{"data":{"slug":"kaden-schutt-hipfire","name":"hipfire","tagline":"RDNA-native LLM inference engine in Rust","github_url":"https://github.com/Kaden-Schutt/hipfire","owner":"Kaden-Schutt","repo":"hipfire","owner_avatar_url":"https://avatars.githubusercontent.com/u/151092359?v=4","primary_language":"Rust","stars":554,"forks":61,"topics":["amd-gpu","gpu-computing","hip","llm-inference","machine-learning","quantization","rdna","rocm","rust"],"archived":false,"github_pushed_at":"2026-08-25T03:19:07+00:00","maintenance_label":"Very active","stars_delta_30d":63,"url":"https://www.graphcanon.com/tools/kaden-schutt-hipfire","markdown_url":"https://www.graphcanon.com/tools/kaden-schutt-hipfire.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/kaden-schutt-hipfire","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=kaden-schutt-hipfire","description":"RDNA-native LLM inference engine in Rust. ","homepage_url":null,"license":"Other","open_issues":96,"watchers":5,"ai_summary":"A tool for RDNA-native machine learning inference optimized for AMD GPUs using Rust and ROCM.","readme_excerpt":"## Install\n\nLinux with ROCm 6 or newer:\n\n```bash\ncurl -fsSL https://raw.githubusercontent.com/warpfront/hipfire/master/scripts/install.sh | bash\n\n---\n\n# Or install the integration branch for testing:\ncurl -fsSL https://raw.githubusercontent.com/warpfront/hipfire/master/scripts/install.sh \\\n  | bash -s -- --branch beta\n```\n\nRDNA4 requires ROCm 6.4 or newer. gfx1151 requires ROCm 7.2 or newer.\nRun `hipfire --version` for a concise build ID or `hipfire version` to compare\nthe installed binary, managed source checkout, and daemon. Managed Linux\ninstalls can switch revisions with `hipfire update @beta`,\n`hipfire update --tag v0.2.1`, or `hipfire update --commit <sha>`.\n\nTo uninstall a managed Linux install while keeping downloaded models and\nsettings:\n\n```bash\ncurl -fsSL https://raw.githubusercontent.com/warpfront/hipfire/master/scripts/uninstall.sh | bash\n```\n\nAdd `--dry-run` to preview it. `--purge` also deletes all data under\n`~/.hipfire` and requires explicit confirmation.\n\nFor Windows, source builds, and verifying the install:\n[docs/GETTING_STARTED.md](docs/GETTING_STARTED.md).\n\n---\n\n## Container (podman/docker)\n\nA multi-stage `Containerfile` builds a slim deliverable inference image and a\nfull-toolchain GPU gate-runner for reproducible PR/dev-build validation. See\n[docs/CONTAINER.md](docs/CONTAINER.md).\n\n```bash\npodman build -f Containerfile --target runtime -t hipfire .\npodman run --rm -it --device /dev/kfd --device /dev/dri \\\n  --group-add keep-groups --security-opt seccomp=unconfined \\\n  -v hipfire-models:/root/.hipfire/models \\\n  -v hipfire-kcache:/var/cache/hipfire \\\n  hipfire run qwen3.5:4b \"2+2=\"\n```\n\n---\n\n## License\n\nhipfire is licensed under **Apache-2.0** as of v0.3.0. See\n[LICENSE](LICENSE), [LICENSE-APACHE](LICENSE-APACHE), and\n[NOTICE](NOTICE) for details. Releases up to and including v0.2.1 were\ndual MIT/Apache-2.0; that grant is not revoked.\n\nIndividual files whose substantive authors have not elected Apache-2.0\nremain MIT-licensed (see [LICENSE-MIT](LICENSE-MIT)) and are identified\nby their per-file `SPDX-License-Identifier`. No contributor's file was\nrelicensed in absentia. New contributions default to Apache-2.0 via DCO\nsign-off. See [CONTRIBUTING.md](CONTRIBUTING.md) for the contributor\nside and\n[docs/governance/relicense-2026-05.md](docs/governance/relicense-2026-05.md)\nfor the decision record (including the 2026-05-19 course correction and\nthe v0.3.0 move to outbound Apache-2.0).\n\nOriginal architectural innovations originating in hipfire are\ncatalogued in [PRIOR-ART.md](PRIOR-ART.md); derivative works\n(including reimplementations informed by hipfire's design) should\nattribute the corresponding inventions per [AGENTS.md](AGENTS.md).","github_created_at":"2026-03-21T02:59:11+00:00","created_at":"2026-07-11T11:46:06.183923+00:00","updated_at":"2026-08-25T12:01:01.485175+00:00","categories":[{"slug":"inference-serving","name":"Inference & Serving","url":"https://www.graphcanon.com/categories/inference-serving","markdown_url":"https://www.graphcanon.com/categories/inference-serving.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/inference-serving"}],"tags":[{"slug":"amd-gpu","name":"amd-gpu"},{"slug":"gpu-computing","name":"gpu-computing"},{"slug":"hip","name":"hip"},{"slug":"llm-inference","name":"llm-inference"},{"slug":"machine-learning","name":"machine-learning"},{"slug":"quantization","name":"quantization"},{"slug":"rdna","name":"rdna"},{"slug":"rocm","name":"rocm"}],"trust":{"provenance":{"is_fork":false,"github_id":1187689258,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-08-25T12:01:00.455Z","maintenance":{"label":"Very active","score":96,"methodology":"github_public_v1","releases_90d":2,"days_since_push":0,"last_release_at":"2026-06-10T03:22:12Z","stars_delta_30d":63,"open_issues_delta_30d":25},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-11T11:46:07.352Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"scan":{"source":"repo_scan","observed_at":"2026-08-25T12:01:01.022Z"},"languages":{"value":["rust"],"source":"github.language","observed_at":"2026-08-25T12:01:01.022Z"},"license_spdx":{"value":"Other","source":"github.license","observed_at":"2026-08-25T12:01:01.022Z"}},"decision_facts":{"hosting":null,"pricing":null,"requirements":null,"constraints":null,"when_to_use":["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."],"when_not_to_use":["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."],"source":"enrich:decision_facts","observed_at":"2026-07-16T20:53:45.789Z"},"constraint_facets":null,"decision_summary":[{"label":"Adopt for","value":"HIPFire is an RDNA-native LLM inference engine in Rust, specifically optimized for AMD GPUs using ROCM."}]}}