<Data Center> Datacenter Compute SoC Performance Modeling Architect

Job description

Role Summary

We are looking for a highly skilled Datacenter Compute SoC Performance Modeling Engineer. In this role, you will be at the forefront of defining our next-generation cloud and datacenter processors. You will develop and maintain high-speed, cycle-accurate, and transaction-level architectural simulators to evaluate design trade-offs, identify performance bottlenecks, and project system performance across a wide range of datacenter workloads.

You will collaborate closely with CPU/Compute core architects, SoC architects, IP architects, workload characterization teams, and RTL designers to ensure our silicon delivers industry-leading performance-per-watt and scales efficiently to meet the demands of modern hyper-scale infrastructure.

Simulator Development

• Design, develop, and maintain cycle-accurate and transaction-level performance models (C++/SystemC) for multi-core datacenter SoCs, including CPU cores, memory subsystems, interconnects (NoC), and high-speed I/O (PCIe/CXL).

Workload Analysis & Profiling

• Analyze and characterize cloud-native workloads, microservices, databases, and AI/ML data processing pipelines to drive architectural requirements.

Performance Exploration

• Conduct what-if analyses and architectural design sweeps to optimize cache sizes, memory bandwidth, latency, interconnect topologies, and core counts.

Bottleneck Identification

• Identify micro-architectural and system-level performance bottlenecks and propose innovative hardware or software solutions.

RTL & Silicon Correlation

• Work with RTL verification and post-silicon validation teams to correlate the performance model against RTL simulations, emulators, and actual silicon, ensuring model accuracy.

Tooling & Infrastructure

• Build and enhance tracing tools, data visualization dashboards, and automated performance regression pipelines using Python and data analysis libraries. #LI-JK1