End to End Data Science Without Leaving The GPU

End to End Data Science Without Leaving the GPU

As machine learning workflows have become more popular, they have also become overly complicated - requiring too many tools, which require shuffling data around between environments (CPU to GPU to CPU and back). The better solution is to leverage GPU memory end-to-end, and GOAI’s GPU Data Frame spec (based on Apache Arrow) is the open source, industry standard way to do it. We’ll deep dive on that standard, and how it supports the modern data science workflow entirely in GPU memory. We will use tools including MapD’s open source GPU database, Jupyter, and machine learning algorithms from H2O.

Randy Zwitch, Senior Developer Advocate at MapD

Randy Zwitch is a Senior Developer Advocate at MapD, enabling customers and community users alike to utilize MapD to its fullest potential. With broad industry experience in Energy, Digital Analytics, Banking, Telecommunications and Media, Randy brings a wealth of knowledge across verticals as well as an in-depth knowledge of open-source tools for analytics.

View the recording here: https://www.youtube.com/watch?v=zy9ipXv0cz0

Email community@mapd.com if you’re interested in collaborating!

Try our 14-day trial: https://www.mapd.com/meet-cloud

Transcript

1. © MapD 2018 @randyzwitch End to End Data Science Without Leaving The GPU Randy Zwitch | July 18, 2018
2. © MapD 2018 @randyzwitch 2 Agenda Introduction (5 mins) Apache Arrow (5 mins) GOAI and the GPU DataFrame (5 mins) Live Code Example (20 mins) Questions
3. © MapD 2018 @randyzwitch 3 About Randy Zwitch - Senior Developer Advocate at MapD • 15 years predictive modeling and data engineering experience across energy, banking, and media verticals • Contributor to R, Python and Julia open-source communities • Started at MapD in March 2018 to support the MapD user community and to publicly demonstrate the power of GPUs for business analytics and data science • Professional inquiries: randy.zwitch@mapd.com
4. © MapD 2018 @randyzwitch 4 Apache Arrow: Shared Memory Layout Source: https://arrow.apache.org/
5. © MapD 2018 @randyzwitch 5 GOAI and the GPU DataFrame GPU Open Analytics Initiative – fusing Machine Learning and GPU analytics Data Warehouse Data Lake/HDFS
6. © MapD 2018 @randyzwitch DEMO https://github.com/mapd/AICamp_demo_201807 6
7. Questions/Comments? randy.zwitch@mapd.com https://community.mapd.com/