Getting Started Kit
Evaluate your workloads for GPU acceleration and learn to configure your cluster for cost savings.
The NVIDIA RAPIDS™ Accelerator for Apache Spark automatically replaces viable operations with GPU-accelerated versions without any code changes. Iteratively design your workloads to optimize for performance and budget without any risk, since you can revert back to CPU-only infrastructure at any time.
Ease the migration of your Apache Spark workloads to GPU-based instances with the Accelerated Spark Analysis tool. The tool guides you through the three steps to implementation.
Identify and estimate the cost savings and acceleration potential of your Spark workloads based on an analysis of the log files. You’ll receive a list of apps recommended for GPU acceleration with estimated savings and speedups.
Get recommended configuration parameters for GPU acceleration on your Spark cluster based on the cluster shape.
Optimize the GPU-acceleration configuration based on the initial (bootstrap) job run using Spark event logs. You’ll receive recommended configuration settings for each application.
Watch Video ›
Apply for a free consultation by filling out the interest form below.
While waiting to be contacted by a member of our team, you will have access to getting started resources.
A member of our team will be in contact with you to determine if you qualify for a free consultation with NVIDIA experts.
NVIDIA Privacy Policy