Charles (Chase) Block

I am a PhD candidate working with the i-acoma group in the Computer Science department at the University of Illinois Urbana-Champaign. My interests are mainly in computer architecture and hardware/software systems development, although I have worked on a variety of other projects in the past. Previously, I studied Electrical & Computer Engineering at the University of Texas at Austin.

In the past, I have worked on a range of topics and with several groups, including firmware development for Amazon’s Scout, databases at Microsoft Azure, CPU verification with Intel’s Advanced Architecture Development Group, automotive electronics with the University of Texas Solar Vehicles Team (a part of Longhorn Racing), and bioelectronics with UT Austin’s Lu Group.

You can see my full CV here.

news

Jun 28, 2024	Our work, titled Taming the Acceleration Tax: Enabling New Opportunities for Fine-Grained, Disaggregated Accelerator-Level Parallelism was just accepted to PACT’24!
Jun 14, 2024	Our work titled Distributed-Memory Parallel Algorithms for Sparse Matrix and Sparse Tall-and-Skinny Matrix Multiplication, led by Isuru Ranawaka, has been accepted to SC’24!
May 01, 2024	The Lu Group at UT Austin just published a work on stretchable pressure sensors in Matter! You can find it here.
Apr 04, 2024	I’m excited to be presenting our work on communication in distributed sparse matrix multiplication, called Two-Face, at ASPLOS in San Diego this month! See you there!

selected publications

ASPLOS
Two-Face: Combining Collective and One-Sided Communication for Efficient Distributed SpMM

Charles Block, Gerasimos Gerogiannis, Charith Mendis, Ariful Azad, and Josep Torrellas

In Proceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2, La Jolla, CA, USA, Apr 2024

Abs Bib PDF

Sparse matrix dense matrix multiplication (SpMM) is commonly used in applications ranging from scientific computing to graph neural networks. Typically, when SpMM is executed in a distributed platform, communication costs dominate. Such costs depend on how communication is scheduled. If it is scheduled in a sparsity-unaware manner, such as with collectives, execution is often inefficient due to unnecessary data transfers. On the other hand, if communication is scheduled in a fine-grained sparsity-aware manner, communicating only the necessary data, execution can also be inefficient due to high software overhead.We observe that individual sparse matrices often contain regions that are denser and regions that are sparser. Based on this observation, we develop a model that partitions communication into sparsity-unaware and sparsity-aware components. Leveraging the partition, we develop a new algorithm that performs collective communication for the denser regions, and fine-grained, one-sided communication for the sparser regions. We call the algorithm Two-Face. We show that Two-Face attains an average speedup of 2.11x over prior work when evaluated on a 4096-core supercomputer. Additionally, Two-Face scales well with the machine size.
@inproceedings{block24twoface, author = {Block, Charles and Gerogiannis, Gerasimos and Mendis, Charith and Azad, Ariful and Torrellas, Josep}, title = {Two-Face: Combining Collective and One-Sided Communication for Efficient Distributed SpMM}, year = {2024}, month = apr, isbn = {9798400703850}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3620665.3640427}, doi = {10.1145/3620665.3640427}, booktitle = {Proceedings of the 29th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2}, pages = {1200–1217}, numpages = {18}, keywords = {high-performance computing, distributed algorithms, sparse matrices, SpMM}, location = {La Jolla, CA, USA}, series = {ASPLOS '24} }