This parallelism is why GPUS are really good for machine learning/ deep learning tasks: https://medium.com/@shachishah.ce/do-we-really-need-gpu-for-deep-learning-47042c02efe2
drewkaul
Overall, GPUs have proven to be well-suited for deep learning tasks because they have high bandwidth main memory, thread parallelism, and a large, fast L1 and register memory.
kavimehta
Google has gone a step further and developed TPUs specifically for machine learning https://cloud.google.com/tpu/
This parallelism is why GPUS are really good for machine learning/ deep learning tasks: https://medium.com/@shachishah.ce/do-we-really-need-gpu-for-deep-learning-47042c02efe2
Overall, GPUs have proven to be well-suited for deep learning tasks because they have high bandwidth main memory, thread parallelism, and a large, fast L1 and register memory.
Google has gone a step further and developed TPUs specifically for machine learning https://cloud.google.com/tpu/