Xiaoming Li

Associate Professor
Department of Electrical and Computer Engineering

201D Evans Hall
University of Delaware
Newark, DE 19716

Phone: 302-831-0334
Fax: 302-831-4316
E-mail: xli AT ece.udel.edu


No matter. Try again. Fail again. Fail better.
- Samuel Becket

Research Interest

My research interests are code generation and optimization, compilers, and interaction between hardware and software. The main goal of my research is to make programs run faster and use less resource. I am interested in all aspect of compiler optimizations and transformations, both statically and dynamically. In particular, I am interested in code generation techniques that combine machine learning and architectural models to automatically deliver higher quality code. I am also interested in compilers for non-traditional executing engines, such as SAT solvers and virtual machines.

Representative Publications

(full list of publications)

"G-Code Re-compilation and Optimization for Faster 3D Printing". Xiaoming Li. The 33rd Workshop on Languages and Compilers for Parallel Computing (LCPC), Stony Brook, NY, USA. October, 2020. [bib] 

					@incollection{li2020gcode,
						year={2020},
						booktitle={Languages and Compilers for Parallel Computing (LCPC)},
						title={G-Code Re-compilation and Optimization for Faster 3D Printing},
						author={Li, Xiaoming},
						month={October},
						location = {Stony Brook, NY, USA},
					}

"Fast Convolutional Neural Networks with Fine-Grained FFTs". Yulin Zhang and Xiaoming Li. The 29th ACM International Conference on Parallel Architectures and Compilation Techniques (PACT). Atlanta, GA, USA. October, 2020. [bib] 

					@inproceedings{zhang2020cnn,
						author = {Zhang, Yulin and Li, Xiaoming},
						title = {Fast Convolutional Neural Networks with Fine-Grained FFTs},
						year = {2020},
						isbn = {9781450380751},
						publisher = {Association for Computing Machinery},
						address = {New York, NY, USA},
						url = {https://doi.org/10.1145/3410463.3414642},
						doi = {10.1145/3410463.3414642},
						booktitle = {Proceedings of the ACM International Conference on Parallel Architectures and Compilation Techniques (PACT)},
						pages = {255–265},
						numpages = {11},
						keywords = {gpu, fast fourier transform, convolutional neural network, algorithm optimization},
						location = {Virtual Event, GA, USA},
						series = {PACT '20}
					}

"Input-adaptive Parallel Sparse Fast Fourier Transform for Stream Processing". Shuo Chen and Xiaoming Li. The 28th International Conference on Supercomputing (ICS 2014), Munich, Germany, June, 2014. [bib] 

					@inproceedings{chen2014input,
						author = {Chen, Shuo and Li, Xiaoming},
						title = {Input-adaptive Parallel Sparse Fast Fourier Transform for Stream Processing},
						booktitle = {Proceedings of the 28th ACM International Conference on Supercomputing},
						series = {ICS '14},
						year = {2014},
						isbn = {978-1-4503-2642-1},
						location = {Munich, Germany},
						pages = {93--102},
						numpages = {10},
						url = {http://doi.acm.org/10.1145/2597652.2597669},
						doi = {10.1145/2597652.2597669},
						acmid = {2597669},
						publisher = {ACM},
						address = {New York, NY, USA},
						keywords = {input adaptive, parallel algorithm, sparse fft, stream processing},
					}

"A Hybrid GPU/CPU FFT Library for Large FFT Problems". Shuo Chen and Xiaoming Li. Proceedings of The 32nd IEEE International Performance Computing and Communications Conference, IPCCC 2013, San Diego, CA, USA. [bib] 

					@INPROCEEDINGS{chen2013hybrid,
					 author={Shuo Chen and Xiaoming Li},
					 booktitle={Performance Computing and Communications Conference (IPCCC), 2013 IEEE 32nd International},
					 title={A hybrid GPU/CPU FFT library for large FFT problems},
					 year={2013},
					 month={Dec},
					 pages={1-10},
					 keywords={fast Fourier transforms;graphics processing units;peripheral interfaces;3D FFT;CPU FFT library;CPU computing power;FFT computation;FFT decomposition paradigms;GPU memory;GPU performance;GeForce GTX480;Intel MKL;PCI channel;Tesla C2070;Tesla C2075;current GPU based FFT implementation;data transfer;empirical tuning process;flexible partitioning;graphic processing units;hybrid GPU;hybrid parallel framework;large FFT problems;large size fast Fourier transform;limited memory size;memory-transfer controller;multicore CPU;optimal load balancing;performance modeling;tailor computation;Graphics processing units},
					 doi={10.1109/PCCC.2013.6742796},}

"Using GPUs to Compute Large Out-of-card FFTs", Liang Gu, Jakob Siegel and Xiaoming Li. 25th International Conference on Supercomputing (ICS 2011), Tucson, Arizona, USA, June, 2011. [bib] 

