Ars Technica

Researchers upend AI status quo by eliminating matrix multiplication in LLMs

Researchers claim to have developed a new way to run AI language models more efficiently by eliminating matrix multiplication from the process. This fundamentally redesigns neural network operations ...
VentureBeat

New Transformer architecture could enable powerful LLMs without GPUs

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Matrix multiplications (MatMul) are the ...
Semiconductor Engineering

Lower Energy, High Performance LLM on FPGA Without Matrix Multiplication

A new technical paper titled “Scalable MatMul-free Language Modeling” was published by UC Santa Cruz, Soochow University, UC Davis, and LuxiTech. “Matrix multiplication (MatMul) typically dominates ...
HotHardware

LLMs On Lightbulb Power? Researchers Develop Low Energy AI Breakthrough

Researchers at the University of California, Santa Cruz have made a breakthrough by creating a large language model (LLM) running on custom hardware that only sips a mere 13 watts, which is the ...
Ars Technica

Researchers upend AI status quo by eliminating matrix multiplication in LLMs | Page 8 | Ars OpenForum

I was under the opinion back when convolutional neural nets were all the rage that this stuff was way too expensive to be practical and something else would take over. LLMs are a step in that ...