Hi there. I am curious about using GPU acceleration for Data Science. Due to GPU's high concurrency level and memory bandwidth, I think that possibility would be very useful. Can I apply it for my models training, querying or processing large datasets, or some kind of visualization? If it's possible, what should I start with? It would be nice if you provide a few technologies or libraries examples.
Hi. There are several options. Check out Rapids. This is an API providing tools, facilitating using GPU for machine learning, data processing, etc. It has its own Python library - cuDF. https://rapids.ai/ https://github.com/rapidsai/cudf
BlazingSQL https://blazingsql.com/ is a tool for querying arbitrary datasets with SQL using GPU. It's based on Rapids. Also, TensorFlow supports executing on GPU. You can learn about it here https://www.tensorflow.org/install/gpu . Good luck!
Just drop us an email to ...
Show more 
Log in with Google