Overview

NeuReality, based in Caesarea, Israel, is an innovative AI infrastructure company founded in 2019 by a team of experienced system engineers. The company develops purpose-built system solutions that eliminate bottlenecks and maximize GPU utilization for next-generation data centers.

With a deep technical expertise and ambitious team, the company has quickly become a significant player in the AI market.

Its flagship products - the NR1® AI-CPU, which unlocks full GPU performance for inference, and the NR2 AI-SuperNIC, which delivers purpose-built networking and connectivity for scalable AI inference and training - work together to enable smarter, faster, and more efficient AI infrastructure, turning innovation into measurable business impact

Challenges

NeuReality was exploring ways to optimize their verification process while uncovering corner case bugs. Their challenges included:
  • Scalability: NeuReality wanted to optimize regression turnaround time TAT for thousands of tests to facilitate resource scaling.
  • Corner case bugs: Early detection of bugs is crucial, as missed issues can lead to costly re-spins.
  • Verification quality: Eliminate redundancies in the regressions and target new design areas to exercise.

Solution

To address these challenges, NeuReality adopted Synopsys VCS® Intelligent Constraint Optimization (ICO) and VSO.ai™ (Verification Space Optimization). Key features of the solution included:

  • High-Quality Stimuli Generation: Bias input stimuli to diversify constraint random stimuli creation, improving the likelihood of hitting corner cases and uncovering new bugs.
  • Regression Turnaround Time (TAT): Prioritize the high ROI tests initially to achieve the same coverage in shorter regression cycles.
  • Seamless Integration: Integrates into existing VCS or VC Execution Manager verification environment.
  • Improved HW Utilization: Maximizing coverage within the available compute budget.

Results

Through the implementation of VCS and VSO.ai, NeuReality achieved remarkable results:

  • Regression TAT Optimization: Achieved the same coverage with 2X fewer tests.
  • Early Testbench Stabilization: Uncovered a missing constraint on the address bus in the testbench.
  • Design Bug Detection: In a directed scenario, sequencing three parallel events exposed a corner case unaccounted for which needed an RTL fix.

NeuReality’s collaboration with Synopsys on deployment of these advanced technologies led to a more efficient, scalable, and high-quality verification framework. This success underscores the impact of advanced AI/ML technologies in driving innovation and operational excellence in software development.