The preceding move was aimed at helping advance the adoption of efficient AI inferencing in custom hardware designs, NVIDIA said. The same NVDLA is shipped in the Nvidia Jetson AGX Xavier Developer Kit, where it is said to offer peak efficiency of 7.9 TOPS/W for AI.
The compiler is a key component of the NVDLA software stack. It generates optimized execution graphs which map the tasks defined in the layers of pre-trained neural network models onto the various execution units in NVDLA.
The open-source release of NVDLA’s optimizing compiler is now available on GitHub.
With this compiler release, NVDLA users have full-access to the software and hardware source needed to integrate, grow, and explore the NVDLA platform.