Accelerating Data Science Workflows with RAPIDS

[Pages:53]HIGH-PERFORMANCE DATA SCIENCE WITH RAPIDS

Zahra Ronaghi

AI Infrastructure Manager

END-TO-END ACCELERATED GPU DATA SCIENCE

Data Processing Evolution

Faster data access, less data movement

Hadoop Processing, Reading from disk

HDFS Read

Query

HDFS HDFS Write Read

Spark In-Memory Processing

ETL

HDFS HDFS Write Read

HDFS Read

Query

ETL

ML Train

Traditional GPU Processing

HDFS Read

GPU Read

Query WCrPitUe

GPU Read

ETL

CPU Write

GPU ML Read Train

5-10x Improvement More code

Language rigid Substantially on GPU

ML Train

25-100x Improvement Less code

Language flexible Primarily In-Memory

3

Data Movement and Transformation

The bane of productivity and performance

APP B

Read Data

CPU

APP B

Copy & Convert APP A

Copy & Convert Copy & Convert

APP B

GPU Data

GPU Data

APP A

GPU

APP A

Load Data

4

Data Movement and Transformation

What if we could keep data on the GPU?

APP B

Read Data

CPU

APP B

Copy & Convert APP A

Copy & Convert Copy & Convert

APP B

GPU Data

GPU Data

APP A

GPU

APP A

Load Data

5

Learning from Apache Arrow

From Apache Arrow Home Page -

6

Data Processing Evolution

Faster data access, less data movement

Hadoop Processing, Reading from disk

HDFS Read

Query

HDFS HDFS Write Read

Spark In-Memory Processing

ETL

HDFS HDFS Write Read

HDFS Read

Query

ETL

ML Train

Traditional GPU Processing

HDFS Read

GPU Read

Query WCrPitUe

GPU Read

ETL

CPU Write

GPU ML Read Train

RAPIDS

Arrow Read

Query ETL

ML Train

5-10x Improvement More code

Language rigid Substantially on GPU

50-100x Improvement Same code

Language flexible Primarily on GPU

ML Train

25-100x Improvement Less code

Language flexible Primarily In-Memory

7

Open Source Data Science Ecosystem Familiar Python APIs

Data Preparation

Pandas Analytics

Dask

Scikit-Learn Machine Learning

Model Training

Visualization

NetworkX Graph Analytics

PyTorch Chainer MxNet Deep Learning

Matplotlib/Seaborn Visualization

CPU Memory

8

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download