Automatic Full Compilation of Julia Programs and ML Models to Cloud TPUs
This enables Julia users to efficiently run ML workloads on TPUs without modifying core compiler or packages, though it is incremental as it builds on existing APIs and compiler techniques.
The authors tackled the problem of compiling Julia programs and machine learning models to Google Cloud TPUs, achieving a forward pass evaluation of a VGG19 model on a batch of 100 images in 0.23s compared to 52.4s on CPU.
Google's Cloud TPUs are a promising new hardware architecture for machine learning workloads. They have powered many of Google's milestone machine learning achievements in recent years. Google has now made TPUs available for general use on their cloud platform and as of very recently has opened them up further to allow use by non-TensorFlow frontends. We describe a method and implementation for offloading suitable sections of Julia programs to TPUs via this new API and the Google XLA compiler. Our method is able to completely fuse the forward pass of a VGG19 model expressed as a Julia program into a single TPU executable to be offloaded to the device. Our method composes well with existing compiler-based automatic differentiation techniques on Julia code, and we are thus able to also automatically obtain the VGG19 backwards pass and similarly offload it to the TPU. Targeting TPUs using our compiler, we are able to evaluate the VGG19 forward pass on a batch of 100 images in 0.23s which compares favorably to the 52.4s required for the original model on the CPU. Our implementation is less than 1000 lines of Julia, with no TPU specific changes made to the core Julia compiler or any other Julia packages.