warpt 0.3.0

Performance monitoring and system utilities

warpt: EarthFrame's Workload and Resource Performance and Transparency Toolkit

warpt is a unified command-line tool for hardware discovery, stress testing, and performance monitoring.

warpt provides a vendor-agnostic reporting interface for software and hardware resources. warpt brings transparency to system configuration, benchmarking, and stress testing, answering questions such as:

• "What hardware do I have?",
• "How much power is it actively using?",
• "Is it working correctly?", and
• "How fast is it?"

Installation

warpt requires Python 3.8 or newer, though we recommend 3.10+. warpt has full support on Mac OS X and Linux; Windows support is currently limited.

Note: We recommend using warpt in a virtualenv. Before installing warpt, you can create and activate a virtualenv by running:

python -m venv warpt-env
source warpt-env/bin/activate

You can install the basic warpt command set like so:

pip install warpt

Stress testing has some vendor-specific code at the moment; you can install this to enable all compatible stress tests by running:

pip install warpt[stress]

Quick Start

# Discover your hardware
warpt list

# Run CPU stress tests
warpt stress -c cpu

# Monitor system in real-time
warpt monitor

# Check power consumption (Linux/macOS)
warpt power

Features

Command	Description
warpt list	Detect CPU, GPU, memory, storage, and installed ML frameworks
warpt stress	Run stress tests across CPU, GPU, RAM, storage, and network
warpt monitor	Real-time system monitoring with TUI dashboard
warpt power	Power consumption monitoring and per-process attribution
warpt carbon	Track energy consumption, CO2 emissions, and estimated cost
warpt benchmark	Performance benchmarking suite
warpt integrate	AI-assisted hardware backend integration

Documentation

• Getting Started — Installation and first steps
• CLI Reference — Complete command and option reference
• Support Matrix — System requirements and platform compatibility

Platform Support

Platform	Status
Linux	Full support
macOS	Full support (power monitoring requires sudo)
Windows	Limited support (see Known Limitations)

GPU Support: NVIDIA GPUs supported. AMD, Intel, and Apple Silicon GPU support coming soon.

Carbon Tracking

warpt automatically tracks energy usage and CO2 emissions during stress tests and power monitoring. You can also track any workload manually:

# Automatic — built into stress tests
warpt stress -c cpu -d 30
# [carbon] 30.2s | 23.8W avg | 199.7 mWh | 0.08g CO2 | $0.0000 | less than breathing for a minute

# Manual — track any workload
warpt carbon start
# ... run your workload ...
warpt carbon stop

# View history and totals
warpt carbon history
warpt carbon summary

# Check available grid regions and carbon intensities
warpt carbon regions

Carbon calculations use regional grid intensity data to estimate CO2 emissions from energy consumption. Configure your region with --region (defaults to US).

Backend Integration

warpt integrate uses an AI agent (Claude Code SDK) to generate new hardware backend implementations from vendor SDK documentation. It creates the backend, power backend, factory registration, and tests automatically.

Prerequisites

warpt integrate requires Claude Code (Anthropic's CLI) and the Claude Code Python SDK.

1. Install Claude Code CLI:

   npm install -g @anthropic-ai/claude-code
   

2. Authenticate — choose one:

   • Anthropic API key (recommended for CI/automation):

     export ANTHROPIC_API_KEY=sk-ant-...
     

   • Interactive login (opens browser, uses your Claude account):

     claude login
     

3. Install warpt with integration dependencies:

   pip install warpt[integrate]
   

4. Verify setup:

   claude --version          # confirm CLI is installed
   warpt integrate --vendor test --sdk-docs ./some-docs/ --dry-run
   

   The --dry-run validates everything without starting the agent.

Quick Start

# Start a new backend integration
warpt integrate --vendor amd --sdk-docs ./amd-sdk/python/

# Check integration status
warpt integrate status

# Process answered questions
warpt integrate

# Validate generated code (ruff + pytest)
warpt integrate validate

# Reset and start over
warpt integrate reset

Dry Run

Use --dry-run to validate your setup before spending tokens on an agent session:

warpt integrate --vendor amd --sdk-docs ./amd-sdk/python/ --dry-run

This runs all validation (vendor name, existing backend check, doc loading, token limit) and prints a summary:

==================================================
DRY RUN SUMMARY
==================================================
  SDK docs:      ~18,500 tokens from 42 file(s)
  System prompt: ~12,000 tokens
  Total context: ~30,500 tokens (limit ~200k)

  Files to generate:
    warpt/backends/amd.py
    warpt/backends/power/amd_power.py
    tests/test_amd_backend.py
    questions.yaml

  Branch: backend/amd
==================================================

Press Enter to start the agent, or Ctrl+C to abort:

Press Enter to proceed into the full agent run, or Ctrl+C to abort without creating anything.

Post-Agent File Audit

After the agent session completes, warpt integrate automatically audits the generated files:

Generated files:
  questions.yaml: created (8 question(s))
  tests/test_amd_backend.py: created (320 lines)
  warpt/backends/amd.py: created (245 lines)
  warpt/backends/power/amd_power.py: created (180 lines)

If the agent modified files outside the expected set (factory.py, pyproject.toml, and the generated files), a warning is printed:

Warning: Unexpected file modifications:
  warpt/utils/helpers.py

Workflow

1. Init (--sdk-docs): Agent reads SDK docs, generates backend + tests, logs questions to questions.yaml
2. Review: Engineer reviews generated code and answers questions (set status: answered)
3. Iterate (no --sdk-docs): Agent reads answers, updates code, runs tests
4. Validate: Run warpt integrate validate to confirm ruff + pytest pass

Example Output

$ warpt list

CPU Information:
  Make:               Intel
  Model:              Xeon W-2295
  Architecture:       x86_64

Topology:
  Total Sockets:      1
  Total Phys Cores:   18
  Total Logic Cores:  36

Memory Information:
  Total:              128.0 GB
  Type:               DDR4

GPU Information:
  GPU 0: NVIDIA RTX 4090
    Memory: 24576 MB
    Driver: 545.23.08

Alpha Release

This is an alpha release. Some features are still in development:

• Carbon tracking — new in v0.2.0
• AMD GPU support (ROCm) — in progress
• Intel GPU support (oneAPI) — in progress
• Apple Neural Engine — in progress
• Additional benchmarks — expanding

See the Support Matrix for full details.

Feedback

We'd love to hear from you:

• Report bugs: GitHub Issues
• Feature requests: GitHub Issues

License

warpt is licensed under the MIT License — see LICENSE for details.