Talks

2024

I/O Benchmarking

2023

Empowering Domain Experts through Automatic Code Transformation of HPC kernels

		climate ====

2022

		models, using in-situ measurements ====

2021

Long-term analysis of the North Sea light climate, using a 3D coupled ecosystem model

Automatic distribution of OpenMP kernels using CATO

2020

Coupled Storage System for Efficient Management of Self-Describing Data Formats (CoSEMoS)

Storage and I/O

R&D Line: Accelerators and Parallelization

Spack Package Manager: Introduction and Best Practices

2019

Storage and I/O

Domain-Specific Programming for Climate and Weather

Enhanced Adaptive Compression in Lustre

Optimizing Memory Bandwidth Efficiency with User-Preferred Kernel Merge

Scalable Parallelization of Stencils using MODA

		for Manycore or Accelerators (Pˆ3MA)|

ESiWACE, High-Resolution Demonstrators and Scalability

Project ESiWACE: Towards Global Kilometre-Scale Weather and Climate Predictions

Spack Community BoF

Managing HPC Software Complexity with Spack

Performance Modeling and Scalability for the ICON Model

Exploiting Architectural Capabilities using Higher Semantics

		Simulations ====

Automatic Vectorization of Stencil Codes with the GGDML Language Extensions

Advanced Computation and I/O Methods for Earth-System Simulations Status update

2018

Molecular-Continuum Simulation Software For Massively Parallel Multiscale Fluid Dynamics

High-Level Coding with GGDML and Code Optimization

Load Balancing and Auto-Tuning in Particle Simulations

PetaFLOP Molecular Dynamics for Engineering Applications

		distributed memory ====

		Methods for Nanoflow Simulation ====

Earth System Modeling with User-Controlled GGDML code Translation

		Translation ====

		2018|

Data-Intensive High-Performance Computing

		System Models ====

Towards Global Cloud-Resolving Weather and Climate Prediction at Exascale

Performance Predictions for Storm-Resolving Simulations of the Climate System

Convergence of High Performance Computing and Big Data

The ESiWACE Demonstrators: Scalability, Performance Prediction, Evaluation

Using Higher-Level Language Extensions to Support Earth-System Modeling

Using Higher-Level Language Extensions to Support Earth-System Modeling

Using Higher-Level Language Extensions to Support Earth-System Modeling

Convergence of Weather and Climate Simulation

Coupling Software for Massively Parallel Molecular-Continuum Flow Simulation

Performance Predictions for Storm-Resolving Simulations of the Climate System

Advanced Computation and I/O Methods for Earth-System Simulations Status update

		Simulations ====

Using Higher-Level Language Extensions to Support Earth-System Modeling

2017

Exploiting Weather and Climate Data at Scale (WP4)

PeCoH - Raising Awareness for Costs and Performance

The Virtual Institute for I/O and the IO-500

Establishing the IO-500 Benchmark

Transient Multi-Instance Molecular-Continuum Flow Simulation on Supercomputers

High Performance Computing and I/O

Towards new storage interfaces – chance or curse?

Empowering Scientists with Domain Specific Languages

Accelerating Storage System Research Through a Common Framework

PeCoH - Performance Conscious HPC

Anderson-Accelerated Spatial Coupling of Lattice Boltzmann and Navier-Stokes Solvers

IO-500 Status

Metadata Issues

JULEA: A Flexible Storage Framework for HPC

MPI-IO In-Memory Storage with the Kove XPD

Lustre at DKRZ

Fortrantestgenerator: Automatic and Flexible Unit Test Generation for Legacy HPC Code

		Scientific Compression Library (SCIL) ====

BoF: The Virtual Institute for I/O and the IO-500

Towards Performance Portability for Atmospheric & Climate Models with the GGDML DSL

Transient Molecular-Continuum Flow Simulation on 32 768 cores

Research Achievements and Perspectives for HPC and Earth Science

High Performance I/O

Big Data Analytics

Analysis of Input/Output Phenomena using Statistics and Machine Learning

Visual Analytics

The Case for a Flexible HPC Storage Framework

ESiWACE

		system modelling ====

Metadata/Small File Benchmarking and Status of the IO-500

Advanced Computation and I/O Methods for Earth-System Simulations Status update

AIMES: Status of the Project

AIMES Status Report: University of Hamburg

Enabling Transient Multi-Instance Molecular-Continuum Simulation on Supercomputers

