[LA-SiGMA-TESC] Fwd: [Allcct] Reminder: HPC user symposium keynote speech "Programming for Hybrid Architectures Today and in the Future" on Thursday at 8:30AM
Le Yan
lyan1 at tigers.lsu.edu
Wed Jun 4 21:32:22 CDT 2014
---------- Forwarded message ----------
From: Alex Pacheco <apacheco at lsu.edu>
Date: Wed, Jun 4, 2014 at 8:43 PM
Subject: Fwd: [Allcct] Reminder: HPC user symposium keynote speech
"Programming for Hybrid Architectures Today and in the Future" on Thursday
at 8:30AM
To: Le Yan <lyan1 at tigers.lsu.edu>, Bhupender Thakur <bthakur at cct.lsu.edu>
------------------------------
*From:* Honggao Liu <honggao at cct.lsu.edu>
*Sent:* June 4, 2014 7:57:42 PM CDT
*To:* allcct at cct.lsu.edu
*Subject:* [Allcct] Reminder: HPC user symposium keynote speech
"Programming for Hybrid Architectures Today and in the Future" on Thursday
at 8:30AM
To ALLCCT:
>
> Mr. John Stone from University of Illinois, the keynote speaker for the 2014 LONI HPC user symposium, will give a talk at the DMC theater on Thursday morning at 8:30. The title is "Programming for Hybrid Architectures Today and in the Future”.
>
> Mr. Sone's bio and the abstract for the talk are attached below. You are encouraged to attend his talk and the LONI HPC user symposium on June 4-5. The schedule of the symposium can be find here: http://www.hpc.lsu.edu/users/symposium/index.php
>
> Bio
>
> Mr. John Stone is the lead developer of VMD, a high performance molecular visualization tool used by researchers all over the world. His research interests include molecular visualization, GPU computing, parallel
> processing, ray tracing, haptics, and virtual environments. Mr. Stone was inducted as an NVIDIA CUDA Fellow in 2010. He also provides consulting services for projects involving computer graphics, GPU computing, and high performance computing.
>
> Abstract
>
> Hybrid or heterogeneous computing architectures combine conventional multi-core CPUs with massively parallel microprocessors such as GPUs to create a platform that is well suited to the needs of many high performance computing workloads. Hybrid architectures have demonstrated significant benefits for application performance and also energy efficiency over conventional architectures. As of November 2013, several of the top ten fastest supercomputers in the world were based on hybrid many-core and GPU architectures, and all of the top ten systems in the Green500 list were hybrid GPU-accelerated systems. The most efficient Green500 system, Tsubame-KFC, achieved a record 4.5 GFLOPS/Watt. While it is clear that
> hybrid systems offer tremendous potential for increased performance and energy efficiency, they also pose challenges in terms of the adaptation of legacy application software to fully exploit massively parallel microprocessors such as those in GPUs. This presentation will outline the development and subsequent evolution of hybrid computing approaches in HPC, describe challenges and solutions for adapation of legacy software to current state-of-the-art hybrid computing platforms, and forecast hybrid computing technologies and changes to HPC workloads and usage models that are expected to impact the HPC user community in the next two to five years.
>
>
>
------------------------------
Allcct mailing list
Allcct at cct.lsu.edu
https://mail.cct.lsu.edu/mailman/listinfo/allcct
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
--
Le Yan, Ph.D.
Manager of User Services
High Performance Computing
Louisiana State University
Tel: 225-578-7524
Fax: 225-578-6400
AIM: lyanlsu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mail.loni.org/mailman/private/lasigma-gpu/attachments/20140604/c0c1277f/attachment.html
More information about the LASiGMA-gpu
mailing list