white paper

High-Level Synthesis, it’s still hardware design

High-Level Synthesis, It’s Still Hardware Design: 3d illustration of futuristic microchip city.

This paper talks about who the key individuals are that need to be involved in a successful High-Level Synthesis (HLS) hardware design flow and the challenges of designing using HLS.

Who are the key players for successful HLS design flow

Hardware design using HLS is no different than the typical ASIC/FPGA design flow with the exception that C++/SystemC is being used along with HLS to create the RTL instead of hand coding it. The advantage of using HLS is that it speeds up RTL creation time and reduces verification time by producing bug free RTL quickly from a fully verified C++/SystemC source. The misconception that still exists today is that anyone can design using HLS to create optimal hardware. The reality is that you still need to involve all of the people in a typical hardware design flow including the algorithm engineers, the hardware/system architects, and the RTL designers if the expectation is to achieve results as good or better than hand-coded RTL.

Hardware design is hard

Not anyone can do hardware design using HLS for production ASIC/FPGA that is efficient for power, performance, and area (PPA). If this was true there would be no need for hardware/system architects or RTL designers. In fact, hardware design requires an in-depth understanding of memory architecture, concurrent processes, and optimization of algorithms running in software to create efficient implementations running in hardware.

Share

Related resources

Google develops WebM video decompression hardware IP using High-Level Synthesis
White Paper

Google develops WebM video decompression hardware IP using High-Level Synthesis

This paper will describe the actual use of Catapult High-Level Synthesis (HLS) by the WebM team in the successful implementation of the G2 VP9 and share results and impressions.

Working smarter, not harder: NVIDIA closes design complexity gap with HLS
White Paper

Working smarter, not harder: NVIDIA closes design complexity gap with HLS

Discover the challenges NVIDIA faces in the ever evolving world of video, camera, and display standards and the reasons an HLS/C-level flow make it possible for them to succeed in this context.

Machine Learning at the Edge: Using HLS to Optimize Power and Performance
White Paper

Machine Learning at the Edge: Using HLS to Optimize Power and Performance

Moving machine learning to the edge has critical requirements on power and performance. Using off-the-shelf solutions is not practical.