Uploaded October 2026 | Updated November 2026, 1 day ago
SPEAKER:
Werner Bachhuber, Siemens EDA
This session was a part of the Hardware track at the RISC-V Developer Workshops held at RISC-V Summit Europe 2026 in Bologna, Italy.
WORKSHOP DETAILS:
RISC-V's open architecture enables the creation of processors highly tailored for specific computational challenges. While this can be done through instruction extensions, even greater levels of performance and efficiency can be achieved by offloading larger computations to co-processors or bus-based accelerators. But designing and implementing a co-processor can be a daunting task. This hands-on session will introduce you to High-Level Synthesis (HLS), a technology that converts a C++ function into a hardware implementation, suitable for deployment in ASIC or FPGA. HLS simplifies and accelerates the hardware design and verification process.
In this session, you’ll use High-Level Synthesis to create a RISC-V co-processor and integrate it into a RISC-V Rocket Core design. You’ll start with a Python implementation of a “Wake Word” algorithm. Profiling it, you’ll discover the bottlenecks in the algorithm, specifically, the inferencing algorithm. Using High-Level Synthesis, you’ll migrate the computational “heavy lifting” into a Verilog implementation of an inferencing accelerator where increased parallelism and customization will deliver greater performance while reducing energy consumption. Then, you’ll integrate the co-processor into the Rocket Core design, at both the hardware and software levels. Finally, you’ll run the new processing platform in simulation to compare power, performance, and area against the pure software implementation.
RESOURCES: drive.google.com/drive/u/0/folders/1YkiJnJctTHGlzSGidRkE1F3qhk1naE38
SPEAKER:
Werner Bachhuber, Siemens EDA
This session was a part of the Hardware track at the RISC-V Developer Workshops held at RISC-V Summit Europe 2026 in Bologna, Italy.
WORKSHOP DETAILS:
RISC-V's open architecture enables the creation of processors highly tailored for specific computational challenges. While this can be done through instruction extensions, even greater levels of performance and efficiency can be achieved by offloading larger computations to co-processors or bus-based accelerators. But designing and implementing a co-processor can be a daunting task. This hands-on session will introduce you to High-Level Synthesis (HLS), a technology that converts a C++ function into a hardware implementation, suitable for deployment in ASIC or FPGA. HLS simplifies and accelerates the hardware design and verification process.
In this session, you’ll use High-Level Synthesis to create a RISC-V co-processor and integrate it into a RISC-V Rocket Core design. You’ll start with a Python implementation of a “Wake Word” algorithm. Profiling it, you’ll discover the bottlenecks in the algorithm, specifically, the inferencing algorithm. Using High-Level Synthesis, you’ll migrate the computational “heavy lifting” into a Verilog implementation of an inferencing accelerator where increased parallelism and customization will deliver greater performance while reducing energy consumption. Then, you’ll integrate the co-processor into the Rocket Core design, at both the hardware and software levels. Finally, you’ll run the new processing platform in simulation to compare power, performance, and area against the pure software implementation.
RESOURCES: drive.google.com/drive/u/0/folders/1YkiJnJctTHGlzSGidRkE1F3qhk1naE38










