# SparCML: High-Performance Sparse Communication for Machine Learning

## TL;DR

SparCML introduces scalable sparse communication protocols that significantly improve the efficiency of distributed machine learning by exploiting the sparsity in gradient data, extending MPI with new features.

## Contribution

The paper presents novel communication protocols and a library, SparCML, that enable efficient sparse data exchange in distributed machine learning frameworks.

## Key findings

- Achieves reduced communication overhead for sparse gradients.
- Supports asynchronous and low-precision operations.
- Lays foundation for future scalable ML systems.

## Abstract

Applying machine learning techniques to the quickly growing data in science and industry requires highly-scalable algorithms. Large datasets are most commonly processed "data parallel" distributed across many nodes. Each node's contribution to the overall gradient is summed using a global allreduce. This allreduce is the single communication and thus scalability bottleneck for most machine learning workloads. We observe that frequently, many gradient values are (close to) zero, leading to sparse of sparsifyable communications. To exploit this insight, we analyze, design, and implement a set of communication-efficient protocols for sparse input data, in conjunction with efficient machine learning algorithms which can leverage these primitives. Our communication protocols generalize standard collective operations, by allowing processes to contribute arbitrary sparse input data vectors. Our generic communication library, SparCML, extends MPI to support additional features, such as non-blocking (asynchronous) operations and low-precision data representations. As such, SparCML and its techniques will form the basis of future highly-scalable machine learning frameworks.