NVIDIA Developer Forums

Not enough GPU memory? Not a problem

AI & Data Science Data Science of the Day

jimscott January 6, 2021, 9:16pm 1

Did you know reading more data than there is #GPU memory isn't a problem? Here, we see that reading 100 million records (58 GB) from a CSV file 2–3x faster than with CPU.
Click the image to read the article
Find more #DSotD posts

1 Like

Topic		Replies	Views	Activity
Jax - Numphy on GPUs and TPUs Data Science of the Day fun-facts , gpu , ml	0	1259	April 21, 2022
An Introduction to GPU DataFrames for Pandas Users Data Science of the Day fun-facts , dataframes-preprocessing-cudf , rapids	0	1392	March 12, 2021
An Introduction to Distributed Computing with GPUs in Python Data Science of the Day python , fun-facts , dask , data-science	0	1988	March 29, 2021
A Decent GPU is Crucial for Machine Learning Data Science of the Day fun-facts , gpu , ml	0	751	May 17, 2022
Driving XGBoost in the GPU Fastlane Data Science of the Day fun-facts , dask , gpu , xgboost	0	1267	February 9, 2021
How to Track Your GPU Usage During Machine Learning Data Science of the Day fun-facts , machine-learning , mlops	0	701	November 15, 2021
Event Streaming on the GPU... Yep, It's a Thing! Data Science of the Day python , fun-facts , gpu	0	851	February 15, 2021
GPUs May Be Better, Not Just Faster, at Training Deep Neural Networks Data Science of the Day ai , fun-facts , neutral-network	0	1235	September 20, 2021
Multi-Instance GPUs, with Kevin Klues and Pradeep Venkatachalam Data Science of the Day cuda , fun-facts , kubernetes	0	1031	July 19, 2021
Embed Your SQL Query Into Your Python Code and Let It Rip on a GPU Data Science of the Day python , fun-facts	0	1948	March 16, 2021