apache / tvm

Open deep learning compiler stack for cpu, gpu and specialized accelerators
https://tvm.apache.org/
Apache License 2.0
11.81k stars 3.48k forks source link

[mrvl][runtime]: Support Marvell Hardware Runtime #17498

Closed f2013519 closed 2 weeks ago

f2013519 commented 3 weeks ago

This PR adds support for the Marvell Hardware runtime, which enables executing compiled models on the Octeon target device.

This change is part of Phase 3 as outlined in the pre-RFC documentation: https://discuss.tvm.apache.org/t/prerfc-byoc-integrating-marvell-ml-ai-accelerator-to-the-tvm-byoc-framework/16155

The user can specify whether to compile for the simulator or for the hardware using a command line option. The documentation has been updated as well to describe the new feature.