Blaize released a computing architecture that, according to the company, meets the demands of new computational workloads found in artificial intelligence (AI) applications.
The Blaize GSP architecture and Blaize Picasso software development platform aids in computational efficiency by blending dynamic data flow methods and graph computing models with programmable proprietary SOCs. Allowing Blaize computing platforms to exploit the native graph structure inherent in neural network workloads through runtime. The efficiency multiplier is delivered via a data streaming mechanism, where non-computational data movement is minimized or eliminated.
With the graph-native structure, developers can build multiple neural networks and entire workflows on a single architecture that is applicable to many markets and use cases. End-to-end applications can be built integrating non-neural network functions such as Image Signal Processing with neural network functions. And AI Application developers can build entire applications, optimize these for edge deployment constraints, and run them using automated data-streaming methods.
For more information, please visit: https://www.blaize.com/