ST: CDA 6958
Multi-core/Many-core Architectures and Programming (Spring 2008)
Office Hours: 2:30pm-4:00pm F. at ENG3 243
Announcements
·
Homework
#1 is assigned (First due date: 2/1/08, Second due date: 2/8/08)
·
AMD/ATI
guest lecture #1 by Mike Mantor (slides)
·
AMD/ATI
guest lecture #2 by Justin Hensley and Jason Yang (slides part a, part b)
·
Lab set up for Nvidia G80 processors (by Jingfei)
·
Lab set up for Cell processors (by Hongliang)
·
AMD/ATI
Guest lecture #3 by Jason Yang (Brook+, AMD IL)
·
Homework
#2 is assigned (First due date: 2/15/08, second due date: 2/22/08, third due
date: 2/29/08)
·
Lab set up for ATI GPUs (by Martin)
·
AMD/ATI
Guest lecture #4 by Justin Hensley and Jason Yang (Performance
analysis, case study: H.264 and AES)
·
Homework
#3 is assigned (due date: 3/14/08)
·
Project
is posted (first due date: 3/14/08)
·
Hongliang’s
optimized matrix multiplication (code, makefile) using CUDA, which outperforms the
matrix multiplication algorithm from Nvidia CUBLAS by around 56% (168 GFLOPS including
data transmission between CPU and GPU for 4kx4k matrices).
·
Homework
#4 is assigned (due date: 4/20/08)
·
Project
Presentation Time: 9:30am-4:30pm, 4/25/08
·
Project Presentation Schedule
Related Courses in Other
Institutions:
Programming
Massively Parallel Processors (UIUC)
Multicore
Programming Premier: Learn and Compete Programming for the PS3 Cell Processors
(MIT)
Multicore
and GPU Programming for Video Games (GaTech)
Useful Links:
