Tensor Streaming processorMXM: For performing matrix opsSXM: For shift and rotate operations on vectorsMEM: Memory read/write opsVXM: Arithmetic ops on the vectorsICU: Instruction Control Unit ArchitectureA Deep Dive into the Underlying Architecture of Groq's LPUWhat powers the ground breaking performance of Groq's Langauge Processing Unit?https://codeconfessions.substack.com/p/groq-lpu-design