Speakers
Description
New technologies (non x86_64, GPU, FPGA, APUs, QC) need large SW development investment. How will you handle the effort (manpower, common projects)? Does the effort suffice? Are there high risks of failure to port SW to new platforms?
Use of ML techniques is growing in many science domains, including HEP. How do you see the evolution of resource needs for specialized AI architectures? Would access to HPC sites resources help?
There are many cross-domain forums, advisory groups, bodies, consortia (at least in EU) promoting and shaping HPC future (competence centers, centers of excellence, …). Would WLCG benefit to join them? US, EU, Japan, other countries, is there an organized effort to support HEP on HPCs?
Clouds, what are the expectations for public and private cloud usage in the future?
