|
Congratulations to doctoral researcher Yuxuan Wang, who was the main author of a paper that made the Best Paper Final at ARC25, the 21st International Symposium on Applied Reconfigurable Computing. This was a great achievement at what was her first conference as a PhD researcher! |
|
Congratulations to all the authors of this fascinating paper:
Yuxuan Wang, Cristian Tirelli, Lara Orlandic, Juan Sapriza, Rubén Rodríguez Álvarez, Giovanni Ansaloni, Laura Pozzi, David Atienza
| An MLIR-based Compilation Framework for CGRA Application Deployment The 21st International Symposium on Applied Reconfigurable Computing |
 |
 |
|
Coarse-Grained Reconfigurable Arrays (CGRAs) are a programmable architectural solution, one that that efficiently supports the execution of computation-intensive application functions (kernels).
To enhance the code coverage of CGRA compilers, the end-to-end compilation framework presented in this paper operates at the kernel level, considering both Data Flow Graph (DFG)-based and Control Flow Graph (CFG)-based optimizations.
Its implementation, based on the MLIR infrastructure, includes a front-end for Intermediate Representation (IR) abstraction, an optional middle-end reshaping the IR to facilitate modulo scheduling, a kernel mapper for operation allocation that supports both DFG and CFG mapping optimizations, and a back-end that performs register allocation and generates assembly with the mapped solution. The entire framework pipeline is fully automated and is released as open-source software.
Read more about our work on CGRAs: