Skip to content

Claude Code plugin that turns Claude into an Apple Silicon GPU systems engineer. 753 knowledge findings across 11 GPU computing domains, 3 specialized agents, 6 commands, Metal shader templates, and 212 BATS tests.

active 2026-02-092026-02-26 (UTC)

Complete coverage26,581 / 26,581 hourly files (100%) · 2 absent upstream2023-08-152026-08-26 (UTC)
Events
130
Pushes
46
Pull requests
5
Issues
60
Stars
3
Forks
1

Activity over time

Daily event counts in the loaded window

Line chart, 18 days from 2026-02-09 to 2026-02-26. Pushes: 46 total, peak 8 in a day. Pull requests: 5 total, peak 2 in a day. Issues: 60 total, peak 22 in a day. Comments: 9 total, peak 4 in a day. Stars: 3 total, peak 1 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

ContributorContributionsPushesPRsComments
kavanaghpatrick1204659

Recent activity

Latest issues, pull requests and releases

  • Issue#51kavanaghpatrick2026-02-26 18:36
    ForgeDB Phase 1: Read Data — Parquet reader + streaming chunk pipeline
  • Issue comment#51kavanaghpatrick2026-02-26 18:36
    ForgeDB Phase 1: Read Data — Parquet reader + streaming chunk pipeline
  • Issue#62kavanaghpatrick2026-02-26 18:31
    forge-datafusion: audit and replace CPU-bound DataFusion operators with GPU primitives
  • Issue#58kavanaghpatrick2026-02-26 13:58
    tpch-bench: document hand-tuning vs realistic SQL execution gap
  • Issue#56kavanaghpatrick2026-02-26 13:27
    Direct atomic GROUP BY AVG (f32): single-dispatch average aggregation
  • Issue comment#56kavanaghpatrick2026-02-26 13:27
    Direct atomic GROUP BY AVG (f32): single-dispatch average aggregation
  • Pull request#46kavanaghpatrick2026-02-26 12:04
  • Issue#53kavanaghpatrick2026-02-26 11:19
    ForgeDB Phase 3: Fuse Data — DataFusion GPU executors + fused kernels
  • Issue#52kavanaghpatrick2026-02-26 11:19
    ForgeDB Phase 2: Skip Data — zone maps, partition pruning, GPU decompression
  • Issue#49kavanaghpatrick2026-02-26 11:18
    Parquet reader + Arrow IPC for columnar file access
  • Issue#45kavanaghpatrick2026-02-25 23:20
    SDK query operators: expression evaluator, predicate fusion, top-k fix, SQL execution
  • Issue#40kavanaghpatrick2026-02-25 23:03
    Out-of-core chunk streaming: process datasets larger than RAM
  • Issue comment#37kavanaghpatrick2026-02-25 23:03
    Sort-based GROUP BY: compose existing primitives into high-cardinality aggregation
  • Issue#35kavanaghpatrick2026-02-25 23:02
    forge-runtime: Optimization roadmap from forge-bench v1 (20 experiments, M4 Pro)
  • Issue comment#35kavanaghpatrick2026-02-25 23:02
    forge-runtime: Optimization roadmap from forge-bench v1 (20 experiments, M4 Pro)
  • Issue#37kavanaghpatrick2026-02-25 17:04
    Sort-based GROUP BY: compose existing primitives into high-cardinality aggregation
  • Issue#37kavanaghpatrick2026-02-25 17:04
    Sort-based GROUP BY: compose existing primitives into high-cardinality aggregation
  • Issue#36kavanaghpatrick2026-02-25 13:50
    SDK Primitive Upgrades from Forge-Bench v2 Experiments (7 improvements, 2x-130x validated)
  • Issue#23kavanaghpatrick2026-02-23 12:02
    KB Gap: Chunk-wise linear attention algorithms (FLA kernel internals)
  • Issue comment#19kavanaghpatrick2026-02-23 12:02
    KB Gap: CPU→GPU algorithm transformation patterns (recursion, allocation, collections, control flow)
  • Issue#18kavanaghpatrick2026-02-23 12:02
    KB Gap: Rust-to-GPU compilation toolchains (rust-gpu, CubeCL, Naga, MSL codegen)
  • Issue comment#17kavanaghpatrick2026-02-23 12:02
    KB Gap: SDF-based GPU-native GUI rendering (fonts, UI primitives, compute rasterization)
  • Issue comment#16kavanaghpatrick2026-02-23 12:02
    KB Gap: vLLM-MLX scheduler and KV cache internals
  • Issue comment#15kavanaghpatrick2026-02-23 12:02
    KB Gap: M5 Neural Accelerator constraints and operational limits
  • Issue#11kavanaghpatrick2026-02-23 12:02
    KB Gap: Flash Attention MSL implementation patterns (llama.cpp + MLX)

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 3 stars here means stars gained during the window, not the repo's star count.