Come for Table discussions, Member Self-Intro, What's New, Application Showcase, and Advanced Application Development Techniques! Exchange ideas, meet experts, share code... all HPC & GPU, all practical, all cutting-edge.
Agenda:
General Discussions:
6:15-6:35pm What’s new
6:35-6:45pm Member self-intros: 30 seconds for each member
Main Program:
6:45-7:45pm OpenCL for FPGAs: Overview and Early Successes (Alex Grbic, Product Marketing Director, Altera)
7:50-8:30pm CUDA on ARM (Don Becker, Engineering, NVIDIA)
Refreshments sponsored by Altera
________________________________________________
OpenCL for FPGAs: Overview and Early Successes
Alex Grbic and Nick Finamore, Altera Corporation
OpenCL continues to grow in popularity as a parallel programming language for heterogeneous systems due to its distinction as an open standard and its ability to target a variety of acceleration devices. To enable powerful system acceleration, Altera has developed an OpenCL for FPGAs program, which combines the parallel performance capability of FPGAs with the OpenCL standard. The OpenCL for FPGAs program enables a broad segment of designers in a variety of end markets including high-performance computing, military, medical and broadcast to dramatically increase the performance of their end systems using the latest generation of FPGAs and a highly productive development flow. Altera is currently engaged with a variety of customers to implement designs in FPGAs using an OpenCL flow. In this talk, we will describe Altera’s OpenCL for FPGAs tool and target application domains.
Alex Grbic Bio: Alex Grbic is a Product Marketing Director at Altera. He is responsible for Software, DSP and IP products, which includes Altera’s OpenCL for FPGAs program. At Altera, Alex has held management positions in the Software and IP R&D organization, where he headed up external memory interfaces IP development, performance analysis of Altera’s products and work on visualization/debug tools. More recently, he was Director of Applications Engineering in the Technical Services Department, responsible for escalated customer issue support, initiatives for early adoption of new products, and technical collateral. Alex holds a Ph.D. in Computer Engineering from the University of Toronto.
________________________________________________
CARMA CUDA on ARM Architecture: Developing Accelerated Applications on ARM
NVIDIA’s Donald Becker has a long history in compilers, device drivers, networking, parallel computing and cluster computing. He co-founded the Beowulf Project at NASA's Goddard Space Flight Center in 1994 and received the Gordon Bell Prize in 1997. He was a major developer of the early Linux kernel networking subsystem, writing essentially all of the network device drivers through 2000.
Best meetup I've attended. Excellent speakers. Another hardware talk but from a perspective useful for software developers.
My conclusion: performance and power efficiency are not free. Savings and gains in hardware are often offset by higher risk and costs in software. No free lunch (although the meetup has free pizza and drinks - good stuff too).
September 18
Good session. The two talks are closely related for the future of embedded, although the second talk seemed to emphasize HPC cloud services. I did not have the time available to go this time but made it anyways and was glad I put the extra effort in to attend.
September 17
Just found this group and am new to the area. I'm curious about the meetups. Is there an agenda? What goes on during the meeting? How long does it last? Are there any parallel compiler centric or gpu compiler centric discussions?
August 17
Join or login to comment.