EnergyIR
Measurement-first efficiency tooling for numerical and ML Python.
pip install energyir installs the energy CLI — diagnostic tools that measure-verify
performance/energy fixes in your numpy / PyTorch / JAX / pandas code, plus BLAS, dependency-upgrade,
and CVE-reachability checks, and a reuse runtime. Diagnosis and measurement are free; advanced
optimization features are licensed.
- Install:
pip install energyir(Python 3.11+) - Docs: https://energyir.io/docs
- License: Apache-2.0
- Contact: david@energyir.io
EnergyFlow — whole-program compilation
energyir.flow sees a whole application rather than one call: what depends on
what, which state continues and which resets, which draws are independent, which
stages are barriers, and which groupings are legal. It then lowers every
candidate to the residency compiler, which remains the only thing that decides
where values live and what crossing the boundary costs.
from energyir import flow
with flow.capture(name="bundle") as f: # existing code, unchanged
x, _ = thermo_solve(A, b)
d, _ = thermo_logdet(A)
compilation = flow.compile_program(f.build())
print(flow.explain(compilation)) # every decision, every refusal
It reports hard counts, never a projected joule, and it refuses rather than guessing: an unsupported state mode, a random stream two "independent" replicas would share, a commit that could race its own verification, or a comparison between a measured number and a modelled one all produce a named refusal.
energy flow check|compile|explain|run|verify|targets|capture-registry
1 maintainer
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file energyir-1.23.1.tar.gz.
File metadata
- Download URL: energyir-1.23.1.tar.gz
- Upload date:
- Size: 2.8 MB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
f62b1eb69e08b41e244332cbc42167b7290ed2a8987a18bf235398deb0974ada
|
|
| MD5 |
b197810912d3bf6da5fe62224c272f53
|
|
| BLAKE2b-256 |
3042bfcaca0fea0df426538d71cd92942c63c722f4fbf70885f525a6ec0ce8c3
|
File details
Details for the file energyir-1.23.1-py3-none-any.whl.
File metadata
- Download URL: energyir-1.23.1-py3-none-any.whl
- Upload date:
- Size: 3.3 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
47fc6ec06c94eb9656bc1ba84030ba5468679c79c2d4dcca402f7f17791b8bca
|
|
| MD5 |
bac74997b2ed5ea627538d6467f2c824
|
|
| BLAKE2b-256 |
4c539a5de1a30a3482ceb54020acde4e78329ff79b582c7d4802a7aa6669ac5e
|
