Max-Planck-Institut für Informatik
max planck institut
informatik
mpii logo Minerva of the Max Planck Society
 

MPI-INF or MPI-SWS or Local Campus Event Calendar

<< Previous Entry Next Entry >> New Event Entry Edit this Entry Login to DB (to update, delete)
What and Who
Title:Transparent Scaling of Deep Learning Systems through Dataflow Graph Analysis
Speaker:Jinyang Li
coming from:New York University
Speakers Bio:Jinyang Li is a professor of computer science at New York University.  Her research is focused on developing better system

infrastructure to accelerate machine learning and web applications. Most recently, her group has released DGL, an open-source library
for programming graph neural networks.  Her honors include a NSF CAREER award, a Sloan Research Fellowship and multiple Google
research awards.  She received her B.S. from National University of Singapore and her Ph.D. from MIT, both in Computer Science.

Event Type:SWS Distinguished Lecture Series
Visibility:D1, D2, D3, INET, D4, D5, SWS, RG1, MMCI
We use this to send out email in the morning.
Level:AG Audience
Language:English
Date, Time and Location
Date:Friday, 17 May 2019
Time:10:30
Duration:90 Minutes
Location:Saarbrücken
Building:E1 5
Room:002
Abstract
As deep learning research pushes towards using larger and more sophisticated models, system infrastructure must use many GPUs efficiently. Analyzing the
dataflow graph that represents the DNN computation is a promising avenue for optimization. By specializing execution for a given dataflow graph, we can
accelerate DNN computation in ways that are transparent to programmers. In this talk, I show the benefits of dataflow graph analysis by discussing two
recent systems that we've built to support large model training and low-latency inference. To train very large DNN models, Tofu automatically re-writes a
dataflow graph of tensor operators into an equivalent parallel graph in which each original operator can be executed in parallel across multiple GPUs.  To
achieve low-latency inference, Batchmaker discovers identical sub-graph computation among different requests to enable batched execution of requests
arriving at different times. 
Contact
Name(s):Annika Meiser
Phone:93039105
EMail:--email address not disclosed on the web
Video Broadcast
Video Broadcast:YesTo Location:Kaiserslautern
To Building:G26To Room:111
Meeting ID:SWS Space 2 (6312)
Tags, Category, Keywords and additional notes
Note:
Attachments, File(s):

Created:Annika Meiser/MPI-SWS, 04/02/2019 02:17 PM Last modified:Uwe Brahm/MPII/DE, 04/03/2019 07:01 AM
  • Annika Meiser, 04/02/2019 02:23 PM
  • Annika Meiser, 04/02/2019 02:22 PM -- Created document.