Samenvatting
High-performance computing with FPGAs is gaining momentum with
the advent of sophisticated High-Level Synthesis (HLS) tools. The performance of
a design is impacted by the input-output bandwidth, the code optimizations and
the resource consumption, making the performance estimation a challenge. This
paper proposes a performance model which extends the roofline model to take into
account the resource consumption and the parameters used in the HLS tools. A
strategy is developed which maximizes the performance and the resource utilization
within the area of the FPGA. The model is used to optimize the design exploration
of a class of window-based image processing application.
the advent of sophisticated High-Level Synthesis (HLS) tools. The performance of
a design is impacted by the input-output bandwidth, the code optimizations and
the resource consumption, making the performance estimation a challenge. This
paper proposes a performance model which extends the roofline model to take into
account the resource consumption and the parameters used in the HLS tools. A
strategy is developed which maximizes the performance and the resource utilization
within the area of the FPGA. The model is used to optimize the design exploration
of a class of window-based image processing application.
| Originele taal-2 | English |
|---|---|
| Titel | Parallel Computing: Accelerating Computational Science and Engineering (CSE) |
| Redacteuren | Michael Bader, Arndt Bode, Hans-Joachim Bungartz, Michael Gerndt, Gerhard R. Joubert, Frans Peters |
| Plaats van productie | Amsterdam |
| Uitgeverij | IOS Press |
| Pagina's | 523-531 |
| Aantal pagina's | 9 |
| ISBN van geprinte versie | 978-1-61499-380-3 |
| Status | Published - 1 mrt 2014 |
Publicatie series
| Naam | Advances in Parallel Computing |
|---|---|
| Volume | 25 |
| ISSN van geprinte versie | 0927-5452 |
| ISSN van elektronische versie | 1879-808X |