GPU Performance Verification Engineer

Overview

Performance verification of GPU graphics logic components (3D graphics, media and display) within Intel's Data Center Group, supporting GPUs for AI, machine learning and high-performance computing workloads.

What you'll do

• Perform performance verification of graphics logic components, including 3D graphics, media and display, to ensure the design meets specification requirements.
• Define and develop scalable and reusable IP verification plans, test benches and verification environment architecture to confirm coverage against graphics microarchitecture specifications.
• Execute verification plans, define and run system simulation models to verify the design, analyse power and timing, and uncover bugs.
• Replicate, root-cause and debug issues in the presilicon environment, and implement corrective measures to resolve failing tests.
• Collaborate with GPU architects, RTL developers and physical design teams to improve verification of complex architectural and microarchitectural features and meet functional, performance and power goals.
• Document test plans and drive technical reviews of plans and proofs with design and architecture teams.
• Maintain and improve existing performance verification infrastructure and methodology.
• Participate in defining verification infrastructure and related TFMs needed for performance design verification.

What you'll need

• Bachelor's or Master's degree with 2-10 years of experience in design verification.
• Experience with verification of subsystem and/or SoC having GPU/CPU.
• Experience writing UVM-based tests, assembly-based tests, and tests using high-level languages such as OpenGL and/or OpenCL.
• Knowledge of graphics 3D and/or compute pipeline.
• Knowledge of cache controllers and memory protocols such as HBM, LPDDR and GDDR.
• Knowledge of industry-standard emulation tools such as Zebu, Veloce or Palladium.

Nice to have

• Experience of performance verification.
• Knowledge of performance analysis or bottleneck analysis of 3D and/or compute workloads.

Details

• Location: Bangalore, India.
• Work model: hybrid, splitting time between on-site at the assigned Intel site and off-site.
• Job type: experienced hire; Shift 1 (India).