What makes OxCaml different from standard OCaml?

OxCaml is a specialized fork that includes Jane Street extensions designed for high-performance computing, such as stack allocation, unboxed types, and a more granular modal type system.

When should I use the 'exclave_' keyword?

You use 'exclave_' when you need to return a value that was allocated on the stack to the caller's frame, preventing it from escaping and causing memory errors.

How does the OxCaml skill help reduce Garbage Collection overhead?

The skill provides patterns for using stack allocation (local values) and unboxed types, which keep data off the OCaml heap, thereby reducing the frequency and duration of GC cycles.

Does this skill support SIMD operations?

Yes, it includes detailed guidance for using 128-bit and 256-bit SIMD vector types and intrinsics for parallel numeric processing.

What are 'Modes' in the context of this skill?

Modes are annotations that track runtime properties of values, such as their locality (stack vs heap), uniqueness (number of references), and portability across threads.

OxCaml Performance Development

Name: OxCaml Performance Development
Author: avsm

byavsm

•

개발자 도구

Optimizes OCaml applications using Jane Street extensions for stack allocation, unboxed types, and SIMD-driven parallelism.

This skill empowers Claude to master the OxCaml compiler, a performance-centric fork of OCaml featuring Jane Street's advanced extensions. It provides precise implementation patterns for high-performance systems, focusing on manual memory management through stack allocation, reduction of GC overhead via unboxed types, and data-race-free parallelism through modal types. Whether you are building low-latency trading engines or data-intensive numeric libraries, this skill ensures your code leverages modern CPU features like SIMD and zero-allocation annotations for maximum efficiency.

주요 기능

01Expert implementation of stack allocation (local_, stack_, exclave_)

02Zero-allocation code path verification and [@zero_alloc] annotations

03Optimization via unboxed numeric types and mixed-block records

0413 GitHub stars

05Advanced modal type system management (local, unique, once, portable)

06SIMD vector integration for 128-bit and 256-bit parallel operations

사용 사례

01Optimizing numeric compute kernels with SIMD and unboxed float arrays

02Developing low-latency financial trading systems requiring minimal GC pauses

03Implementing memory-safe, data-race-free parallel algorithms using modal types

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add avsm/ocaml-claude-marketplace oxcaml

For use in Claude.ai and ChatGPT

Download Skill