py-veo | Python bindings for VE Offloading for SX-Aurora

py-veo is a Python library. py-veo has no bugs, it has no vulnerabilities, it has build file available and it has low support. You can download it from GitHub.

The NEC Aurora Tsubasa Vector Engine (VE) is a very high memory bandwidth vector processor with HBM2 memory in the form-factor of a PCIe card. Currently up to eight VE cards can be inserted into a vector host (VH) which is typically a x86_64 server. The primary usage model of the VE is as a standalone computer which uses the VH for offloading its operating system functionality. Each VE card behaves like a separate computer with its own instance of operating system (VEOS), it runs native VE programs compiled for the vector CPU that are able to communicate with other VEs through MPI. A second usage model of VEs lets native VE programs offload functionality to the VH with the help of the VHcall mechanisms. The VH is used by the VE as an accelerator for functions it is better suited for, like unvectorizable code. The third usage model is the classical accelerator model with a main program compiled for the VH running high speed program kernels on the VE. A mechanism for this usage model is the VE Offloading (VEO) library provided by the veofload and veoffload-veorun RPMs. This Python module is an implementation of the VEO API for Python programs. It is an extension to the C API and exposes the mechanisms through Python objects.