BEGIN:VCALENDAR VERSION:2.0 PRODID:Linklings LLC BEGIN:VTIMEZONE TZID:Europe/Stockholm X-LIC-LOCATION:Europe/Stockholm BEGIN:DAYLIGHT TZOFFSETFROM:+0100 TZOFFSETTO:+0200 TZNAME:CEST DTSTART:19700308T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:+0200 TZOFFSETTO:+0100 TZNAME:CET DTSTART:19701101T020000 RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20210916T132446Z LOCATION:Henry Dunant DTSTART;TZID=Europe/Stockholm:20210705T133000 DTEND;TZID=Europe/Stockholm:20210705T140000 UID:submissions.pasc-conference.org_PASC21_sess104_pap128@linklings.com SUMMARY:Stream-AI-MD: Streaming AI-driven Adaptive Molecular Simulations f or Heterogeneous Computing Platforms DESCRIPTION:Paper\n\nStream-AI-MD: Streaming AI-driven Adaptive Molecular Simulations for Heterogeneous Computing Platforms\n\nBrace, Salim, Subbiah , Ma, Emani...\n\nEmerging hardware tailored for artificial intelligence ( AI) and machine learning (ML) methods provide novel means to couple them w ith traditional high performance computing (HPC) workflows involving multi -scale molecular dynamics (MD) simulations. We propose Stream-AI-MD, a novel instance of applying deep learning methods to drive adaptive M D simulation campaigns in a streaming manner. We leverage the abi lity to run ensemble MD simulations on GPU clusters, while the data from a tomistic MD simulations are streamed continuously to AI/ML approaches to g uide the conformational search in a biophysically meaningful manner on a w afer-scale AI accelerator. We demonstrate the efficacy of Stream-AI-MD simulations for two scientific use-cases: (1) folding a small protot ypical protein, namely ββα-fold (BBA) FSD-EY and (2) under standing protein-protein interaction (PPI) within the SARS-CoV-2 proteome between two proteins, nsp16 and nsp10. We show that Stream-AI-MD simulations can improve time-to-solution by ~50X for BBA protein folding.& nbsp;Further, we also discuss performance trade-offs involved in implement ing AI-coupled HPC workflows on heterogeneous computing architectures.\n\n Domain: CS and Math, Life Sciences END:VEVENT END:VCALENDAR