Accelerating PCB Layout Editor using modern GPU architecture for complex designs
TimeThursday, December 9th11:00am - 11:30am PST
Event Type
Special Session (Research Track)
Virtual Programs
Hosted in Virtual Platform
Presented In-Person
Machine Learning/AI
DescriptionAs the complexity of modern electronics continuously grows, EDA products are getting challenged to keep up with performance. The traditional CPU rendering approach is not going to be sufficient to be able to process 10s and 100s of millions of objects at a reasonable speed. Users are demanding a responsive tool running on modern hardware platforms. This talk describes the approach of using GPUs to accelerate Cadence Allegro rendering, augmenting the accuracy and quality of the graphics and improving other highly expensive tool functions. With millions of objects to be rendered and manipulated in Allegro, we will discuss the following optimizations that needed to be considered: (1) How to efficiently render different database element types; (2) Scene graph and memory cache, and why they are a fit for the GPU; (3) Incremental database updates; (4) Technology that scales with the hardware used; and (5) the NVIDIA Path rendering extension.