PyRe GPU guide: run local Python scripts on Windows
PyRe GPU is a Windows workbench for adding and running local Python scripts. Review the resource allocation before starting, then inspect the output and run record afterward. CPU execution is available; compatible GPU execution is optional.
This guide covers the current public workflow. Your script and its required packages must be ready in the selected environment before you launch a run.
1. Prepare the script and its environment
Choose a local Python script you understand and install its required packages in the environment you intend to use. Review the code and any commands before running them. The published example Python script is a starting point for checking a compatible run.
PyRe coordinates selected runs; it does not rewrite your script or turn CPU-only code into GPU code. Read the step-by-step local script guide when adding your first workload.
2. Check the resource allocation before starting
Add the script to the workbench and review its execution plan and proposed resource allocation. Leave capacity for Windows and other work. If the app refuses a run, read its admission or environment message, reduce the request or free resources, and review the plan again.
The pre-run checklist covers the environment, resources and selected run or queue. Resource planning is a check to perform, not a guarantee that every workload will fit your machine.
3. Use a GPU only when the workload supports it
GPU execution requires compatible Python code, hardware, drivers and an environment with the needed packages. A GPU in the PC does not make an ordinary CPU script accelerated. Review your script's device settings and the output from its compatibility checks.
Start with a small workload and inspect its result. Do not infer a speedup from a device name or utilization percentage alone.
4. Read the completed run record
Open the run record and check its terminal outcome and exit code. Read the output log and any recorded resource observations. If a run failed, correct the script or environment and launch a fresh run.
The run-result guide explains these checks. A completed process and a correct result are separate questions; validate the output that matters for your work.
Download and current requirements
The Microsoft Store download is free. Check the Store listing for current system requirements and download terms, and the installed app for any timed full-feature trial details. GPU compatibility varies by device and workload.
Follow the current instructions for your installed version and selected Python environment when planning the next run.