2

This repository has no description

2

remoras / bench /

at main 6 folders 2 files

byarielm.fyi Rework benchmark design as architecture-neutral suite with operational companion 3hrs ago

Add llm docs and bench spike code - llm/: design docs, specs, evidence, openspec, and archive - bench/: performance spike harnesses (oplog, yrs, loro, kleppmann fuzz) - .gitignore: exclude .DS_Store, Rust target/, node_modules/

5 days ago

ios-spike-oplog

Add llm docs and bench spike code - llm/: design docs, specs, evidence, openspec, and archive - bench/: performance spike harnesses (oplog, yrs, loro, kleppmann fuzz) - .gitignore: exclude .DS_Store, Rust target/, node_modules/

5 days ago

Add llm docs and bench spike code - llm/: design docs, specs, evidence, openspec, and archive - bench/: performance spike harnesses (oplog, yrs, loro, kleppmann fuzz) - .gitignore: exclude .DS_Store, Rust target/, node_modules/

5 days ago

Publish op-log-bench do_op per-scale results (10k–1M); all ops constant-time O(1)

4 days ago

Add PM spike: reactive invalidation, CPM scheduling, SQLite recursive CTE Benchmarks for project/task management feasibility: - Reactive invalidation: 241ns at 10k blocks, O(affected) not O(n) - Critical path method (CPM): 9.6ms at 10k tasks via Steel scheduler - SQLite recursive CTE for dependency queries: 7.7µs at 10k tasks - Graph generator for synthetic DAG workloads Results documented in evidence/30-pm-spike-results.md.

4 days ago

yrs-per-block-spike

Add llm docs and bench spike code - llm/: design docs, specs, evidence, openspec, and archive - bench/: performance spike harnesses (oplog, yrs, loro, kleppmann fuzz) - .gitignore: exclude .DS_Store, Rust target/, node_modules/

5 days ago

Rework benchmark design as architecture-neutral suite with operational companion

3 hours ago

benchmark-suite.md

Rework benchmark design as architecture-neutral suite with operational companion

3 hours ago

README.md

bench — research & validation crates#

Standalone Rust/JS crates that validate architecture decisions before production implementation. Each crate proves a specific property of the op-log + Kleppmann tree + Yrs architecture (Path C+).

These crates are LLM-generated with minimal human editing, same as the llm/ docs. They serve as validation evidence, not production code.

Benchmark suite#

The full benchmark suite — including planned crates, classification (spike vs. requires-model-settled vs. built-implementation), and implementation phases — is documented in ./benchmark-suite.md. The design spec for the architecture-neutral realistic workload benchmark is ../llm/design/35-realistic-workload-benchmark.md.

Current (Path C+ — op log as source of truth)#