There seems to be an ongoing attempt by spammers to get access to this blog.  Dozens of registrations with simple alphanumeric user names (such as wi784568@yahoo.com)  keep coming in.  I will continue to keep comments are turned off and posting privledges tightly secured, but I may also have to turn off new logins, mostly so I don’t have to wade through them.  My apologies for all of the restrictions.

An Information Week blog discusses the ‘Multicore Crisis’: Intel Blog Warns Of Multicore Crisis

From EDN:  Intel backs multicore effort with $10M investment in Endeca

In the EE Times article Intel CTO presses software developers to keep pace, Intel CTO Justin Rattner complains about the slow pace of software application development for new multicore devices for PCs.  Rattner complains that “innovation seems to have moved on to the handheld [market]“, where issues such as power efficiency are helping to spur multicore acceptance.

While specific to game software development, Emergent Game Technology has released its Floodgate technology which makes use of the multicore hardware in game consoles such as the PS3 and XBox.  An interview with Gamasutra answers a few questions about Floodgate:  Emergent’s Dan Amerson Opens The Floodgates

Jack Ganssle has a nice article on what 1,000,000 million lines of code looks like.  Not directly related to multicore (unless you are looking at legacy code that needs to be ‘multicored’):  A Million Lines of Code

A fairly practical, mainstream article about multicore processors from Electronic Design magazine:  Multicore Projects Mean Multiple Choices

Google Research has an older paper that has been generating some interest as a technique for programming large scale multicore devices.  This is the Map-Reduce technique as described in the paper:  MapReduce: Simplified Data Processing on Large Clusters.  Interestingly, it refers to scans and parallel prefix operators popularized in the 1980s on machines such as the Connection Machine from Thinking Machines.  These same parallel operators were later reapplied to FPGA-based reconfigurable architectures in the 1990s.

American Scientist has an excellent article on multicore devices and parallel programming titled Programming in a Parallel Universe.  It ties back to some of the more interesting work of the 1980s, including Hillis’ Connection Machine (a favorite of mine).

The Many-core and Reconfigurable Supercomputing Conference (MRSC) will be held at Queen’s University in Belfast, Northern Ireland on April 1 – 3, 2008.  While multicore architectures can be viewed as an extension to traditional multiprocessors, it can also be very useful to view these architectures as reconfigurable systems.  This perspective will make MRSC an interesting conference for anyone looking to explore multicore architecture and particularly software issues more deeply.