hipfire

Triage and repair hipfire runtime failures such as daemon hangs, stale serve.pid, port 11435 conflicts, ROCm include-path problems, missing precompiled kernels, VRAM OOM, kernel JIT failures, and multi-turn recall regressions. Use after diagnostics identify a likely runtime issue or when the user asks to fix a broken hipfire serve/run flow.

Build terminal UIs in Rust with Ratatui. Use when creating TUI applications, immediate-mode rendering, high-performance terminal interfaces, or production Rust CLIs.

Port hipfire compute kernels to a new RDNA / CDNA architecture (gfx1201/gfx1200/gfx94x/gfx1150/etc.). Use when adding support for a new GPU arch, fixing arch-specific kernel codegen failures (e.g. "Cannot select intrinsic %llvm.amdgcn.wmma..."), or refactoring dispatch.rs's arch-conditional branches. Captures the WMMA operand-shape matrix, builtin name table per arch, dispatch routing convention, validation procedure (channel-test / coherence-gate / speed-gate), contributor onboarding workflow, and known correctness traps. Triggers on phrases like "port to gfx12", "9070 XT support", "R9700 support", "WMMA gfx12", "Cannot select intrinsic wmma", "amdgcn.wmma", "new arch port", "cross-arch kernel".

Optimize hipfire HIP/compute kernels — pick a tuning lever (multi-row, K-tile depth, prefetch, wave-size port, WMMA/MFMA, fused projections, ISA flags) and validate the win across the supported RDNA arch matrix. Use when you've identified a hot kernel, want to land a real perf win, and need to NOT regress on archs you don't have hardware for. Codifies the methodology from this repo's actual perf history — wave64 CDNA3 port (commit 4105035, 2× decode), nontemporal-load revert (34eb024, -13% caught only by clean-baseline bisect), gfx12 WMMA port (PR

Use when porting a hipfire feature/fix branch authored against pre-0.1.20 master onto post-modular master. Walks through the engine→hipfire-runtime + per-arch-crate split mechanically, then surfaces semantic conflicts that need human judgment.

Use Kernel Atlas to collect phase-aware hipfire measurements and render ISA Fit View visualizations for AMD GPU kernels, quant formats, and architectures. Use when a user asks how MQ/HFQ/HFP/Q8 quants occupy hardware, asks for an ASCII ISA visualization, wants to compare gfx1010/gfx1030/gfx11/gfx12 kernel fit, or wants an agent-readable "left on table" summary from Atlas rows.

Use for any question about a codebase, its architecture, file relationships, or project content — especially when graphify-out/ exists, where the question should be treated as a graphify query first. Turns any input (code, docs, papers, images, videos) into a persistent knowledge graph with god nodes, community detection, and query/path/explain tools.

Use when the user wants to design, redesign, shape, critique, audit, polish, clarify, distill, harden, optimize, adapt, animate, colorize, extract, or otherwise improve a frontend interface. Covers websites, landing pages, dashboards, product UI, app shells, components, forms, settings, onboarding, and empty states. Handles UX review, visual hierarchy, information architecture, cognitive load, accessibility, performance, responsive behavior, theming, anti-patterns, typography, fonts, spacing, layout, alignment, color, motion, micro-interactions, UX copy, error states, edge cases, i18n, and reusable design systems or tokens. Also use for bland designs that need to become bolder or more delightful, loud designs that should become quieter, live browser iteration on UI elements, or ambitious visual effects that should feel technically extraordinary. Not for backend-only or non-UI tasks.

Compile MLIR-AIE kernels for AMD XDNA NPUs via IRON + aiecc. Use when adding or iterating on an NPU compute kernel that feeds libhipfire_xdna1.so.

Run a hipfire model end-to-end via the daemon JSON-lines protocol. Use when the user wants to load a model and generate output, run a batch-prefill session, or validate a model loads and produces tokens on this machine.

Use when running or interpreting the unified Hipfire eval harness, choosing hipfire eval tiers/batteries/suites, validating quantized model candidates, collecting admission evidence, or deciding whether fast/medium/long/extensive eval results are sufficient for a quantization claim.

Use Hipfire's vendored AMD Matrix Instruction Calculator in ./third_party/amd_matrix_instruction_calculator to inspect WMMA, SWMMAC, MFMA, and SMFMAC instruction details, lane/register mappings, matrix layouts, operand modifiers, and throughput/register-use facts for AMD RDNA/CDNA kernel work. Use when designing or debugging WMMA/MFMA kernels, especially BF16/F16 16x16x16 mappings, C/D accumulator layouts, OPSEL/NEG modifiers, or arch-specific matrix-instruction availability.