产品76°
PyTorch用FBTriton替代CUDA提升TBE性能
精选理由
PyTorch团队用FBTriton替代CUDA，推荐系统模型速度翻倍，开发者效率大增。
PyTorch团队用FBTriton替代CUDA实现Table Batched Embedding(TBE)内核。该优化使推荐系统模型前向速度提升最高1.28倍，后向速度提升2倍。开发者工作效率也获得显著提升。
原文 · PyTorch
Replacing CUDA with FBTriton for Table Batched Embedding (TBE) kernels delivered massive gains for our rec sys models.
@Meta contributors achieved up to 1.28x faster forward passes, 2x faster backward passes, and a huge boost in developer velocity. Check out the engineering deep-dive: https://t.co/qdP0nnoLgF
✍️ Daohang Shi, Oleksandr Stashuk, Rupert Wu, Liangbei Xu, Rich Zhu