Introduction to Special Section on FCCM 2019

The International Symposium on Field-Programmable Custom Computing Machines (FCCM) is the original and premier forum for presenting and discussing new research related to computing that exploits the unique features and capabilities of field-programmable gate arrays (FPGAs) and other reconfigurable hardware. Since its first meeting in 1993, FCCM has been the place to present the latest work in architectures, tools, and programming models for field-programmable custom computing machines as well as applications that use such systems. The 27th FCCM met April 28 to May 1, 2019 on the campus of UCSD in San Diego, California. In this special ACM TRETS section, it is our pleasure to present extended versions of a selected set of 2 of the 31 full and 7 short papers accepted and presented at FCCM 2019. With the end of processor frequency scaling and the end of Dennard Scaling, the broader architecture community is starting to acknowledge the role of “accelerators” and applicationcustomized parallel compute engines. FCCM has highlighted and nurtured the body-of-knowledge and tools for these application-customized accelerators throughout its history. In this special section, we present two article that provide tools to address the growing challenges in large-scale FPGAs and computing systems. Both come with open-source releases for their code so they also contribute to the technical infrastructure and building blocks that allow our community to collaboratively advance the field. One of the common complaints holding back FPGA computing is the slow runtime for the basic mapping tools. Compared to processors and GPUs, the hour-long runtimes for FPGA mapping serve to discourage potential adopters. As FPGAs grow in capacity with Moore’s Law, FPGA mapping times have continued to grow. Routing is often one of the large components of FPGA