Proceedings of the 44th Annual IEEE/ACM International Symposium on Microarchitecture

This year, we present to you a technical program consisting of 44 technical papers, selected from among 209 submissions by a Program Committee composed of 43 active researchers and practitioners in our field. The challenge of selecting which papers to accept among the 209 submissions was significant. To support this selection process, we solicited at least four reviews for each paper, and the vast majority of the papers had five reviews. A typical paper was reviewed by three program committee members and two other experts in our field. Overall, 1018 reviews were generated (on average, 4.9 reviews per paper), 629 of which came from the program committee members and 389 from experts outside of the program committee. For all papers but one, at least two of the reviewers indicated that they are at least Knowledgeable: know most if not all of the relevant work, understand the problem very well, and for 81% of the papers the average level of expertise among its reviewers was at least Knowledgeable. This year we tried a new approach to how author responses to reviews are handled - we asked both PC and external reviewers to read all of the reviews and the authors responses, and provide a separate post-rebuttal score that will be used to decide which papers will be discussed in the program committee meeting. We are especially grateful to our external reviewers for doing this - some of them reviewed and then graded four papers, which makes their workload about 25% of that assigned to the average PC member. We introduced a set of detailed formatting rules that made it possible to set a reasonable limit of 28 pages for submissions. Our goal was to reduce the variance in the amount of work each PC member and each reviewer had to perform while encouraging authors to succinctly present their work.