PACT

It is our pleasure to welcome you to the 21st International Conference on Parallel Architectures and Compilation Techniques (PACT 2012) in Minneapolis, Minnesota, the Land of Ten Thousand Lakes.

This year the conference received 207 complete paper submissions, which is almost exactly the same number received last year. These papers were rigorously reviewed by a combination of a 42-member technical program committee and more than 300 additional external reviewers. More than 800 total reviews were submitted with a typical program committee member personally reviewing 11-12 papers each. After almost all of the reviews had been received, the authors were given several days to read and respond to the reviewers' comments. The program committee members evaluated these responses during an online discussion period before the formal program committee meeting.

On June 9, 2012, the program committee met in Minneapolis for a full day's discussion of the submitted papers. Committee members who had conflicts-of-interest with papers as they were brought up for discussion were asked to leave the room. After thorough discussions of the reviews and author responses, 39 papers were selected for the main conference program. The final acceptance rate of 19% was similar to previous years. The authors of an additional 41 papers were invited to submit posters of their papers during a special poster session at the conference, and to prepare an extended abstract for publication in the conference proceedings. A total of 25 authors accepted this poster invitation.

We are pleased to have three distinguished keynote speakers for the conference, two of them are from academia. They are Dr. Bill Cramer from the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign and Dr. Kathy Yelick (University of California, Berkeley, also with Lawrence Berkeley National Laboratory). One keynote speaker is from industry, Peter J. Ungaro from Cray. They provide their insights into the future directions of parallel computing.