@inproceedings{ author={ C. Chen, J. Yao, H. Subramoni, D. Panda }, title={ Design and Optimization of GPU-Aware MPI Allreduce Using Direct Sendrecv Communication }, conference={ NVIDIA GTC AI Conference 2026 }, year={ 2026 }, month={ March }, location={ San Jose, California }, source={ http://nowlab.cse.ohio-state.edu/publications/ }, }