Yifan Li 李轶凡
I'm currently a 1st-year PhD student in Computer Science from Cornell University. My advisor is Prof. Giulia Guidi. I was an undergraduate student majoring in Computer Science and Technology at Tsinghua University (China). I graduated with a GPA of 3.98/4.00 and was selected as an Outstanding Graduate (3/180).
My research interests revolve around parallel computing. Currently I'm exploring (1) core parallel computing and sparse linear algebra, and (2) applied parallel computing and computational science, potentially with heterogeneous hardware.
I was a member of the champion-winning Tsinghua Student Cluster Competition team during my undergraduate studies.
You can contact me via l@iyi.fan.
Picture Credit to Jiayuan.
News
2026-05-15: I am awarded the SIGHPC travel grant to ICS2026, with an award amount of $1600.
2026-04-07: Our paper "Ocean: Fast Estimation-Based Sparse General Matrix-Matrix Multiplication on GPU" was accepted to ICS2026; paper is now available at arxiv.
2025-08-15: I have started my PhD at Cornell University, advised by Giulia Guidi.
2024-11-23: Our team was the overall winner of Student Cluster Competition 24 at SC24! I was responsible for the Reproducibility Challenge.
2024-07-01: I was admitted to the highly competitive Summer@EPFL program ( 1.6% acceptance rate ), with a stipend of 1800CHF/month. I'll work at Prof. Sanidhya Kashyap's RS3Lab for the summer.
2024-06-13: Our paper "High-Performance Sorting-Based K-mer Counting in Distributed Memory with Flexible Hybrid Parallelism" was accepted to ICPP24; paper is now available at arxiv.
2024-05-20: Talked about "Counting K-mers on distributed memory efficiently with sorting and task-based parallelism" at MemPanG24. Slides.
Publication
Y. Li and G. Guidi, "Ocean: Fast estimation-based sparse general matrix-matrix multiplication on GPU," arXiv preprint arXiv:2604.19004, 2026.
Y. Li et al., "Critique of “Data Flow Lifecycles for Optimizing Workflow Coordination” by SCC Team From Tsinghua University" in IEEE Transactions on Parallel & Distributed Systems, doi: 10.1109/TPDS.2025.3627225.
Y. Li and G. Guidi, "High-performance sorting-based k-mer counting in distributed memory with flexible hybrid parallelism," in Proc. 53rd Int. Conf. Parallel Process. (ICPP), 2024, pp. 919–928.
Research
Ongoing Projects
2025-: Faster GRG
GRG (Genotype Representation Graph) is a new data structure for representing genomic variation. Compared to traditional methods, GRG speedups operations on the genotype matrix by 1-2 orders of magnitude. We're working on further optimize GRG, from a computer science perspective.
Past Projects
2025-2026: Ocean: Single-GPU SpGEMM kernel
We accelerated sparse matrix operations on a single GPU using estimation-based techniques. We were the first to introduce HyperLogLog estimators to sparse linear algebra, and used HLL to replace the symbolic computation in SpGEMM. We designed a novel workflow to predict the price of estimation and dynamically select the optimal workflow. Compared to previous state-of-the-arts, our approach achieved speedups of 1.4x-2.8x.
2023-2024: HySortK: High Performance Distributed K-mer Counting
We worked on distributed memory k-mer counting. With novel design and careful implementation, our application successfully scaled to 128 nodes on Perlmutter, and achieved a speedup of 2x compared to previous state-of-the-arts.