d-Matrix, a provider of high-efficiency AI computation and inference, partners with Microsoft to use its low-code reinforcement learning (RL) platform, Project Bonsai, enabling a trained compiler to AI for in-memory digital computation of d-Matrix (DIMC).
The easy-to-use Project Bonsai platform accelerates time to value, with an out-of-the-box solution that reduces development effort using an AI-based compiler that leverages d’s ultra-efficient DIMC technology. -Matrix, depending on the supplier.
The combination of d-Matrix technology with Project Bonsai enables the efficient creation of a compiler for the DIMC platform. The Bonsai project is accelerating the rapid prototyping, testing, and deployment of RL agents trained in the compiler stack to take full advantage of d-Matrix’s low-power AI inference technology that can deliver up to ten times the energy efficiency of older architectures.
“d-Matrix has built the world’s most efficient computing platform for large-scale AI inference,” said Sudeep Bhoja, co-founder, CTO at d-Matrix. “What attracted us to Project Bonsai was its product features and ease of use. Microsoft’s unique offering is built around machine learning and the Inkling language, making RL constructs fully explainable.
d-Matrix transforms the economics of complex processors and generative AI with a scalable platform designed to handle the immense data and power demands of inference AI, making power-hungry data centers more effective, according to the company.
This new AI computing platform from d-Matrix uses an ingenious combination of smart ML tools and integrated software architectures using chips in a grid formation of Lego blocks, which enables the integration of multiple engines programming in a common package.
The RL-based compiler is expected to become a key differentiator in d-Matrix’s first-generation DIMC product offering, CORSAIR, set to ship in late 2023.
“We worked together to develop the RL-based compiler,” said Kingsuk Maitra, Principal Applied AI Engineer at Microsoft, with the Project Bonsai team. “We made it a point to have a product mindset from the start. The embodiments, including the instruction set architecture, were checked and validated on two d-Matrix test chips, NightHawk and JayHawk, and integrated into the RL training environment.The low-code attributes of the Bonsai project facilitated early development work, as well as the ability to integrate statistical control parameters and simplify the integration of other real chip design constraints, with very promising results so far.
For more information on this news, visit www.d-matrix.ai.