SparCML: High-Performance Sparse Communication for Machine Learning

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 gl...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:SC19: International Conference for High Performance Computing, Networking, Storage and Analysis s. 1 - 15
Hlavní autoři: Renggli, Cedric, Ashkboos, Saleh, Aghagolzadeh, Mehdi, Alistarh, Dan, Hoefler, Torsten
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: ACM 17.11.2019
Témata:
Distributed databases
High performance computing
Industries
Libraries
Machine learning
Machine learning algorithms
Protocols
Scalability
Sparse AllGather
Sparse AllReduce
Sparse Input Vectors
Vectors
ISSN:2167-4337
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí: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 1 1 Stands for Sparse Communication layer for Machine Learning, to be read as sparse ML., 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.
ISSN:2167-4337
DOI:10.1145/3295500.3356222