					@inproceedings{gu2011using,
					 author = {Gu, Liang and Siegel, Jakob and Li, Xiaoming},
					 title = {Using GPUs to Compute Large Out-of-card FFTs},
					 booktitle = {Proceedings of the International Conference on Supercomputing},
					 series = {ICS '11},
					 year = {2011},
					 isbn = {978-1-4503-0102-2},
					 location = {Tucson, Arizona, USA},
					 pages = {255--264},
					 numpages = {10},
					 url = {http://doi.acm.org/10.1145/1995896.1995937},
					 doi = {10.1145/1995896.1995937},
					 acmid = {1995937},
					 publisher = {ACM},
					 address = {New York, NY, USA},
					 keywords = {fft dft library gpu cuda},
					}

"An Empirically Tuned 2D and 3D FFT Library on CUDA GPU", Liang Gu, Xiaoming Li and Jakob Siegel. Proceedings of International Conference on Supercomputing (ICS 2010). Tsukuba, Japan. June, 2010. [bib] 

					@inproceedings{gu2010empirically,
					 author = {Gu, Liang and Li, Xiaoming and Siegel, Jakob},
					 title = {An Empirically Tuned 2D and 3D FFT Library on CUDA GPU},
					 booktitle = {Proceedings of the 24th ACM International Conference on Supercomputing},
					 series = {ICS '10},
					 year = {2010},
					 isbn = {978-1-4503-0018-6},
					 location = {Tsukuba, Ibaraki, Japan},
					 pages = {305--314},
					 numpages = {10},
					 url = {http://doi.acm.org/10.1145/1810085.1810127},
					 doi = {10.1145/1810085.1810127},
					 acmid = {1810127},
					 publisher = {ACM},
					 address = {New York, NY, USA},
					 keywords = {2D FFT, 3D FFT, CUDA, GPU, empirical tuning, library generation},
					}

"DFT Performance Prediction in FFTW", Liang Gu and Xiaoming Li. Proceedings of Languages and Compilers for Parallel Computing, 22nd International Workshop, (LCPC 2009), Newark, Delaware, USA, October, 2009. [bib] 

					@incollection{gu2010dft,
					year={2010},
					isbn={978-3-642-13373-2},
					booktitle={Languages and Compilers for Parallel Computing},
					volume={5898},
					series={Lecture Notes in Computer Science},
					editor={Gao, GuangR. and Pollock, LoriL. and Cavazos, John and Li, Xiaoming},
					doi={10.1007/978-3-642-13374-9_10},
					title={DFT Performance Prediction in FFTW},
					url={http://dx.doi.org/10.1007/978-3-642-13374-9_10},
					publisher={Springer Berlin Heidelberg},
					author={Gu, Liang and Li, Xiaoming},
					pages={140-156}
					}

"A Control-structure Splitting Optimization for GPGPU", Snaider Carillo, Jakob Siegel and Xiaoming Li, Proceedings of ACM International Conference on Computing Frontier (CF) 2009, March, 2009. [bib] 

					@inproceedings{carrillo2009control,
					 author = {Carrillo, Snaider and Siegel, Jakob and Li, Xiaoming},
					 title = {A Control-structure Splitting Optimization for GPGPU},
					 booktitle = {Proceedings of the 6th ACM Conference on Computing Frontiers},
					 series = {CF '09},
					 year = {2009},
					 isbn = {978-1-60558-413-3},
					 location = {Ischia, Italy},
					 pages = {147--150},
					 numpages = {4},
					 url = {http://doi.acm.org/10.1145/1531743.1531766},
					 doi = {10.1145/1531743.1531766},
					 acmid = {1531766},
					 publisher = {ACM},
					 address = {New York, NY, USA},
					 keywords = {cuda, gpgpu, optimizations},
					}

"Optimizing Sorting with Genetic Algorithm". Xiaoming Li, María Jesús Garzarán, and David Padua.  In Proc. of the 3rd International Symposium on Code Generation and Optimization (CGO-2005), pages 99-110, San Jose, CA, USA, 2005. [bib] 

				@inproceedings{li2005optimizing,
				author = {Li, Xiaoming and Garzaran, Maria Jesus and Padua, David},
				title = {Optimizing Sorting with Genetic Algorithms},
				booktitle = {Proceedings of the International Symposium on Code Generation and Optimization},
				series = {CGO '05},
				year = {2005},
				isbn = {0-7695-2298-X},
				pages = {99--110},
				numpages = {12},
				url = {http://dx.doi.org/10.1109/CGO.2005.24},
				doi = {10.1109/CGO.2005.24},
				acmid = {1048979},
				publisher = {IEEE Computer Society},
				address = {Washington, DC, USA},
				}

