workflow-llm-eval

Serves AI models on AMD Instinct GPU hardware using vLLM. Use this skill whenever the user wants to run, serve, deploy, start, host, or launch a language model on an AMD GPU, AMD Instinct, MI300X, MI325X, MI350X, or MI355X. Also use when the user mentions vLLM on ROCm, vLLM on AMD, serving on HBM, or asks how to get a model running on AMD data center hardware. Use when the user asks "run Qwen3", "serve DeepSeek", "start a vLLM endpoint", "get a model running on my AMD machine", or any similar phrasing. Handles the full flow: GPU detection, environment validation, vLLM configuration, launch, and health verification. Do not use for NVIDIA GPUs, consumer AMD GPUs (RX series, Radeon), Ryzen AI, NPU, MI250X, or MI100.

Recommends which rocprofiler-systems runtime profiling preset (balanced, trace-gpu, trace-hpc, sys-trace, workload-trace, etc.) to pass to `rocprof-sys-run --preset=<name>` or `rocprof-sys-sample --preset=<name>` based on the user's workload and profiling goal, and returns the ready-to-run command. Use when the user asks "which preset should I use", describes a workload (GPU kernel, MPI, OpenMP, AI/ML training, production overview) and wants profiling guidance, or invokes `rocprof-sys-run`/`rocprof-sys-sample` without picking a preset. Do NOT use for CMake build presets (`cmake --preset debug|release|ci`) used to configure/compile rocprofiler-systems itself — that is a build-configuration concern, not a profiling-run concern.

Answers questions about what rocprofiler-systems configuration options exist and how to use them — e.g. "what hardware counters are available on my system", "what ROCm APIs can I trace", "what settings control sampling/output", "how do I persist my configuration" — by mapping the question to the right `rocprof-sys-avail` or `rocprof-sys-run --help=<topic>` command. Use when the user wants to discover or explore ROCPROFSYS_* settings, Timemory components, hardware counters, ROCm API tracing domains, or GPU metrics, rather than apply them. Do NOT use this for picking a starting --preset for a workload — that's `choosing-a-profiling-preset`.

Guides instrumentation with `rocprof-sys-instrument` - choosing runtime instrumentation vs binary rewrite (`-o`), including or excluding functions and modules with regex options, previewing what will be instrumented with a dry run (`--simulate`, `--dump-info`, `--print-*`), tuning granularity (`--min-instructions`, `--min-address-range`, `--instrument-loops`), rewriting libraries, and diagnosing why a function is missing from, or too numerous in, a trace. Use when the user asks about rocprof-sys-instrument, binary rewrite, instrumenting a shared library, function or module include/exclude regexes, dry runs, or the difference between a function and a module. Do NOT use to pick a `--preset` (use `choosing-a-profiling-preset`), to list settings, counters or ROCm domains (use `discovering-profiling-options`), or to attach to an already-running process (a separate tool, `rocprof-sys-attach`).

Serves an LLM on a supported AMD EPYC server CPU using vLLM with zentorch, in Docker, Podman, or conda. Use for "vLLM on CPU", "zentorch serving", or an EPYC CPU endpoint, including on a host that also has AMD Instinct GPUs. Detects the EPYC generation, validates the runtime, checks model support and RAM fit, sizes threads/KV/NUMA, confirms the plan, launches, and verifies the endpoint. Runs one instance on one socket and its memory. Reports and stops on failure; does not retry or debug. Use serving-llms-on-instinct when the endpoint should run on a GPU. Excludes multi-node, EPYC 4000, and pre-Zen4 EPYC without AVX-512.

See LICENSE for license information. -->

Inspects and tunes the shared-vs-dedicated memory split on AMD Ryzen APUs with unified memory (UMA) so larger LLMs and image-gen models fit on the iGPU, or so reserved GPU memory is returned to the CPU. Use when the user mentions Ryzen AI, Strix Halo / Strix Point / Krackan / Phoenix / Hawk Point, Ryzen AI Max, gfx1150 / gfx1151 / gfx1152, integrated Radeon, iGPU memory, UMA Frame Buffer Size, AMD Variable Graphics Memory, VGM, GTT, GART, TTM, pages_limit, amd-ttm, amd-debug-tools, "shared GPU memory", "dedicated GPU memory", carve-out, "not enough VRAM", "out of VRAM", "GPU OOM", llama.cpp on iGPU, ROCm on APU; or asks how much memory the iGPU can use, how to give the iGPU more memory, how to balance memory between CPU and GPU on UMA, or how to change the BIOS UMA reservation. Read-only diagnostics work everywhere; tuning runs automatically on Linux via `amd-ttm` and prints guided BIOS steps on Windows. Do not use for discrete Radeon cards, Intel iGPUs, or Apple Silicon -- it is APU-only.

Profiles a target application using rocprofv3 with program counter sampling enabled and then analyzes the results. Use when the user has a program that runs on an AMD GPU and asks to perform PC sampling, to determine the runtime performance characteristics of their application, to determine stall reasons, or to determine hotspots in the code.

Configure this repo for the engineering skills: set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.

Teach the user a new skill or concept, within this workspace.

Ask which skill or flow fits your situation. A router over the skills in this repo.

Plan a huge chunk of work (more than one agent session can hold) as a shared map of decision tickets on your issue tracker, and resolve them one at a time until the way to the destination is clear.

Hand the current conversation off to a fresh background agent that picks up the work immediately.

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.

Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions

Use when executing implementation plans with independent tasks in the current session