Skip to content

[DLRM/PyT] Readme fixes #655

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
nv-kkudrynski merged 1 commit into from
Aug 18, 2020
Merged

[DLRM/PyT] Readme fixes #655

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion PyTorch/Recommendation/DLRM/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -139,7 +139,7 @@ TF32 is supported in the NVIDIA Ampere GPU architecture and is enabled by defaul

### Hybrid-parallel multiGPU with all-2-all communication

Many recommendation models contain very large embedding tables. As a result the model is often too large to fit onto a single device. This could be easily. This could be easily solved by training in a model-parallel way, using either the CPU or other GPUs as "memory donors". However, this approach is suboptimal as the "memory donor" devices' compute is not utilized. In this repository we use the model-parallel approach for the bottom part of the model (Embedding Tables + Bottom MLP) while using a usual data parallel approach for the top part of the model (Dot Interaction + Top MLP). This way we can train models much larger than what would normally fit into a single GPU while at the same time making the training faster by using multiple GPUs. We call this approach hybrid-parallel.
Many recommendation models contain very large embedding tables. As a result the model is often too large to fit onto a single device. This could be easily solved by training in a model-parallel way, using either the CPU or other GPUs as "memory donors". However, this approach is suboptimal as the "memory donor" devices' compute is not utilized. In this repository we use the model-parallel approach for the bottom part of the model (Embedding Tables + Bottom MLP) while using a usual data parallel approach for the top part of the model (Dot Interaction + Top MLP). This way we can train models much larger than what would normally fit into a single GPU while at the same time making the training faster by using multiple GPUs. We call this approach hybrid-parallel.

The transition from model-parallel to data-parallel in the middle of the neural net needs a specific multiGPU communication pattern called [all-2-all](https://en.wikipedia.org/wiki/All-to-all_\(parallel_pattern\)) which is available in our [PyTorch 20.06-py3] NGC docker container. In the [original DLRM whitepaper](https://arxiv.org/abs/1906.00091) this has been also referred to as "butterlfy shuffle".

Expand Down
2 changes: 1 addition & 1 deletion PyTorch/Recommendation/DLRM/triton/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
# Deploying the DLRM model using Triton Inference Server

The [NVIDIA Triton Inference Server](https://github.com/NVIDIA/trtis-inference-server) provides a datacenter and cloud inferencing solution optimized for NVIDIA GPUs. The server provides an inference service via an HTTP or gRPC endpoint, allowing remote clients to request inferencing for any number of GPU or CPU models being managed by the server.
The [NVIDIA Triton Inference Server](https://github.com/NVIDIA/triton-inference-server) provides a datacenter and cloud inferencing solution optimized for NVIDIA GPUs. The server provides an inference service via an HTTP or gRPC endpoint, allowing remote clients to request inferencing for any number of GPU or CPU models being managed by the server.

This folder contains instructions for deploment and exemplary client application to run inference on
Triton Inference Server as well as detailed performance analysis.
Expand Down