Collect community CPU/GPU benchmark results #7

Community benchmark collection

We are collecting reproducible CPU/GPU benchmark results for the tensor factor engine. Slow or negative results are useful too; the goal is to understand machine and runtime differences, not to cherry-pick the fastest number.

Canonical CPU run: protocol v1

From the repository root, on Windows, macOS, or Linux:

python scripts/benchmark_tensor_factors.py --json-out artifacts/benchmark-v1.json

If GNU Make is installed, make benchmark is an optional shortcut for the same protocol.

Protocol v1 fixes:

• 750 dates ×ばつ 1,000 stocks;
• 20-day rolling window;
• seed 42;
• 3 warmup runs and 10 measured repetitions;
• one PyTorch intra-op thread and one inter-op thread.

The console output and JSON report include the protocol, Python, PyTorch, platform, CPU, logical CPU count, thread counts, CUDA availability, panel shape, window, warmup/repeat count, seed, and per-case mean/std timings.

Useful variants

After the canonical CPU run, variants are welcome:

python scripts/benchmark_tensor_factors.py --device cuda --json-out artifacts/benchmark-cuda.json
python scripts/benchmark_tensor_factors.py --device cpu --n-dates 1500 --n-stocks 3000 --json-out artifacts/benchmark-large.json

Please label GPU, larger-panel, and multi-thread runs separately from protocol v1.

What to submit

1. Commit SHA.
2. Exact command.
3. The generated JSON report as an attachment.
4. Complete console environment and result tables.
5. CPU/GPU model, CUDA details where relevant, and any thermal, memory, or workload notes.

See the benchmarking guide, benchmark board, and benchmark result form.