What historical constraints and design choices lead to the design of Tensor/Storage (and their Impl variants) as they are today? Why do we use intrusive refcounting? Why are we trying to get rid of virtual methods on TensorImpl? Why are there so many frickin' bitfields?Further reading.<ul><li>PyTorch internals blog post <a href="http://blog.ezyang.com/2019/05/pytorch-internals/">http://blog.ezyang.com/2019/05/pytorch-internals/</a></li><li>Writing Python in C++, a manifesto <a href="https://github.com/pytorch/pytorch/wiki/Writing-Python-in-cpp-(a-manifesto)">https://github.com/pytorch/pytorch/wiki/Writing-Python-in-cpp-(a-manifesto)</a></li><li>At time of writing, the breakdown of all fields on TensorImpl <a href="https://github.com/pytorch/pytorch/blob/71f4c5c1f436258adc303b710efb3f41b2d50c4e/c10/core/TensorImpl.h#L2155-L2177">https://github.com/pytorch/pytorch/blob/71f4c5c1f436258adc303b710efb3f41b2d50c4e/c10/core/TensorImpl.h#L2155-L2177</a></li></ul>

Conjugate views

The PyTorch Developer Podcast is a place for the PyTorch dev team to do bite sized (10-20 min) topics about all sorts of internal development topics in PyTorch.