异步 GPU 的计算与内存解耦
5 分•作者: yiyingzhang•4 个月前
一个很棒的开源项目，它引入了一种新的编程模型，用于解耦 NVIDIA GPU 的计算和内存，并支持异步内存操作（例如，Hopper）。与当前最先进的技术相比，性能提升了 12%，内核代码减少了 67%。
论文：《VDCores: Resource Decoupled Programming and Execution for Asynchronous GPU》 arXiv:2605.03190
查看原文
Cool open-source project that introduces a new programming model for decoupling compute and memory for NVIDIA GPUs that supports asynchronous memory operations (e.g., Hopper). 12% perf improvement over SOTA and 67% less kernel code.<p>Paper: "VDCores: Resource Decoupled Programming and Execution for Asynchronous GPU" arXiv:2605.03190