High-Level FPGA Accelerator Design for Structured-Mesh-Based Explicit Numerical Solvers

Kamalakkannan, Kamalavasan; Mudalige, Gihan R.; Reguly, István Z.; Fahmy, Suhaib A.

doi:10.1109/ipdps49936.2021.00117

Cited by 4 publications

(20 citation statements)

References 27 publications

(30 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Rannacher smoothing available in the original application has been switched off in our evaluation. The hv_pred* and hv_matrices are explicit loops each using 10 point stencils, requiring a window buffer implementation [13] for data reuse. The 9 kernels in Algo.…”

Section: Stochastic Local Volatilitymentioning

confidence: 99%

“…In comparison, the HLS-based synthesis presented in this paper, targets the solution of multiple tridiagonal systems and in multiple dimensions as commonly found in real-world applications. It uses the Thomas algorithm demonstrating that together with techniques such as batching of systems [13], high throughput for small and medium sized systems can be achieved. The Thomas algorithm uses fewer resources than the more computationally intensive PCR algorithm.…”

Section: Related Workmentioning

confidence: 99%

“…This enables FPGA resource utilization to be maximized by implementing 6 CUs each having 8 Thomas solvers synthesized as a vectorized solver. The RHS calculation, which is a 3D explicit stencil loop was implemented using techniques similar to those in [13], as a separate module. The intermediate results between CUs and RHS module were written/read to/from external memory.…”

mentioning

confidence: 99%

See 2 more Smart Citations

High throughput multidimensional tridiagonal system solvers on FPGAs

Kamalakkannan

Mudalige

Reguly

et al. 2022

Proceedings of the 36th ACM International Conference on Supercomputing

Self Cite

View full text Add to dashboard Cite

We present a high performance tridiagonal solver library for Xilinx FPGAs optimized for multiple multi-dimensional systems common in real-world applications. An analytical performance model is developed and used to explore the design space and obtain rapid performance estimates that are over 85% accurate. This library achieves an order of magnitude better performance when solving large batches of systems than previous FPGA work. A detailed comparison with a current state-of-the-art GPU library for multidimensional tridiagonal systems on an Nvidia V100 GPU shows the FPGA achieving competitive or better runtime and significant energy savings of over 30%. Through this design, we learn lessons about the types of applications where FPGAs can challenge the current dominance of GPUs. CCS CONCEPTS• Computer systems organization → Reconfigurable computing; Multicore architectures; • Mathematics of computing → Mathematical software performance.

show abstract

Section: Stochastic Local Volatilitymentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

High throughput multidimensional tridiagonal system solvers on FPGAs

Kamalakkannan

Mudalige

Reguly

et al. 2022

Proceedings of the 36th ACM International Conference on Supercomputing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Field Programmable Gate Arrays (FPGAs) have gained traction as accelerator devices, particularly due to their low power consumption and re-programmability. Their utility in accelerating scientific computing workloads have recently been gaining interest in the high performance computing (HPC) community with several studies showing competitive or better performance compared to traditional CPU/GPU architectures, for select classes of applications [9,13,19,27].…”

Section: Introductionmentioning

confidence: 99%

“…(1) Based on design strategies developed in previous work [19,20], in this paper we develop a generalized workflow for formulating optimized FPGA designs with SYCL, targeting Intel FPGAs for two classes of structured-mesh based solvers: (1) stencil applications based on explicit numerical methods and…”

Section: Introductionmentioning

confidence: 99%

FPGA Acceleration of Structured-Mesh-Based Explicit and Implicit Numerical Solvers using SYCL

Kamalakkannan

Mudalige

Reguly

et al. 2022

International Workshop on OpenCL

Self Cite

View full text Add to dashboard Cite

We explore the design and development of structured-mesh based solvers on current Intel FPGA hardware using the SYCL programming model. Two classes of applications are targeted : (1) stencil applications based on explicit numerical methods and (2) multidimensional tridiagonal solvers based on implicit methods. Both classes of solvers appear as core modules in a wide-range of realworld applications ranging from CFD to financial computing. A general, unified workflow is formulated for synthesizing them on Intel FPGAs together with predictive analytic models to explore the design space to obtain near-optimal performance. Performance of synthesized designs, using the above techniques, for two non-trivial applications on an Intel PAC D5005 FPGA card is benchmarked. Results are compared to performance of optimized parallel implementations of the same applications on a Nvidia V100 GPU. Observed runtime results indicate the FPGA providing better or matching performance to the V100 GPU. However, more importantly the FPGA solutions provide 59%-76% less energy consumption for their largest configurations, making them highly attractive for solving workloads based on these applications in production settings. The performance model predicts the runtime of designs with high accuracy with less than 5% error for all cases tested, demonstrating their significant utility for design space explorations. With these tools and techniques, we discuss determinants for a given structuredmesh code to be amenable to FPGA implementation, providing insights into the feasibility and profitability of a design, how they can be codified using SYCL and the resulting performance.

show abstract

FA-LAMP: FPGA-Accelerated Learned Approximate Matrix Profile for Time Series Similarity Prediction

Kalantar

Zimmerman

Brisk

2021

2021 IEEE 29th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM)

View full text Add to dashboard Cite

High-Level FPGA Accelerator Design for Structured-Mesh-Based Explicit Numerical Solvers

Abstract: Please refer to published version for the most recent bibliographic citation information. If a published version is known of, the repository item page linked to above, will contain details on accessing it.

Cited by 4 publications

References 27 publications

High throughput multidimensional tridiagonal system solvers on FPGAs

High throughput multidimensional tridiagonal system solvers on FPGAs

FPGA Acceleration of Structured-Mesh-Based Explicit and Implicit Numerical Solvers using SYCL

FA-LAMP: FPGA-Accelerated Learned Approximate Matrix Profile for Time Series Similarity Prediction

Contact Info

Product

Resources

About