CGO 2021
Sat 27 February - Wed 3 March 2021
Tue 2 Mar 2021 13:00 - 13:10 - CGO Student Research Competition

Just-in-time (JIT) code generation techniques are increasingly gaining traction due to their ability to significantly speedup computations. Small and medium matrix multiplication has wide uses in the domains of Machine Learning (ML) and scientific simulations in high performance computing. These computations can benefit greatly from JIT compilation. Prior runtime code generation approaches have primarily focused on loop unrolling and address compression techniques that focus on optimizing the instruction flow. In contrast, we consider a unique code generation approach where data values are dynamically embedded in instructions as immediate values, effectively transforming a memory load into an immediate load. By utilizing Intel’s AVX-512 vector extensions, we show that our technique achieves geometric mean speedups of 1.12$\times$ and 1.05$\times$ over MKL and MKL JIT, the current state of the art JIT library, for 32 input channels. We intend to open source our JIT library and plan to apply our approach to produce high performance convolution libraries.

Tue 2 Mar

Displayed time zone: Eastern Time (US & Canada) change

12:30 - 14:30
CGO Student Research CompetitionMain Conference / Student Research Competition
12:30
10m
Talk
A New Memory Layout for Self-Rebalancing Trees
Student Research Competition
Paul Iannetta ENS Lyon
12:40
10m
Talk
Automatic Inspection of Program State for Debugging and Verification Purposes
Student Research Competition
José Wesley de Souza Magalhães Federal University of Minas Gerais
12:50
10m
Talk
Compiler Framework for Low Overhead Fork-Join Parallelism
Student Research Competition
13:00
10m
Talk
Data vs. Instructions: Runtime Code Generation for Convolutions
Student Research Competition
Malith Jayaweera Northeastern University
13:10
10m
Talk
Fast Structural Register Allocation
Student Research Competition
William Zhang Carnegie Mellon University, Pranav Kumar Carnegie Mellon University
13:20
10m
Talk
Fine Grained Control of Program Transformations via Strategic Rewriting in MLIR
Student Research Competition
Martin Lücke University of Edinburgh
13:30
10m
Talk
Towards an Exploration Tool for Program Optimization Using Heuristic Search Algorithms
Student Research Competition
Johannes Lenfers University of Münster
13:40
10m
Talk
When Binary Optimization Meets Static Profiling
Student Research Competition
Angelica Aparecida Moreira Universidade Federal de Minas Gerais