TurboPrefill
TurboPrefill accelerates multi-GPU LLM prefill by pipelining microbatches across GPUs, reducing idle time and inter-GPU communication bottlenecks.
TurboPrefill accelerates multi-GPU LLM prefill by pipelining microbatches across GPUs, reducing idle time and inter-GPU communication bottlenecks.
Log in or sign up for Devpost to join the conversation.