Last week I attended the 28th International Conference on Architectural Support for Programming Languages and Operating Systems, which was held in Vancouver, Canada. ASPLOS is the premier forum for interdisciplinary systems research, intersecting computer architecture, hardware and emerging technologies, programming languages and compilers, operating systems, and networking.
The conference featured talks on various research topics, covering both classic topics like memory consistency, compiler techniques and optimization, and virtualization, as well as emerging topics such as sustainability and quantum computing. I was particularly interested in two sessions that had talks on topics related to my research area, namely disaggregated memory systems and persistence. Two of the talks were about Compute Express Link (CXL), a new memory fabric technology that enables the disaggregation of memory from compute, which can greatly improve the performance of workloads that need a lot of data.
There were three keynotes one for each day of the conference. On conference opening day, Azalia Mirhoseini shed light on the important role that generative AI itself can play in designing the next generation of computing systems and hardware that in turn would fuel the next generation of AI breakthroughs. She talked about how a generalizable deep reinforcement learning method for chip floorplanning reduced the design time for Google TPUs. She also discussed the opportunities and challenges for future computing systems in the era of large generative models. On the second conference day, Abhishek Bhattacharjee talked about his group’s work on a standardized and general computer architecture for future brain interfacing, and painted a broad vision for how such an architecture can enable the treatment of several neurological disorders, most notably, epilepsy and movement disorders. Finally, on the third and final day’s keynote, Bryan Catanzaro discussed the work they have been doing at NVIDIA to optimize systems for Large Language Model training and inference, and highlighted some of the challenges that remain for future work.
On a more personal note, it was a joy to see my brother Stavros and his ex-colleagues from EPFL receive the Influential Paper Award for their paper Clearing the clouds: a study of emerging scale-out workloads on modern hardware, which they published at ASPLOS 2012.
Overall, I felt thrilled to attend the conference, as it was my first post-COVID event where I could interact with people in person. I enjoyed catching up with old friends and meeting new people. I found the conference quite energetic with lively discussions in the sessions or in the hallways. After the conference, I visited Simon Fraser University in the neighboring city of Burnaby. There, I enjoyed meeting Tianzheng Wang and his students, and presenting our work on persistent scripting.