[Theory] REMINDER: 1/23 Talks at TTIC: Naama Ben-David, Carnegie Mellon University

Mary Marre mmarre at ttic.edu
Wed Jan 22 14:19:38 CST 2020


*When:*      Thursday, January 23rd at 11:00 am



*Where:*     TTIC, 6045 S. Kenwood Avenue, 5th Floor, Room 526



*Who: *        Naama Ben-David, Carnegie Mellon University




*Title:*         Theoretical Foundations for Practical Distributed Computing

*Abstract: *Over time, we have grown to rely on distributed computation in
our day-to-day interactions with computers; most laptops and phones
nowadays have multiple cores, and behind the scenes, data is often stored
on several machines, either due to its size, or to improve storage
reliability and access speed. This leads to challenges when designing
algorithms on such systems; tasks that may be trivial on a single processor
can become difficult or even impossible in a distributed setting. To help
navigate this space and ensure the design of correct algorithms,
theoretical tools have been developed. However, recent trends in hardware
have yielded new capabilities and semantics for distributed systems, that
have not been reflected in theoretical models. This gap has led to several
buggy systems and missed optimization opportunities.

In this talk, I will highlight some of my work on bringing the theory and
practice of distributed computing closer together. In particular, I will
focus on my work on Remote Direct Memory Access (RDMA), which is a
communication technology recently introduced in data centers. Recent
practical experience demonstrates that RDMA can lead to impressive
performance benefits, but it is unclear whether these are inherent to
RDMA’s non-traditional features, or whether they are just an artifact of
the way RDMA is currently being used. I’ll present a theoretical model of
RDMA, and use it to prove that RDMA has some fundamental advantages over
previous communication mechanisms. I’ll also present experimental results
showing that algorithms developed under our theoretical model outperform
the state-of-the-art in practice.


Host: Avrim Blum <avrim at ttic.edu>



Mary C. Marre
Administrative Assistant
*Toyota Technological Institute*
*6045 S. Kenwood Avenue*
*Room 517*
*Chicago, IL  60637*
*p:(773) 834-1757*
*f: (773) 357-6970*
*mmarre at ttic.edu <mmarre at ttic.edu>*


On Thu, Jan 16, 2020 at 4:08 PM Mary Marre <mmarre at ttic.edu> wrote:

> *When:*      Thursday, January 23rd at 11:00 am
>
>
>
> *Where:*     TTIC, 6045 S. Kenwood Avenue, 5th Floor, Room 526
>
>
>
> *Who: *        Naama Ben-David, Carnegie Mellon University
>
>
>
>
> *Title:*         Theoretical Foundations for Practical Distributed
> Computing
>
> *Abstract: *Over time, we have grown to rely on distributed computation
> in our day-to-day interactions with computers; most laptops and phones
> nowadays have multiple cores, and behind the scenes, data is often stored
> on several machines, either due to its size, or to improve storage
> reliability and access speed. This leads to challenges when designing
> algorithms on such systems; tasks that may be trivial on a single processor
> can become difficult or even impossible in a distributed setting. To help
> navigate this space and ensure the design of correct algorithms,
> theoretical tools have been developed. However, recent trends in hardware
> have yielded new capabilities and semantics for distributed systems, that
> have not been reflected in theoretical models. This gap has led to several
> buggy systems and missed optimization opportunities.
>
> In this talk, I will highlight some of my work on bringing the theory and
> practice of distributed computing closer together. In particular, I will
> focus on my work on Remote Direct Memory Access (RDMA), which is a
> communication technology recently introduced in data centers. Recent
> practical experience demonstrates that RDMA can lead to impressive
> performance benefits, but it is unclear whether these are inherent to
> RDMA’s non-traditional features, or whether they are just an artifact of
> the way RDMA is currently being used. I’ll present a theoretical model of
> RDMA, and use it to prove that RDMA has some fundamental advantages over
> previous communication mechanisms. I’ll also present experimental results
> showing that algorithms developed under our theoretical model outperform
> the state-of-the-art in practice.
>
>
> Host: Avrim Blum <avrim at ttic.edu>
>
>
>
>
> Mary C. Marre
> Administrative Assistant
> *Toyota Technological Institute*
> *6045 S. Kenwood Avenue*
> *Room 517*
> *Chicago, IL  60637*
> *p:(773) 834-1757*
> *f: (773) 357-6970*
> *mmarre at ttic.edu <mmarre at ttic.edu>*
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cs.uchicago.edu/pipermail/theory/attachments/20200122/db8ea125/attachment.html>


More information about the Theory mailing list