Tensilica IP division of Cadence is looking for a computer architecture analyst to join our architecture performance team to work on highly configurable and extensible architectures ranging from domain specific DSPs used in vision, audio applications to high end ML accelerator IPs. The candidate will be intimately involved in analysis that will go into the design and optimization of future architectures, extensions and implementations for high performance and low power compute architectures and systems.


In this role, your responsibilities will include:

Architectural exploration and path finding studies for next generation DNA 100 IP for ML/AI applications
Understanding of CNN based workloads used in classification, object detection, segmentation and all similar ML/AI application
Creating and maintaining speed-of-light (SOL) performance models and conducting roofline analysis to understand performance related bottlenecks
Identification of architectural, micro-architectural and ISA enhancements for cycle, power and frequency optimizations
Collaboration with hardware and CAD teams to evaluate the impacts of proposed features
Identification and prototyping of new analysis tools
Improvements to existing trace-based and simulation-based tools
Collaboration with software development tools team to ensure that solutions work well with software toolchains
Collaboration with target software development teams to work with benchmark suites.
Trace-based and simulation based analysis of proposed architectures
Presentation of analysis and proposed architectures to stakeholders
Help guide early products definition by providing alternative PPA design options for the product and marketing teams

For more details, hit “Apply for job”