research

An in depth look at Tenstorrent Wormhole, originally posted on corsix.org

An in depth look at Tenstorrent Wormhole, originally posted on corsix.org

When implementations of the Transformer's self-attention layer utilize SRAM instead of DRAM, they can achieve significant speedups.

Today we are pleased to announce the release of our RISC-V Architectural Compatibility Suite, now available in our GitHub repository.

Training directed neural networks typically requires forward-propagating data through a computation graph..

We extend the capabilities of neural networks by coupling them to external memory resources.

A lot of our code is in the process of being transitioned from Software 1.0 (code written by humans) to Software 2.0 (code written by an optimization, commonly in the form of neural network training)...

This week at the ASPLOS 2021 conference, Dr. Chris Lattner gave the keynote address to open the event with a discussion of the Golden Age of Compilers in an era of hardware and software co-design.