Structure-Aware Adaptive Data Placement

Parallel I/O and Performance Evaluation

2016

TaLPas: Task-basierte Lastverteilung und Auto-Tuning in der Partikelsimulation

BoF: The Virtual Institute for I/O and the IO-500

Statistical File Characterization and Status Update: Monitoring at DKRZ

Middleware for Earth System Data

MPI-IO In-Memory Storage with the Kove XPD

BigStorage

Enhanced Adaptive Compression in Lustre

Lustre usage and compression at DKRZ

		Learning ====

Performance Conscious HPC (PeCoH) Kickoff

		Multi-Instance Sampling ====

Additional I/O Considerations

		Scientific File Formats and Compression Features ====

Molecular-Continuum Coupling Software: Flexibility, Parallelism, Multi-Instance Sampling

		Applications|

Lustre at DKRZ

		Scientific File Formats and Compression Features ====

I/O Benchmarking

BoF: VI4IO: The I/O Community Hosting the High-Performance Storage List

Steady-State and Transient Coupling of Lattice Boltzmann and Navier-Stokes Solvers

I/O Perspectives for Earth Science Data

		Parallelism, and Multi-Instance Sampling ====

Compression of Scientific Data

2018

		coastal hypoxia ====

		within an OSPAR frame ====

		Ecosystems (EUTRO2018)|

		ecosystems through element tracing: A case study for the northern Gulf of Mexico ====

2017

		nutrient transports for the ecological management of low oxygen conditions ====

		for the North Sea ====

2016

The influence of riverine nitrogen on the North Sea oxygen minimum zone

Exascale Software

Enhanced Adaptive Compression in Lustre

Storage expenses and data reduction techniques

Parallel distributed file systems

Advanced Computation and I/O Methods for Earth-System Simulations (AIMES)

AIMES overview, contributions of WR to WPs

Compression (in HDF5) in the projects ICOMEX and AIMES

2015

WP2: Efficient input/output of very large data sets

Interaktiver C Kurs (ICP)

		Boltzmann-Navier-Stokes Simulations ====

		and Information Technology of the University of Stuttgart|

SIOX: A flexible approach

Monitoring plans at DKRZ

Future CS Trends and IO

High-Performance I/O at Mistral

Molecular-Continuum Simulations in Fluid Dynamics: Methods, Challenges and Software

		Physics|

Teamwork Across Disciplines: High-Performance Computing Meets Engineering

Coupling 4 Molecular Dynamics Codes in a Massively Parallel Molecular-Continuum Framework

Exploiting Semantical Information for Performance Optimization and Data Reduction

Mistral Supercomputer I/O

Analyzing and Assessing I/O Performance

Predicting Performance of Non-Contiguous I/O with Machine Learning

I/O at the German Climate Computing Center

A Best Practice Analysis of HDF5 and NetCDF-4 Using Lustre

MaMiCo: Software Design for Massively Parallel Molecular-Continuum Flow Simulations

Dynamically Adaptable I/O Semantics for High Performance Computing

MaMiCo: Software Design for Massively Parallel Molecular-Continuum Flow Simulations

Coupling 4 Molecular Dynamics Codes in a Massively Parallel Molecular-Continuum Framework

High Performance Visual Computing: An Overview

Advanced Computation and I/O Methods for Earth-System Simulations (AIMES)

I/O Semantics for Future Storage Systems

		models: A case study for the North Sea ====

Domain-specific Language Development

2014

Challenges in HPC I/O

Mesh-to-Particle Coupling for Hybrid Molecular-Continuum Simulations

		Parallelisation ====

		Platforms for Atomic and Molecular Simulations|

		Energy Consumption ====

Multiscale Simulation Approaches for Micro- and Nanoflows

		(FISIOCOMP), Universidade Federal de Juiz de Fora|

A Modular Software for Parallel Hybrid Multiscale Flow Simulations

Parallel Particle Simulation in Multiscale Fluid Representations

Hybrid Multiscale Simulation Approaches for Micro- and Nanoflows

The SIOX Architecture -- Coupling Automatic Monitoring and Optimization of Parallel I/O

