TensorWorld studies sparse and structured computation across HPC, systems, and AI.

CROSS

Cross-layer coordination and optimization for scalable and sparse tensor networks. TensorWorld leads NSF-supported work on abstractions, systems, and performance portability for tensor network workloads.

NSF PPoSS LARGE, 2023-2028

Sparse Tensor Systems

We design algorithms, data structures, and runtime support for sparse tensor contraction, decomposition, and related kernels on CPUs, GPUs, and emerging architectures.

Performance Profiling

Our group builds profiling and analysis tools for modern ML workloads, with an emphasis on data types, memory behavior, GPU execution, and framework-level performance bottlenecks.

Irregular Applications

We study performance optimization for irregular algorithms in graph, hypergraph, tensor, and scientific computing workloads, including auto-tuning and architecture-aware implementation strategies.

STTID

STTID: High-Performance Sparse Tensor-Train Interpolative Decomposition.

RedSan

RedSan: A Redundant Memory Instruction Sanitizer for GPU Programs.

SymProp

SymProp: Scaling Sparse Symmetric Tucker Decomposition via Symmetry Propagation.

HiParTI

A hierarchical parallel tensor infrastructure for sparse and structured tensor computing.

PASTA

A parallel sparse tensor algorithm benchmark suite spanning CPUs and GPUs.

ParTI

A parallel tensor infrastructure for data analysis and tensor kernels.

AdaTM

Adaptive tensor memoization support for CP decomposition.

InTensLi

Input-adaptive and in-place dense tensor-times-matrix multiplication.

SMAT

A sparse matrix-vector multiplication auto-tuner.