"Is Search Really Necessary to Generate High-Performance BLAS?". Kamen Yotov, Xiaoming Li, Gang Ren, Maria Garzaran, David Padua, Keshav Pingali and Paul Stodghill. Proceedings of the IEEE Special Issue on Program Generation, Optimization, and Platform Adaptation, Vol. 93, No. 2, pages 358-386, February, 2005. [bib] 

				@ARTICLE{yotovli2005search,
				 author={Yotov, Kamen and Xiaoming Li and Ren, Gang and Garzaran, Maria Jesus and Padua, David and Pingali, Keshav and Stodghill, Paul},
				 journal={Proceedings of the IEEE},
				 title={Is Search Really Necessary to Generate High-Performance BLAS?},
				 year={2005},
				 month={Feb},
				 volume={93},
				 number={2},
				 pages={358-386},
				 keywords={linear algebra;mathematics computing;optimising compilers;software libraries;ATLAS;BLAS;Basic Linear Algebra Subprograms;code generation;global search engine;library generators;model driven optimization engine;program compilers;program optimization;search based empirical optimization;Analytical models;Computer science;Hardware;Libraries;Linear algebra;Optimizing compilers;Program processors;Programming profession;Search engines;Tiles;Basic Linear Algebra Subprograms (BLAS);compilers;empirical optimization;high-performance computing;library generators;model-driven optimization;program optimization},
				 doi={10.1109/JPROC.2004.840444},
				 ISSN={0018-9219},}

"A Dynamically Tuned Sorting Library". Xiaoming Li, María Jesús Garzarán, and David Padua. In Proc. of the International Symposium on Code Generation and Optimization (CGO-2004), pages 111-124, March 2004. [bib] 

				@inproceedings{li2004dynamically,
				 author = {Li, Xiaoming and Garzar\'{a}n, Mar\'{\i}a Jes\'{u}s and Padua, David},
				 title = {A Dynamically Tuned Sorting Library},
				 booktitle = {Proceedings of the International Symposium on Code Generation and Optimization: Feedback-directed and Runtime Optimization},
				 series = {CGO '04},
				 year = {2004},
				 isbn = {0-7695-2102-9},
				 location = {Palo Alto, California},
				 pages = {111--},
				 url = {http://dl.acm.org/citation.cfm?id=977395.977663},
				 acmid = {977663},
				 publisher = {IEEE Computer Society},
				 address = {Washington, DC, USA},
				}

"A Comparison of Empirical and Model-driven Optimization". Kamen Yotov, Xiaoming Li, Gang Ren, Michael Cibulskis, Gerald DeJong, María Jesús Garzarán, David Padua, Keshav Pingali, Paul Stodghill, and Peng Wu. In Proc. of the International Conference on Programming Language Design and Implementation (PLDI 2003), pages 63-76, June 2003. [bib] 

				@inproceedings{yotovli2003comparison,
				 author = {Yotov, Kamen and Li, Xiaoming and Ren, Gang and Cibulskis, Michael and DeJong, Gerald and Garzaran, Maria and Padua, David and Pingali, Keshav and Stodghill, Paul and Wu, Peng},
				 title = {A Comparison of Empirical and Model-driven Optimization},
				 booktitle = {Proceedings of the ACM SIGPLAN 2003 Conference on Programming Language Design and Implementation},
				 series = {PLDI '03},
				 year = {2003},
				 isbn = {1-58113-662-5},
				 location = {San Diego, California, USA},
				 pages = {63--76},
				 numpages = {14},
				 url = {http://doi.acm.org/10.1145/781131.781140},
				 doi = {10.1145/781131.781140},
				 acmid = {781140},
				 publisher = {ACM},
				 address = {New York, NY, USA},
				 keywords = {BLAS, blocking, code generation, compilers, empirical optimization, memory hierarchy, model-driven optimization, program transformation, tiling, unrolling},
				}

Research Group

Current students:

Ryan Taylor
Sha Li
Shuo Chen
Yuanfang Chen
Chao Yang
Chaoyu Chen

Graduates:

Liang Gu. Ph.D. thesis: "FFT Libraries on Heterogeneous Computers". First employment: Bloomberg.

Jakob Siegel. Ph.D. thesis: "A Tasking Framework to Hanedle Loadbalancing and Node Level Optimizations on Heterogeneous GPGPU Clusters". First employment: Ion Torrent.

Murat Bolat. Ph.D. thesis: "Context-award Compilation". First employment: ARM.

Contact Information


Mail: Department of Electrical and Computer Engineering
University of Delaware
140 Evans Hall
Newark, DE 19716

Office: 201D Evans Hall

E-Mail: xli@udel.edu
Phone: 302-831-0334
Fax: 302-831-4316

Bio

Xiaoming Li obtained his B.S. and M.E. in Computer Science at Nanjing University in 1998 and 2001, and Ph.D. in Computer Science at the University of Illinois at Urbana-Champaign in 2006. He is currently an Assistant Professor of Computer Engineering at the University of Delaware. He is a recepient of the National Science Foundation CAREER award (2008).