hipfire logo

hipfire

Kaden-Schutt/hipfire

RDNA-native LLM inference engine in Rust

GraphCanon updated 1mo · GitHub synced 1mo

491 stars49 forksLast push 1mo Rust Other

Decision brief

HIPFire is an RDNA-native LLM inference engine in Rust, specifically optimized for AMD GPUs using ROCM.

Good fit when

  • You are working with AMD GPUs and want to optimize your inference tasks with machine learning models on these specific hardware setups.
  • Your project requires low-level optimizations that leverage the capabilities of RDNA architecture directly through Rust, ensuring efficient execution on AMD GPU infrastructure.

Avoid when

  • If you primarily use NVIDIA GPUs or any other non-AMD GPU type for your machine learning inference tasks, HIPFire may not provide optimized results due to its specialization in RDNA architecture.
  • Your environment does not support ROCM software stack; HIPFire requires this infrastructure to function optimally in conjunction with AMD RDNA GPUs.

Observed Jul 16, 2026 · Source: enrich:decision_facts

Verify the decision

Maintenance and security

Full trust report
Maintenance
Very active (0d since push)
As of 1mo
Provenance
Not a fork · Organization account
As of 1mo
Security (OSV)
No lockfile
As of 1mo

Public GitHub metadata and optional OSV scans. Signals, not a guarantee. Trust methodology.

Install

cargo add hipfire
crates.io

Similar tools

Same-category neighbours. No typed graph edges are catalogued for this tool yet.

Evidence and technical details

Sourced facts, taxonomy, compatibility claims, README excerpt, and machine-readable endpoints.

Overview

A tool for RDNA-native machine learning inference optimized for AMD GPUs using Rust and ROCM.

Capability facts

Languages
rust

Source: github.language · Jul 25, 2026

Categories

Tags

README

Install

Linux with ROCm 6+:

curl -L https://raw.githubusercontent.com/Kaden-Schutt/hipfire/master/scripts/install.sh | bash

For Windows, source builds, and verifying the install: docs/GETTING_STARTED.md.


License

hipfire is dual-licensed under MIT or Apache-2.0 at your option. See LICENSE (dual-license pointer), LICENSE-MIT, LICENSE-APACHE, and NOTICE for details.

New contributions default to Apache-2.0 via DCO sign-off; existing contributors' MIT-licensed contributions remain MIT unless they opt in. Each source file carries an SPDX-License-Identifier reflecting actual authorship (MIT, Apache-2.0, or MIT OR Apache-2.0). See CONTRIBUTING.md for the contributor side and docs/governance/relicense-2026-05.md for the decision record (including the 2026-05-19 course correction from a unilateral Apache-2.0 relicense to dual licensing).

Original architectural innovations originating in hipfire are catalogued in PRIOR-ART.md; derivative works (including reimplementations informed by hipfire's design) should attribute the corresponding inventions per AGENTS.md.

For agents

This page has a .md twin and JSON over the API.

Was this helpful?

Anonymous feedback helps us improve pages and translations.