San Diego Super Computer Center Workshop

You're invited to attend a full day workshop presented by staff from the San Diego Super Computer Center. The morning will provide general information and resources, and the afternoon will split into two parallel tracks (Data Science and Compute Intensive). You can select your preferred track during the RSVP process. An overview of the workshop agenda is provided below:

Morning Session:

Overview of SDSC resources and how to access them ~ 1 hour

  • Background on SDSC
  • Quick summaries on Comet, TSCC
  • Some science stories - use cases
  • HPCforUC
  • TSCC access
  • XSEDE allocations
  • Science Gateways

Introduction to SDSC Comet ~ 2hours 

  • Comet architecture details
  • Software stack
  • Running jobs on CPU, GPU resources - hands on examples
  • Overview of containerization on Comet using Singularity
  • Building singularity containers and running jobs - hands on examples


Afternoon Session 1 [Data Analysis and ML/DL Track]

  • HPC and Data Analysis using Python and Jupyter Notebooks [1.5 hours]
  • Use of Jupyter notebooks on Comet
  • Python for HPC
  • Using python for data preparation on, and explore pandas (the Python data analysis toolkit)
  • Hands on exercises


Introduction to Machine Learning and Deep Learning on Comet [1.5 hours]

  • Overview of machine learning / deep learning tools available on Comet
  • Examples using R, Python, and Keras with TensorFlow
  • Hands on exercises


Afternoon Session 2 [HPC Applications Track]


Performance Tuning on HPC resources [1.5 hours]

  • Optimization techniques aim at code developers
  • Compiler options and their limitations
  • Cache, loop-level optimizations
  • Reductions, time-space tradeoffs etc


Profiling and Tracing HPC Applications on Comet [1.5 hours]

  • Overview of tools - mpiP, Allinea MAP, TAU
  • Using mpiP - hands on compilation example and sample results
  • Allinea MAP - profiling examples for CPU, GPU applications
  • Using TAU with library preloading - example
  • Code instrumentation and compiling with TAU wrappers

Friday, November 2, 2018 at 9:30am to 4:30pm

University Laboratory Building, 104
University Laboratory Bldg, Riverside, CA 92507

Event Registration Required

This event requires registration.