The two largest barriers to adoption of FPGA platforms for HPC applications are the difficulty of programming FPGAs and the performance gap when compared to GPUs. To address the first barrier, new ecosystems like Intel oneAPI, and Xilinx Vitis HLS aim to improve programmability for FPGA platforms. From a performance aspect, FPGAs trade off lower compute frequencies for more customized hardware acceleration and power efficiency when compared to GPUs. The performance for memory-bound applications on recent GPU platforms like NVIDIA’s H100 and AMD’s MI210 has also improved due to the inclusion of high-bandwidth memories (HBM), and newer FPGA platforms are also starting to include HBM in addition to traditional DRAM.
To understand the current state-of-the-art and performance differences between FPGAs and GPUs, we consider realized memory bandwidth for recent FPGA and GPU platforms. We utilize a custom STREAM benchmark to evaluate two Intel FPGA platforms, the Stratix 10 SX PAC and Bittware 520N-MX, two AMD/Xilinx FPGA platforms, the Alveo U250 and Alveo U280, as well as GPU platforms from NVIDIA and AMD. We also extract power measurements and estimate memory bandwidth per Watt ((GB/s)/W) on these platforms to evaluate how FPGAs compare against GPU execution. While the GPUs far exceed the FPGAs in raw performance, the HBM equipped FPGAs demonstrate a competitive performance-power balance for larger data sizes that can be easily implemented with oneAPI and Vitis HLS kernels. These findings suggest a potential sweet spot for this emerging FPGA ecosystem to serve bandwidth limited applications in an energy-efficient fashion.
more »
« less
Scalable Window Generation for the Intel Broadwell+Arria 10 and High-Bandwidth FPGA Systems
Emerging FPGA systems are providing higher external memory bandwidth to compete with GPU performance. However, because FPGAs often achieve parallelism through deep pipelines, traditional FPGA design strategies do not necessarily scale well to large amounts of replicated pipelines that can take advantage of higher bandwidth. We show that sliding-window applications, an important subset of digital signal processing, demonstrate this scalability problem. We introduce a window generator architecture that enables replication to over 330 GB/s, which is an 8.7x improvement over previous work. We evaluate the window generator on the Intel Broadwell+Arria10 system for 2D convolution and show that for traditional convolution (one filter per image), our approach outperforms a 12-core Xeon Broadwell E5 by 81x and a high-end Nvidia P6000 GPU by an order of magnitude for most input sizes, while improving energy by 15.7x. For convolutional neural nets (CNNs), we show that although the GPU and Xeon typically outperform existing FPGA systems, projected performances of the window generator running on FPGAs with sufficient bandwidth can outperform high-end GPUs for many common CNN parameters.
more »
« less
- Award ID(s):
- 1738420
- PAR ID:
- 10073108
- Date Published:
- Journal Name:
- 2018 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays, FPGA ’18
- Page Range / eLocation ID:
- 173 to 182
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
FPGAs offer a heterogenous compute solution to the continuous de- sire for increased performance by enabling the creation of application- specific hardware that accelerates computation. While the barrier to entry has historically been steep, advances in High Level Synthe- sis (HLS) are making FPGAs more accessible. Specifically, the Intel FPGA OpenCL SDK allows software designers to abstract away low level details of architecting hardware on an FPGA and allows them to author computational kernels in a higher level language. Furthermore, Intel has developed a system that incorporates both a multicore Xeon CPU and Arria 10 FPGA into the same chip package as part of the Heterogeneous Accelerator Research Program (HARP) that can be targeted by their SDK. In this work, we target the second iteration of the HARP platform (HARPv2) using HLS through porting of OpenCL kernels originally written for FPGAs connected via a PCIe bus. We evaluate the HARPv2 system’s performance against previously reported results, explore the portability of kernels through a hardware design space search, and empirically show the benefits of using the shared virtual memory (SVM) abstraction over explicit reads and writes.more » « less
-
null (Ed.)Large Convolutional Neural Networks (CNNs) are often pruned and compressed to reduce the amount of parameters and memory requirement. However, the resulting irregularity in the sparse data makes it difficult for FPGA accelerators that contains systolic arrays of Multiply-and-Accumulate (MAC) units, such as Intel’s FPGA-based Deep Learning Accelerator (DLA), to achieve their maximum potential. Moreover, FPGAs with low-bandwidth off-chip memory could not satisfy the memory bandwidth requirement for sparse matrix computation. In this paper, we present 1) a sparse matrix packing technique that condenses sparse inputs and filters before feeding them into the systolic array of MAC units in the Intel DLA, and 2) a customization of the Intel DLA which allows the FPGA to efficiently utilize a high bandwidth memory (HBM2) integrated in the same package. For end-to-end inference with randomly pruned ResNet-50/MobileNet CNN models, our experiments demonstrate 2.7x/3x performance improvement compared to an FPGA with DDR4, 2.2x/2.1x speedup against a server-class Intel SkyLake CPU, and comparable performance with 1.7x/2x power efficiency gain as compared to an NVidia V100 GPU.more » « less
-
FPGAs are an excellent target for real-time video processing as they provide large amounts of low-level parallelism, low latency, and high bandwidth. However, creating real-time video processing systems on an FPGA is tedious and requires significant effort and low-level digital design skills. This paper presents a technique for creating complex real-time video processing pipelines relatively quickly and easily using partial reconfiguration (PR). A static FPGA system is created that provides a template for a variety of partially reconfigurable video processing cores. A library of video filters has been created that can be inserted into the template regions. At run-time, the user can select the topology of video cores and customize these cores to create complex and unique video pipelines without any understanding of low-level FPGA details. This paper demonstrates this technique with a library of 11 partial reconfigurable regions and 16 video processing cores operating on the Xilinx PYNQ system.more » « less
-
Over the past few years, there has been an increased interest in including FPGAs in data centers and high-performance computing clusters along with GPUs and other accelerators. As a result, it has become increasingly important to have a unified, high-level programming interface for CPUs, GPUs and FPGAs. This has led to the development of compiler toolchains to deploy OpenCL code on FPGA. However, the fundamental architectural differences between GPUs and FPGAs have led to performance portability issues: it has been shown that OpenCL code optimized for GPU does not necessarily map well to FPGA, often requiring manual optimizations to improve performance. In this paper, we explore the use of thread coarsening - a compiler technique that consolidates the work of multiple threads into a single thread - on OpenCL code running on FPGA. While this optimization has been explored on CPU and GPU, the architectural features of FPGAs and the nature of the parallelism they offer lead to different performance considerations, making an analysis of thread coarsening on FPGA worthwhile. Our evaluation, performed on our microbenchmarks and on a set of applications from open-source benchmark suites, shows that thread coarsening can yield performance benefits (up to 3-4x speedups) to OpenCL code running on FPGA at a limited resource utilization cost.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1145/3174243.3174262
