ASPLOS

It is our great pleasure to welcome you to the Eighteenth International Conference on Architectural Support for Programming Languages and Operating Systems --- ASPLOS 2013! This year's conference continues and reinforces the ASPLOS tradition of publishing innovative multidisciplinary research spanning the boundaries of hardware, computer architecture, compilers, languages, operating systems, networking, and applications.

This year, authors submitted 193 papers, 12% more than last year when 172 papers were submitted, and 7% more than in 2010, the previous record year, when 180 papers were submitted. PC members co-authored 22 submissions. The following ten topic keywords were marked by at least 20 submissions: compilation, parallelization and optimization (40 submissions), programming models (32), OS scheduling and resource management (31), OS abstractions (29), testing and debugging (27), power/energy/thermal management (27), memory optimizations (25), heterogeneous architectures (24), parallel programming languages (23), and virtualization (21).

The 2013 ASPLOS Program Committee (PC) consisted of 35 members; each member was assigned 18 submissions, a workload of four more papers than the previous year. As in previous years, the Program Chair recruited an External Review Committee (ERC) whose expertise complemented that of the PC. The 2013 ERC consisted of 63 members, assigned 5 papers each. Additionally, 90 sub reviewers prepared 143 reviews. In total, we received 943 reviews, an average of 4.89 reviews per paper. 90% of submissions received at least five reviews, and all received at least four.

As in previous years, we employed two rounds of reviews. A significant difference this year was that no papers were rejected after the first round. The two phases were employed merely to spread out the review workload and to allocate reviewer expertise based on first-round reviews. After a three-day author response period, the PC and ERC members entered online discussion lasting about a week. The online discussion, moderated by paper discussion leaders assigned to each paper, was unusually lively. The discussion resulted in 71 papers being selected for deliberation at the physical PC meeting. During the entire review and selection process, scores alone were never used to reject a submission.

The PC gathered in Berkeley, CA, on Nov. 1, 2012, for a one-day meeting. Three committee members were unable to attend, mostly due to travel disruptions from the Hurricane Sandy. These members participated in online discussions prior to the meeting. The meeting followed the protocol of allocating a fixed amount of time for the discussion of each paper. If a consensus was not reached, the paper advanced to a discussion phase, which had the form of small groups composed of paper reviewers and other relevant PC members. In the late afternoon, these groups reported back to the PC with arguments for or against accepting the paper. PC and ERC members volunteered to shepherd 7 papers to ensure reviewer comments were properly addressed. The PC-authored submissions were not discussed at the meeting (see below). PC members with conflicts left the room during discussions, as is customary. Vivek Sarkar, the General Chair, attended the PC meeting as an observer and handled conflicts involving the Program Chair.

In the end, the PC and ERC accepted 44 papers, 7 more than the previous year. The acceptance rate is 22.8%, compared to 21.5% the previous year.

The 22 PC-authored submissions were reviewed only by the ERC. The ERC discussed and accepted these submissions in an online discussion, conducted mostly prior to the physical PC meeting. The ERC held the PC submissions to the same conference standard as other papers. Preliminary acceptance decisions, done by the ERC before the PC meeting, were calibrated against the papers accepted at the PC meeting with the help of the Program and General Chairs. In the end, 5 PC submissions were accepted. The acceptance rate was the same as for non-PC submissions.