amdresearch/intellikit/skill

Metrix: GPU Profiling

Profile AMD GPU kernels and get human-readable metrics (bandwidth, cache, coalescing, FLOPS). Architecture is auto-detected.

When to Use

• User asks to profile a GPU application or kernel
• Performance analysis, optimization, or bottleneck investigation
• Need HBM/L2/L1 bandwidth, hit rates, or compute metrics
• Need timing-only runs (fast, no hardware counters)

Instructions

1. Ensure the target runs on AMD ROCm (e.g. hipcc-built binary or Python script that launches HIP/ROCm kernels).
2. Choose execution path:

• If a Metrix MCP server is available, use list_available_metrics to discover metric names, then profile_metrics to collect them (or omit metrics to collect all).
• Otherwise run the CLI or Python API from the environment where Metrix is installed.

CLI

From the project or install prefix:

# Profile with all metrics (auto-detected arch)
metrix profile ./my_app
 
# Time only (fast, no counters)
metrix profile --time-only -n 10 ./my_app
 
# Filter kernels by name
metrix profile --kernel matmul ./my_app
 
# Specific metrics
metrix profile --metrics memory.l2_hit_rate,memory.coalescing_efficiency,compute.total_flops ./my_app
 
# Save to JSON/CSV
metrix profile -o results.json ./my_app

Options: --profile/-p (run metrix list profiles for names: quick, memory, memory_bandwidth, memory_cache, compute), --metrics/-m, --time-only, --kernel/-k (regular expression), --num-replays/-n, --output/-o, --top, --aggregate, --timeout, --no-counters, --log/-l, --quiet/-q. Discovery: metrix list <metrics|profiles|devices>, metrix info <metric|profile> <name>. Note: metrix list counters and metrix info counter <name> are not implemented yet (CLI reports “not yet implemented”).

Python API

from metrix import Metrix
profiler = Metrix()
results = profiler.profile("./my_app", num_replays=5)
for kernel in results.kernels:
    print(kernel.name, kernel.avg_time_us)
    for metric, stats in kernel.metrics.items():
        print(f"  {metric}: {stats.avg}")

Use metrics=[...] for a subset; omit for all metrics. Use cwd when the binary expects a specific working directory.

Workflow

1. Identify the executable or script to profile (e.g. ./app or python run_kernels.py).
2. If only timing is needed, use --time-only for speed.
3. If full metrics are needed, run metrix profile ./app (or MCP equivalent); optionally restrict with --kernel or --metrics.
4. Interpret results: low L2 hit rate, low coalescing, or low HBM utilization suggest optimization targets.
5. For automation or tooling, use -o results.json and parse the JSON output.

Key Metrics (reference)

• Memory: memory.hbm_bandwidth_utilization, memory.l2_hit_rate, memory.l1_hit_rate, memory.coalescing_efficiency, memory.global_load_efficiency, memory.lds_bank_conflicts, memory.atomic_latency
• Compute: compute.total_flops, compute.hbm_gflops, compute.hbm_arithmetic_intensity, compute.l2_arithmetic_intensity, compute.l1_arithmetic_intensity

Use relative paths for the target binary and output files so the skill is portable across environments.