🚀 The paper “PyPIM: Integrating Digital Processing-in-Memory from Microarchitectural Design to Python Tensors” proposes the first end-to-end integration of digital bitwise PIM architectures from a low-level microarchitecture that supports partition-based operations all the way up to a high-level Python tensor library that is a drop-in replacement to libraries such as NumPy and PyTorch.
You can find the paper here: https://lnkd.in/d-S7ycrV and 💡already try out this new “PyTorch”💡 here: https://lnkd.in/d35bc-Eb
Great work Orian, congrats !!
