Professor for computer science at @unipb. Head High-Performance Computing group @hit_upb and director of Paderborn Center for Parallel Computing @pc2_upb.
At @unipb we are currently seeking to full an Endowed Heinz Nixdorf Full Professorship for "Algorithms and Software Development for Quantum Computers", intended to create a link between High Performance Computing and Quantum Systems at @pc2_upb, PhoQS & @HNI_PB. #Academicjobs
🌿💻Nicht verpassen: Im #cleanIT openXchange Talk spricht Prof. Christian @plessl von der @unipb heute ab 16 Uhr über "Approximate Computing – A Paradigm for Energy Efficient Computations". Den Zoom-Link gibt's über das clean-IT-Forum auf @openHPI: open.hpi.de/courses/cleanit2…#SDGs
DYK @unipb researchers reached the computing milestone of one quintillion operations per second (exaops) when they used @NERSC's Perlmutter #HPC to simulate the COVID-19 spike protein using a novel mixed-precision method?
More: bit.ly/3TRUZtk#ExascaleDay@BerkeleyLab
@LeeEdwardE published a great article on the destructive rejection culture at CS conferences:
sigbed.org/2022/08/22/the-to…
His point is very valid. The primacy of novelty over interestingness has promoted countless articles that are boring or inconsequential. 1/2
The next application deadline for PhD scholarships in the area of HPC funded by @nhr_germany is on 15 Sep.
For conditions and topics see: nhr-verein.de/application
If you are interested in joining @pc2_upb for your PhD please consider applying. And feel free to spread the word.
Scholarship topics for @pc2_upb
1) Sparse linear algebra with FPGAS
2) Scalable and performance portable scientific FPGA applications
3) Methods for atomistic simulations on modern comp arch
4) Electronic structure theory on quantum computers.
Details: nhr-verein.de/sites/default/…
This is very exciting! Bringing us one step closer to reconfigurable fabrics in general-purpose CPUs. And mixed-precision data paths can have much broader applications in scientific computing than AI inference.
Electricity Map shows the real time
electricity production broken down by source and CO2 intensity. It’s surprisingly fun to poke into the data.
apps.apple.com/de/app/electr…
We are thrilled to join the AMD Heterogeneous Accelerated Compute Clusters (HACC) initiative.
Our mission will be to advance the use of FPGA in HPC with improved system integration, toolflows and scalable FPGA-accelerated libraries and applications.
linkedin.com/pulse/announcin…
Moreover, compute bound codes are easily encapsulated into a library (eg LAPACK) due to narrow data interface, but bandwidth bound & latency bound codes are difficult. That is why even to this day many latter codes have not been ported into GPUs which should be rather standard.