Sardar Usman
Papers
4
Total Citations
56
H-Index
3
About
Sardar Usman is a leading researcher in high-performance computing, specializing in the optimization of sparse matrix-vector multiplication (SpMV)—a critical kernel in scientific and engineering applications. His work focuses on leveraging machine learning to automate performance tuning for SpMV computations across distributed and shared memory architectures. Usman’s major contributions include the development of the ZAKI and ZAKI+ tools, which use machine learning to intelligently map SpMV computations onto distributed memory systems, achieving significant performance gains. His AAQAL tool extends this approach to Block CSR formats, while Elegante optimizes thread configurations on shared memory systems. Collectively, his papers have garnered over 56 citations, reflecting their impact on the field. Usman’s research is particularly relevant to cyber-physical systems and smart city applications, where efficient sparse linear algebra is essential. His work stands out for its practical, tool-oriented approach, providing ready-to-use solutions for one of computing’s “seven dwarfs.”
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4