Add meta post about frameworks and platforms #310
Conversation
|
|
||
| To make things clearer let's start by answering questions like _"Which packages make up RAPIDS?"_ and _"Who is in the RAPIDS community?"_. | ||
|
|
||
| ## The many projects of RAPIDS |
There was a problem hiding this comment.
This seems like a really useful way to convey to first-time users the world of RAPIDS (thinking of that PyData ecosystem image that gets passed around a lot), and I'm wondering if any of this can be applied to the about section of RAPIDS' main page, where some of these packages aren't mentioned?
There was a problem hiding this comment.
Thanks @charlesbluca.
What do you think @exactlyallan? The goal of this doc is to help explain some backround about RAPIDS from a deployment perspective, but perhaps some of this copy would be useful to put on https://rapids.ai?
|
|
||
| ## Multi-node and multi-GPU deployments | ||
|
|
||
| As you move to larger datasets you will want to scale out to multiple GPUs and **multi-node deployments**. |
There was a problem hiding this comment.
What's the difference between mult-GPU and multi-node?
There was a problem hiding this comment.
Good point, does this make things clearer?
| As you move to larger datasets you will want to scale out to multiple GPUs and **multi-node deployments**. | |
| As you move to larger datasets you will want to scale out to multiple GPUs which may be spread over multiple nodes with **multi-node deployments**. |
Co-authored-by: Mike McCarty <mmccarty@nvidia.com>
3 participants
I often find myself having conversations around deploying RAPIDS where folks ask questions like:
I thought I'd draft up some general thoughts on answering some of these questions. Title, location, etc all TBC but thought I'd just throw something here for feedback.