Deficiencies in I/O and the Future with E10

Exascale 10 -- High-Level Architecture

Compression By Default -- Reducing Total Cost of Ownership of Storage Systems

Introductions to the SIOX Architecture

Exascale Storage Systems -- An Analytical Study of Expenses

Exascale 10 -- High-Level Architecture

Advancing Independent Non-Contiguous I/O

Adaptive Mesh Refinement in Lattice Boltzmann Simulations

Wie aus Molekülen eine Strömung entsteht

Hochleistungsrechnen: Einblicke in Leistungsanalyse und Speichersysteme der Zukunft

I/O Interfaces Will Change -- Chance or Curse for Programmers?

ICOMEX

Challenges in Understanding, Optimizing and Procuring Storage Systems

		depletion in the North Sea ====

		Simulations ====

Monitoring, Simulation and Autonomous Optimization of Parallel I/O

2013

Guiding Parallel I/O -- The Semantical Gap

Spatial Coupling of Navier-Stokes and Lattice Boltzmann Solvers

Monitoring and Optimization of I/O-Performance with SIOX

Simulation Techniques for Large Molecular Systems

A Semantics-Aware I/O Interface for High Performance Computing

Using Simulation to Validate Performance of MPI(-IO) Implementations

Towards Intelligent Self-Optimisation in HPC I/O

Parameterising Primary Production and Convection in a 3D Model

		Hochleistungsrechnen ====

Massively Parallel Molecular-Continuum Simulations with the Macro-Micro-Coupling Tool

2012

Parameterising Primary Production and Convection in a 3D Model

Towards Flexible Adaptive Molecular-Continuum Simulations

EXIOS -- Unified Exascale I/O System

A Coupling Tool for Parallel Molecular Dynamics-Continuum Simulations

		Patterns ====

Fast Prototyping and Parallel Computations with Peano

Simulating Parallel Programs on Application and System Level

Tracing and Visualization of Energy-Related Metrics

The SIOX Project: Scalable I/O for Extreme Performance

A Semantics-Aware I/O Interface

Towards a Unified Exascale I/O System

Scientific Computing: Performance and Efficiency in Climate Models

		High Performance Computing ====

WiP: Simulating Application and System Interaction with PIOsimHD

Simulation-Aided Performance Evaluation of Server-Side Input/Output Optimizations

		Hardware Architecture ====

2011

eeClust: Energy-Efficient Cluster Computing

Energy-Efficient Cluster Computing

Energy-Efficient Cluster Computing

Simulations to the tides of ancient oceans and the evolution of the Earth-Moon system

Towards an Energy-Aware Scientific I/O Interface

Optimizations for Two-Phase Collective I/O

Visualization of MPI(-IO) Datatypes

Energy-Efficient Cluster Computing

Managing Hardware Power Saving Modes for High Performance Computing

A Flux-Based Approach to Hybrid Molecular Dynamics-Lattice Boltzmann Simulations

Energieeffizienzanalysen von parallelen Programmen

LB at TUM - Outtakes

Simulation der Entwicklung des Erde-Mond-Systems unter dem Einfluss ozeanischer Gezeiten

2010

Collecting Energy Consumption of Scientific Data

Simulation of Power Consumption of Energy Efficient Cluster Hardware

Coupled Approaches for Fluid Dynamic Problems Using the PDE Framework Peano

Auswirkungen von Windfeldänderungen durch Offshore-Windparks auf das marine Umfeld

Ansätze zum Monitoring hierarchischer E/A-Systeme

Energieeffizienz und Hochleistungsrechnen

2009

Tracing Internal Communication in MPI and MPI-I/O

Using Non-blocking I/O Operations in HPC to Reduce Execution Times

2008

Bottleneck Detection in Parallel File Systems with Trace-Based Performance Monitoring

Directory-Based Metadata Optimizations for Small Files in PVFS

Werkzeuge zur Analyse der Speicher- und Energieeffizienz bei Hochleistungsrechnern

Performance Visualization of MPI-IO Operations in Parallel Programs

2007

Performance Evaluation of the PVFS2 Architecture

2006

Performance Analysis of Parallel File Systems

File Systems for Mass Storage of Image Data in Bioinformatics

Migration and Load Balancing Within PVFS2