Unleashing the Power of Parallel Computing with Commodity Radeon 5800 Series GPU!

SpeakerMichael Mantor, AMD Senior Fellow / GPU Compute Architect
Organization AMD Graphics Product Group
Location1021 EB2
Start Date November 20, 2009 2:20 PM
End Date November 20, 2009 3:10 PM

This talk will present a brief introduction to the new RadeonTM 5800 Series Architecture with an emphasis on the compute aspects of this product.  We will cover compute dispatch capabilities, ALU architecture, features and performance, shared memory architecture and shader I/O memory hierarchies.  Brief update on Dx11 DirectCompute and OpenCl software APIs.

About the speaker:  Mike Mantor has been developing software and hardware for computer graphic and imaging systems for 20+ years.  From high end proprietary military imaging systems to commercial development of commodity graphics for chips such as Intel 740, ATI/AMD RadeonTM Series of GPUs and game console systems such as the Xbox 360.  He has been a key contributor in texture subsystems, rasterizers, ATI's first programmable vertex shaders and ATI/AMD's Unified Programmable Shaders systems with a heavy emphasis in the development of the ALU units. Mike has accumulated over 20 issued patents and has spoken at several leading universities such as MIT, Stanford, UCF and leading industry technical talks at Graphics Hardware Conferences and Hot Chips.

  November 2009
Sun Mon Tues Wed Thu Fri Sat
1234567
891011121314
15161718192021
22232425262728
2930