Role Overview
We are looking for a curious, hands-on and self-motivated intern to join the GPU Compiler team in Cambridge for a 3-month placement. The GPU compiler translates shaders and compute kernels into efficient machine code for Qualcomm Adreno GPUs.
Responsibilities
- Implement, test and benchmark compiler features or optimisations within the GPU compiler toolchain.
- Analyse generated GPU code and runtime performance data to understand the impact of compiler transformations.
- Write and run graphics and compute tests to assess functionality and performance.
- Develop scripts and tools for test automation, data processing and performance visualisation.
- Collaborate with GPU architecture, driver and validation teams to understand hardware behaviour.
- Document designs, experiments, results and recommendations.
Requirements
- Good programming skills in C++, plus familiarity with Python or a similar scripting language.
- Understanding of compiler fundamentals, such as parsing, intermediate representations, data-flow analysis, optimisation passes and code generation.
- Basic understanding of computer architecture: pipelines, memory hierarchy, caches and instruction sets.
- Strong analytical, debugging and problem-solving skills.
- Interest in performance and how software maps onto hardware.