High Performance Computing Systems and Architecture Group

Welcome to the High Performance Computing Systems and Architecture Group Website

Our research focuses on optimizing performance of those emerging and mission-critical applications on increasingly heterogeneous hardware. Hardware heterogeneity, commonly evidenced by memory heterogeneity and machine learning accelerators poses challenges on programming methods, performance analysis and modeling, runtime design, and architecture design. HPCSA@Merced is composed of researchers with strong expertise on high performance computing (particularly runtime systems, compiler, interconnect and HPC applications) and accelerator architecture. Our research is widely supported by various government agencies (e.g., NSF, DOE and California) and industry (e.g., NVIDIA and Intel).

News

  • [9/2021] Prof. Changhee Jung from Purdue will give us a talk on Sept 7, and Prof. Jia Rao from UT Arlington will give us a talk on Sept 21.

  • [8/2021] Dr. Sergey Blagodurov from AMD will give us a talk on heterogeneous memory on Aug 31.

  • [7/2021] Prof. Su received his first NSF grant on Jul 21.

  • [6/2021] Prof. Li received an NSF grant on big memory computing.

  • [5/2021] Prof. Daniel Wong from UC Riverside will give us a talk on May 11.

  • [3/2021] Our collaboration work with LLNL on MPI fault tolerance benchmark suite is reported by HPCWire.

  • [3/2021] Wenqian Dong got an internship offer! She will go to the HP labs during the summer working on scientific machine learning.

  • [3/2021] Profs. Jeon and Li received the Academic Senate Faculty Grant for efficient deep learning model training on heterogeneous memory architectures.

  • [3/2021] Prof. Lu received the Academic Senate Faculty Grant for "Networking and Data Processing Infrastructure for Collaborative Research on Agriculture between UC Merced and Merced College.”

  • [1/2021] Our collaboration work with Microsoft on training large NLP models with heterogeneous memory draws some attentions from media (see 1 and 2).

  • [1/2021] A paper "Tahoe: Tree Structure-Aware High Performance Inference Engine for Decision Tree Ensemble on GPU" is accepted into EuroSys'21.

  • [12/2020] Shuangyan Yang joined us as a PhD student. Welcome aboard, Shuangyan!

  • [12/2020] A paper “ ArchTM: Architecture-Aware, High Performance Transaction for Persistent Memory” is accepted in FAST'21!

  • [12/2020] Welcome Yan Li from Western Digital to visit us. She gave a talk, “NAND Flash and its Application”.

  • [12/2020] Ph.D. student Shashank Gugnani has successfully defended his thesis and graduated. Congratulations, Dr. Gugnani!

  • [12/2020] A paper is accepted in IPDPS 2021: NVMe-CR: A Scalable Ephemeral Storage Runtime for Checkpoint/Restart with NVMe-over-Fabrics. Congratulations, Shashank and Tianxi!

  • [12/2020] A paper is accepted in VLDB 2021: Understanding the Idiosyncrasies of Real Persistent Memory. Congratulations, Shashank and Arjun!