Liang, Xin

Introduction

The mission of Performance Optimizations for LARge-scale cyberInfrastructure and Systems (Polaris) Laboratory led by Dr. Xin Liang in the department of Computer Science is to develop high-performance and scalable middleware for large-scale data management, storage, and analytics on cutting-edge computing systems, which will advance multiple domains including scientific simulations, big data analytics, artificial intelligence, and quantum computing. Driven by the actual needs from various science applications, research in the Polaris lab aims to address the performance issues raised by the big data challenges in real-world applications via deep collaborations with computational scientists. The research products are widely used in multiple DOE laboratories and will be validated on current and next-generation supercomputers including Summit, Aurora, and Frontier.

Error-controlled lossy compression for scientific data at scale

This project aims to deliver efficient error-controlled lossy compressors to address the data challenges for a wide range of scientific applications. Research in this project includes design and implementation of novel compression algorithm, performance improvement with accelerators, and I/O studies on parallel file systems. 

Participant

PI - Xin Liang

Graduate Students - Pu Jiao, Mingze Xia, Added on MCC cluster, 09/07/2023 

Xuan Wu, Added on MCC cluster, 07/27/2023 

Software:

SZ (https://github.com/szcompressor)

ZFP (https://github.com/LLNL/zfp)

MGARD (https://github.com/CODARcode/MGARD)

TTHRESH (https://github.com/rballester/tthresh)

IDX2 (https://github.com/sci-visus/idx2)

Collaborators:

Dr. Sheng Di (ANL)

Dr. Franck Cappello (ANL)

Dr. Scott Klasky (ORNL)

Dr. Hanqi Guo (OSU) 

Exascale data management 

This project studies how to performance efficient data management in a scientific workflow on large-scale clusters. Research includes data staging, in-situ data visualization and analytics, and data retrieval and placement on hierarchical storage systems.

Participant

PI - Xin Liang

Graduate Students - Pu Jiao, Mingze Xia, Added on 09/08/2022 on MCC cluster

Software:

ADIOS2 (https://github.com/ornladios/ADIOS2)

HDF5 (https://github.com/HDFGroup/hdf5)

Hermes (https://github.com/HDFGroup/hermes)

Collaborators:

Dr. Sheng Di (ANL)

Dr. Franck Cappello (ANL)

Dr. Scott Klasky (ORNL)

Dr. Hanqi Guo (OSU) 

Scalable algorithms for quantum circuit compilation

This project studies how to leverage hybrid systems to perform scalable quantum circuit compilation. Research includes designing mapping and routing algorithms, parallel implementation on distributed systems, and scheduling and workflows on hybrid systems.

Participant

PI - Xin Liang

Software:

Qiskit (https://github.com/Qiskit)

HPX (https://github.com/STEllAR-GROUP/hpx)

Collaborators:

Dr. Avah Banerjee (Missouri S&T)

Dr. Rod Tohid (LSU)

Dr. Qiang Guan (KSU)

Grants:

Publications: