Role in this project:
Back-end Developer / System Architect Contributions:2 releases, 42 commits, 1 PR in 1 year 11 months
Contributions summary:Prashant focused on implementing and integrating GEM (Graphics Engine Memory) support within the NVDLA (NVIDIA Deep Learning Accelerator) driver, specifically for DMA buffer allocation, mapping, and DMA address retrieval. They implemented core NVDLA HW programming including the engine scheduler for scheduling HW layers on sub-engines. Further contributions involved updating register specifications and adding new features. The user was responsible for registering NVDLA memory for DMA, ensuring memory access by NVDLA, and addressing synchronization issues by modifying the LUT programming strategy.