Data-Intensive Computing and AI/ML Applications at Scale
→
America/Chicago
Room 1501 (Building 240)
Room 1501
Building 240
Description
With Aurora to enter production in the coming months, a new era of computing will dawn at Argonne. Not only is Aurora one of the world's largest supercomputers, it is also designed to support three overlapping activities -- traditional high-performance computing, data-intensive problems, and large-scale AI/ML applications. This meeting will bring together researchers working primarily on the last two topics. The aim is to understand the current scope of activities, the nature of hardware/software challenges going forward, and how to overcome them via collaborative approaches. Additionally, the meeting will also be useful in capturing the status of the current software suite, data-intensive application support services, and what may be required by applications in the future.
Registration
Registration
41
Participants
-
-
1
Informal Introductions/Registration
-
2
IntroductionSpeakers: Salman Habib (CPS), Venkatram Vishwanath (LCF)
- 3
-
4
Toward DMC Accuracy Across Chemical Space with Scalable Δ-QMLSpeaker: Anouar Benali (CPS)
-
5
Data Management, I/O, and scaling for AISpeaker: Huihuo Zheng (LCF)
-
6
Scientifically motivated LLM methodsSpeaker: Azton Wells (CPS)
- 7
-
8
Scalable diverse workloads on AI acceleratorsSpeaker: Varuni Katti Sastry (LCF Division)
-
9
Scalable Graph Neural Networks for Mesh-Based Fluid Flow ModelingSpeaker: Shivam Barwey (LCF Division)
-
10
Data servicesSpeaker: Aditya Tanikanti (LCF)
-
11
Automatic MRI segmentation for upper limb muscles for clinical applicationsSpeaker: Marta Garcia Martinez (CPS)
-
10:21 AM
Coffee Break
-
12
Machine Learning Potentials for Molecules and MaterialsSpeaker: Murat Keceli (CPS)
-
13
Solving the quantum many-body problem with neural network quantum statesSpeaker: Alessandro Lovato (PHY & CPS Divisions)
-
14
Data-driven modeling in CosmologySpeaker: Nesar Ramachandra (CPS)
- 15
-
16
Workflow tools for large scale job campaigns at ALCFSpeaker: Christine Simpson (ALCF)
-
17
HMC with change of variables using ML and scaling upSpeaker: Xiaoyong Jin (CPS)
-
18
AI/ML for generative models and acceleration of scientific workloadsSpeaker: Filippo Simini (LCF)
-
19
Scaling LLMs for Science and Ongoing CollaborationsSpeaker: Sam Foreman (LCF)
- 20
-
21
Integrated Research Infrastructure for DOE / Brain reconstruction from electron microscopy imagingSpeaker: Thomas Uram (LCF)
-
22
Interfacing cosmological simulations and observationsSpeaker: Patricia Larsen (CPS)
- 23
-
24
Coupling MPI with Dask in SituSpeaker: Amal Gueroudji (Argonne)
-
25
Scalable Learning on ALCF SystemsSpeaker: Corey Adams (LCF)
-
11:15 AM
Coffee Break
-
26
Open Discussion
-
12:00 PM
Working Lunch, Collect boxed lunches
-
27
Subgroup Discussions
-
1