Scientific Computing
Institute for Scientific Computing
Scientific Computing
The Institute for Scientific Computing works on various aspects of computational sciences, with a focus on high-performance computing, performance engineering, and algorithmic differentiation. Our research includes tools and methodologies to support domain scientists in developing correct and efficient applications.
The Institute
Scientific Computing was installed as part of the Department of Computer Science when Professor Dr. Bischof was called to the Technical University of Darmstadt. As part of this chair we offer teaching activities on parallel computers, parallel programming, and computer science foundations of simulation technology.
An overview of our research activities is available here. Furthermore, we are always interested in students looking for thesis topics and research assistants.
Office Location
You find us in the Karl Plagge-Haus at campus Stadtmitte, building S1I22, Alexanderstraße 2, 64283 Darmstadt.
Research Members
@ SC
NHR PerfLab Seminar: Performance Measurements in HPC
In this NHR PerfLab Seminar we deep-dive the technology concepts of current performance analysis tools to gain a critical understanding of their capabilities and limitations, and the overheads and perturbation involved. This insight enables one to utilize these tools to their fullest potential in order to write efficient code and optimize HPC programs.
Conference talk online
Our conference presentation of MPI-CorrBench from this year's HPDC is now available on youtube.
Paper accepted
Our paper “Automatic partitioning of MPI operations in MPI+OpenMP applications” was accepted at the C3PO Workshop, held in conjunction with ISC 2021.
Paper accepted
Our paper “Automatic low-overhead load-imbalance detection in MPI applications” was accepted at EuroPar 2021.
Paper accepted
Our paper “MPI-CorrBench: Towards an MPI Correctness Benchmark Suite” was accepted at the ACM Symposium on High-Performance Parallel and Distributed Computing (HPDC).
FOSDEM Talk Video online
The video recording of our FOSDEM 2021talk is now available here.
FOSDEM Talk
In a Talk in the HPC, Big Data and Data Science devroom at this year's FOSDEM, we present our recent open source research tools. The talk is scheduled for Sunday, February 7th at 16:30 (CET).
Paper accepted
Our paper “MetaCG: Annotated call-graphs to facilitate whole-program analysis” was accepted at the Workshop on Tools for Automatic Program Analysis (TAPAS).
Paper accepted
Our paper “Developing Models for the Runtime of Programs With Exponential Runtime Behavior” was accepted at the International Workshop on Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems (PMBS’20).
Paper accepted
Our paper “Towards Compiler-Aided Correctness Checking of Adjoint MPI Applications” was accepted at the Correctness 2020 Workshop. This is a collaboration with the IT Center at RWTH Aachen.
Paper accepted
Our paper “Leveraging Hybrid Cloud HPC in Multitier Reactive Programming” was accepted at the SuperCompCloud Workshop of Supercomputing. This is a collaboration with the group of Prof. Dr. Guido Salvaneschi.
PIRA v0.2 now on github
Version 0.2 of the automatic instrumentation refinement tool PIRA is now available on our github
qTESLA now on github
Implementations of a quantum-attack resistant signature scheme are now available on our github
SimAnMo now on github
The performance modeling tool SimAnMo is now available on our github!
Conference presentation recordings available
The recordings of our presentations at the C3PO workshop are now available on YouTube.
Compiler-assisted type-safe checkpointing
Automatic detection of MPI assertions
Paper accepted
Our papers “Automatic detection of MPI assertions” and “Compiler-assisted type-safe checkpointing” are accepted as full papers at the C3PO workshop.
Paper accepted
Our paper “A Comparison of the Scalability of OpenMP Implementations” is accepted as full paper at the EuroPar conference.
Now on github!
The Institute for Scientific Computing now maintains a public repository of the software tools developed in our group. You can find it at github.com/tudasc
High-Performance Computing in Hesse
The video explains how high-performance computing at TU Darmstadt helps researchers not only at TU Darmstadt but within Hesse in general.
SFB CROSSING & HKHLR – High Performance Computing at TU Darmstadt (youtube)
Invited talk by Thorsten Höfler
Torsten Höfler from ETH Zürich gave an interesting talk on proper benchmarking. The recording of this talk can be downloaded here.