Join

Using a Performance Model to Implement a Superscalar CVA6

Thales built a precise performance model for the CVA6 RISC-V core to evaluate changes before touching RTL. The model achieved COREMARK performance results that were within 0.8% of the RTL, uncovering performance bottlenecks and informing the design of a dual-issue CVA6 core. Outcome? A 40% performance boost with just +7% power and +11% area. This shows how modeling can transform open hardware development.

Project Snapshot

How does performance modeling accelerate real hardware innovation in open RISC-V designs?

In this paper, Thales demonstrates a performance model of the CVA6 RISC-V processor, built to evaluate performance-related modifications before implementing them in RTL. 

By creating a precise cycle-based Python model (99.2% accuracy on Coremark) to detect and fix performance bugs, they guided the implementation of a dual-issue CVA6 core.

The result: +40% higher CoreMark/MHz performance with only a small power (+7%) and area (+11%) increase.

 
In Their Own Words
Share the Post:

Related Posts

The OpenHW Foundation to present at the Latch-Up 2026 conference

See OpenHW Foundation at Latch-Up on Friday to Sunday May 1–3, 2026 in Waterloo, Ontario, Canada. The Latch-Up conference is a weekend of presentations and networking dedicated to free and open source silicon. It’s an event for the open source digital design community, much like its European sister conference ORConf, run by the FOSSi Foundation.

Read More