NVIDIA uses Vera CPU to develop future processors in a ‘continuous feedback loop’
© NVIDIA NVIDIA has started using its Vera CPU across engineering workflows involved in developing future NVIDIA CPUs and GPUs. The company says early testing showed up to 1.5x higher performance in selected chip verification and simulation workloads.
Vera accelerates CPU-dependent chip verification
Processor development relies on simulation, formal verification, regression testing and digital implementation before a design can enter production. GPUs and AI can accelerate parts of this process, but several electronic design automation workloads still depend heavily on CPU core performance, memory access and latency.
NVIDIA tested Vera with Cadence Jasper and Synopsys VCS. Jasper is used for formal verification, while VCS simulates and validates chip designs before fabrication. Both applications recorded gains of up to 1.5x in selected production-class workloads, according to NVIDIA.
© NVIDIA Vera combines 88 custom NVIDIA Olympus CPU cores with LPDDR5X memory and the second-generation NVIDIA Scalable Coherent Fabric. NVIDIA says the design can shorten individual verification jobs and increase throughput across large regression testing farms.
Each NVIDIA CPU generation helps build the next
The company is now deploying Vera throughout the EDA workflows used to create its future processors. Faster verification allows engineers to test more design alternatives and identify problems earlier in the development process.
NVIDIA plans to follow Vera with the Rosa CPU, which will introduce the NVIDIA Rigel core . By using Vera to develop future CPUs and GPUs, NVIDIA is linking its silicon, software and system engineering work in a continuous feedback loop. Each CPU generation will effectively help the company design the hardware that follows it.
Source: NVIDIA Blog
