Nicolas Bohm Agostini PhD Proposal Review

Announcing:
PhD Proposal Review

Name:
Nicolas Bohm Agostini

Title:
Hardware/Software Codesign and Compiler Techniques for Efficient Hardware Acceleration of Dense Linear Algebra Kernels and Machine Learning Applications

Date:
4/4/2024

Time:
3:30:00 PM

Location: Zoom

Committee Members:
Prof. David Kaeli (Advisor)
Prof. Gunar Schirner
Prof. José Luis Abellán (University of Murcia)
Antonino Tumeo (PNNL)

Abstract:
Today’s linear algebra and machine learning applications (ML) continue to grow in size and complexity, placing rapidly increasing demands on the underlying hardware and software systems. To address these issues, hardware designers have proposed using custom accelerators explicitly designed for accelerating these demanding workloads. What needs to be improved is the ability to perform efficient hardware/software (HW/SW) co-design in order to reap the full benefits from these platforms. This thesis presents an integrated solution to facilitate HW/SW accelerator design. We also address issues in accelerator deployment, enabling rapid prototyping, integrated benchmarking, and comprehensive performance analysis of custom accelerators.

In this thesis, we demonstrate the value of a lightweight system modeling library integrated into the build/execution environment, leveraging TensorFlow~Lite for deployment. We also explore efficient design space exploration of different classes of accelerators while considering the impact of parameters. Secondly, we employ the Multi-Level Intermediate Representation (MLIR) compiler framework to automatically partition host code from accelerator code, pre-optimizing the latter for improved high-level synthesis designs and high-quality accelerated kernels. Lastly, we propose compiler extensions to automate the generation and optimization of communication between the host CPU and AXI-based accelerators.

We present novel solutions that enable more efficient and effective design space exploration, optimization, and deployment of custom accelerators. The utility of these approaches is demonstrated through experiments with specific accelerator designs and key linear algebra and ML workloads. Most importantly, these solutions empower high-level language users, such as domain scientists, to participate in the design of new accelerator features.