The final decisions have been made on technical papers for SC18, and the entire program is now online. This year, the committee adopted a new review process, in which some papers were selected for major and minor revisions. The new process combines the timeliness of traditional SC conference submissions with the rigor of journal review.
Overall, 68 papers out of 288 submissions were accepted. Fifty-five were accepted in the first round, some with a minor revision. The authors of 15 more submissions were invited to submit a major revision of their work, in cases when the committee was not able to make a confident decision. Of these, 13 were ultimately accepted.
The committee reported that the revision process improved the overall quality of the papers, and that they were more confident in their decisions as a result of the revisions. The new process allowed us to accept a slightly larger fraction (23.6%) of the papers than in the past years, without compromising quality.
As in past years, the committee selected some of the top papers as finalists for the best paper and best student paper awards.
The best paper candidates are:
- Exploiting Idle Resources in a High-Radix Switch for Supplemental Storage
- Extreme Scale De Novo Metagenome Assembly
The best student paper candidates are as follows:
- Doomsday: Predicting Which Node Will Fail When on Supercomputers
- Distributed-Memory Hierarchical Compression of Dense SPD Matrices
- HiCOO: Hierarchical Storage of Sparse Tensors
- Mitigating Inter-Job Interference Using Adaptive Flow-Aware Routing
- Cooperative Rendezvous Protocols for Improved Performance and Overlap
We congratulate all authors of accepted papers and award finalists! We also thank all submitters, committee members, and track chairs for making SC18 a success!