Split gpu_id.h and GpuIdManager out from build target...
Split gpu_id.h and GpuIdManager out from build target //tensorflow/core:gpu_runtime, to reduce the size of dependencies, so when other lightweight libraries like grappler utils needs the TfToCudaGpuId translation function it doesn't need to depend on things like stream executor and cuda libraries. PiperOrigin-RevId: 185175757
Loading
Please sign in to comment