Back to Conference View
Login

Workshop on Mixed-Precision Computation

→ America/Chicago
Bldg 201, Rm 102

Bldg 201, Rm 102
Esteban Rangel (CPS), Kevin Harms (LCF)
Description

The future generation of computational hardware has added significant computational capability in precisions lower than double precision (64 bit). Modern high-end GPUs have also started to reduce the silicon area dedicated to 64-bit computation, resulting in stagnation or reductions in high-precision performance on a per-GPU basis. In order for high performance computation codes to continue performance improvement on new hardware, application developer must consider the user of lower precision computation. This workshop is an introduction into topic of mixed precision in HPC with presentations on application usage, numerics, and tools.

The workshop is planned to be series of a few workshops with the initial workshop focused within the Argonne computing staff and then growing to larger workshop involving the wider community and vendors.

Mixed_Precision-draft.pdf
Admin Contact: Samantha Tezak
Registration
Registration Form
Register
Participants
41 View full list
Surveys
Attendee Survey: October 2025 Workshop on Mixed-Precision Computation
Fill out
    • 9:30 AM → 9:40 AM
      Welcome & Goals 10m
      Speaker: Salman Habib (CPS)
    • 9:40 AM → 10:00 AM
      Motivation 20m
      Speaker: Paul Hovland (Argonne National Laboratory)
    • 10:00 AM → 10:15 AM
      A Record-and-Replay Workflow for Floating-Point Analysis of GPU Kernels 15m
      Speaker: Esteban Rangel (CPS)
    • 10:15 AM → 10:30 AM
      Half Precision Wave Simulation 15m
      Speaker: Longfei Gao (LCF)
    • 10:30 AM → 11:00 AM
      Coffee Break 30m
    • 11:00 AM → 11:15 AM
      Mixed-Precision Quasi-Newton Methods for Optimization 15m
      Speaker: Sven Leyffer (Argonne National Laboratory)
    • 11:15 AM → 11:30 AM
      What Makes Mixed Precision in PETSc Hard? Attempts, Failures, Successes, and Hopes 15m
      Speaker: Junchao Zhang (MCS)
    • 11:30 AM → 12:00 PM
      Roundoff error prediction and correction using CENA and Poseidon 30m
      Speaker: Jan Hückelheim (Argonne National Laboratory)
    • 12:00 PM → 1:00 PM
      Lunch 1h
    • 1:00 PM → 2:00 PM
      Group Discussion: Next Generation Architectures 1h
    • 2:00 PM → 3:00 PM
      Group Discussion: Methods and Tools 1h
    • 3:00 PM → 3:15 PM
      Workshop Planning and Next Steps 15m
      Speaker: Salman Habib (CPS)