PC
Back to Home
Semiconductor Digest

Hardware Assisted Verification Is Key to AI Chip Design

Semiconductor Digest

By JUERGEN JAEGER, Director of Prototyping Product Strategy, Siemens EDA

A decade ago, only the largest chipmakers could justify investing in hardware emulation to verify their biggest SoC designs. Juergen Jaeger of Siemens EDA explains why hardware-assisted verification — emulation, enterprise prototyping, and FPGA prototyping — has since become a must-have across mainstream IC design, driven largely by the scale and software-defined complexity of AI chips, where the software increasingly dictates what the hardware needs to do before a single transistor is taped out.

The article breaks down how the three HAV technologies fit together: logic emulation validates RTL-level IC and mixed-signal blocks before they’re integrated; enterprise prototyping runs three to five times faster to give whole design teams earlier insight into hardware/software interaction; and FPGA prototyping lets software developers start building middleware and applications on real silicon-equivalent hardware well before tape-out. Jaeger points to Siemens EDA’s Veloce CS platform as an example of unifying these previously siloed systems under a common operating environment and compiler, so errors surface faster and are easier to trace between systems.

Two trends stand out for where HAV is headed next: comprehensive digital twins that extend emulation beyond logic verification into thermal and power behavior for mission-critical systems, and cloud-based emulation that lets startups and small design teams lease time on emulation and prototyping systems rather than buying and housing the custom racks themselves — effectively democratizing a verification capability that was, until recently, available only to the largest chipmakers.

Click here to read the full article in Semiconductor Digest magazine.

The post Hardware Assisted Verification Is Key to AI Chip Design appeared first on Semiconductor Digest.

📄 Disclaimer: This article was originally published on Semiconductor Digest. PartsCube Global republishes this content for reference purposes only. All rights, including copyright, belong to the original publisher.

← Back to PartsCube Global