Vaishaal Shankar and Karl Krauth and Qifan Pu and Eric Jonas and Shivaram Venkataraman and Ion Stoica and Benjamin Recht and Jonathan Ragan-Kelley

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2018-137

October 22, 2018

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-137.pdf

Linear algebra operations are widely used in scientific computing and machine learning applications. However, it is challenging for scientists and data analysts to run linear algebra at scales beyond a single machine. Traditional approaches either require access to supercomputing clusters, or impose configuration and cluster management challenges. In this paper we show how the disaggregation of storage and compute resources in so-called ``serverless'' environments, combined with compute-intensive workload characteristics, can be exploited to achieve elastic scalability and ease of management.

We present numpywren, a system for linear algebra built on a serverless architecture. We also introduce lambdapack, a domain-specific language designed to implement highly parallel linear algebra algorithms in a serverless setting. We show that, for certain linear algebra algorithms such as matrix multiply, singular value decomposition, and Cholesky decomposition, numpywren's performance (completion time) is within 33% of ScaLAPACK, and its compute efficiency (total CPU-hours) is up to 240% better due to elasticity, while providing an easier to use interface and better fault tolerance. At the same time, we show that the inability of serverless runtimes to exploit locality across the cores in a machine fundamentally limits their network efficiency, which limits performance on other algorithms such as QR factorization. This highlights how cloud providers could better support these types of computations through small changes in their infrastructure.

Advisors: Benjamin Recht and Jonathan Ragan-Kelley

---

BibTeX citation:

@mastersthesis{Shankar:EECS-2018-137,
    Author= {Shankar, Vaishaal and Krauth, Karl and Pu, Qifan and Jonas, Eric and Venkataraman, Shivaram and Stoica, Ion and Recht, Benjamin and Ragan-Kelley, Jonathan},
    Title= {numpywren: serverless linear algebra},
    School= {EECS Department, University of California, Berkeley},
    Year= {2018},
    Month= {Oct},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-137.html},
    Number= {UCB/EECS-2018-137},
    Abstract= {Linear algebra operations are widely used in scientific computing and machine learning applications. However, it is challenging for scientists and data analysts to run linear algebra at scales beyond a single machine. Traditional approaches either require access to supercomputing clusters, or impose configuration and cluster management challenges. In this paper we show how the disaggregation of storage and compute resources in so-called ``serverless'' environments, combined with compute-intensive workload characteristics, can be exploited to achieve elastic scalability and ease of management. 

We present numpywren, a system for linear algebra built on a serverless architecture. We also introduce lambdapack, a domain-specific language designed to implement highly parallel linear algebra algorithms in a serverless setting. We show that, for certain linear algebra algorithms such as matrix multiply, singular value decomposition, and Cholesky decomposition, numpywren's performance (completion time) is within 33% of ScaLAPACK, and its compute efficiency (total CPU-hours) is up to 240% better due to elasticity, while providing an easier to use interface and better fault tolerance.
At the same time, we show that the inability of serverless runtimes to exploit locality across the cores in a machine fundamentally limits their network efficiency, which limits performance on other algorithms such as QR factorization.
This highlights how cloud providers could better support these types of computations through small changes in their infrastructure.},
}

EndNote citation:

%0 Thesis
%A Shankar, Vaishaal 
%A Krauth, Karl 
%A Pu, Qifan 
%A Jonas, Eric 
%A Venkataraman, Shivaram 
%A Stoica, Ion 
%A Recht, Benjamin 
%A Ragan-Kelley, Jonathan 
%T numpywren: serverless linear algebra
%I EECS Department, University of California, Berkeley
%D 2018
%8 October 22
%@ UCB/EECS-2018-137
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2018/EECS-2018-137.html
%F Shankar:EECS-2018-137