Hybrid SORS: How to have a Bad Research Visiting Experience/Internship

Abstract: Are you thinking of doing a research stay at another lab? Come to this talk where I will talk about my recent experience at the University of California, Santa Barbara. I will explain what I have done in the last 4 months including the research done, travels in California, American experiences, and differences with BSC labs. During the talk, I will try to give simple tips that you should avoid to have a successful experience (or not).
The research done at UCSB involves rethinking the efficiency of the network of a data centre with a creative alternative to support the new PB datasets. Also, I will explain my recent paper accepted at DATE 23 called Metro-MPI about parallelizing RTL simulations using MPI. Finally, I will explain a new paradigm of connecting coherent hardware accelerators to processors using standard software queues.

Short bio: Guillem López is a second-year PhD student at BSC working in improving the connection of hardware accelerators and traditional CPUs. His advisors are Miquel Moretó and Adria Armejach. In addition, he also collaborates internationally with Jonathan Balkind (UCSB), Stefan Wallentowitz (HM) and Balaji Venu (ARM). At BSC, he actively participates in different projects such as DRAC, EPI, MEEP and the Arm Center of Excellence, developing RISC-V CPUs and helping in different tapeouts. He has some papers at international conferences (DATE 23’, DSD 22', DATE 21', ICPP 21’, DCIS 20', ACACES). In 2021, he received an FI grant in 2021 and was accepted at the Google Summer of Code 21’, which is a very competitive international program by Google to develop software for open-source projects (several thousand applicants, 20% acceptance rate). Additionally, he has received funding from ARM to do research with leading-edge technologies. Finally, whenever he is not doing research he enjoys going to the mountains to climb, ski or do any other related activity.