Josep Torrellas
University of Illinois Urbana-Champaign, IL·Computer Science
Modern computing problems are growing faster than the computers that solve them. This research makes computers work better across many layers at once: algorithms, software, compilers and the processor hardware itself.
- Fields
Computer Science
Engineering
- Topics
computer architecture
non-volatile memory
persistent memory
- Grants
- 2 known grants, through Jul 2028
What they work on
- Develop approximate algorithms for graph clustering in static and dynamic settings, graph construction that preserves key information, and machine learning, for computational biology, social network and software verification problems.
- Build a graph programming framework with a domain-specific language, scalable sparse numerical libraries and an ML-guided compiler with two intermediate representations, generating efficient code for datacenter-scale machines.
- Design a hardware accelerator with a high-level instruction set for tiled sparse matrix operations, plus an auto-tuner that maps graph code across accelerators and general-purpose processors.
- Create an easy-to-use, generic programming framework for non-volatile memory that requires minimal programmer involvement.
- Redesign processor and memory-hierarchy hardware primitives to reduce the cost of persistency operations, and develop distributed data persistency models for a distributed runtime system.
More about the research
One project speeds up huge graph problems (graphs are networks of connected points) on datacenter-scale machines. The other makes non-volatile memory practical. Non-volatile memory is a type of memory that keeps its data when power is off.
Why it matters
Faster, more energy-efficient graph computing could let researchers tackle larger problems in evolutionary biology, social network analysis and software verification. Usable non-volatile memory could make applications faster and easier to write. Both projects aim to help chip makers and datacenter operators, through planned work with industry partners such as Microsoft and Intel.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Nathaniel Bleier
University of Michigan, MI·Computer Science
Combine learning-guided program synthesis with automated logical reasoning to generate instruction sequences, accelerator control codes and microcoded expansions that are formally proven equivalent to the original program.
Shaahin Angizi
New Jersey Institute of Technology, NJ·Engineering·Early-career
Design reconfigurable circuits and architectures for a multi-modal in-sensor accelerator that moves AI computation onto sensor nodes as efficiently as possible.
Minxuan Zhou
Illinois Institute of Technology, IL·Computer Science
Design a domain-specific programming language that makes it easier to write general fully homomorphic encryption (FHE) applications across different algorithmic schemes.
The school
University of Illinois Urbana-Champaign: 41 labs on PI Pages, $907M a year on research, 870 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.