Mastodawn

Good news for researchers! The IWOMP 2026 Call for Papers deadline has been extended to June 12. This year's theme — "OpenMP: Adaptability for Heterogeneous Multi-Device Systems" — covers accelerated computing, performance portability, machine learning, tasking, and more. Don't miss your chance to present at one of HPC's premier workshops.

https://www.iwomp.org/call-for-papers/
#OpenMP #HPC #MachineLearning #tasking #concurrency

Christos Argyropoulos 2d ago

The joy of being able to leverage contiguous memory access and tiling to x3 performance in the #openmp code is somewhat dulled, but not extinguished by the apparent failure to get the compiler to listen to my "#pragma omp simd" requests .
Time to move to portable intrinsics through the #SIMDe (SIMD everywhere) library

OpenMP ARB May 26

The IWOMP 2026 Call for Papers has been extended to June 5th!

The 22nd International Workshop on OpenMP takes place October 7-9, 2026 at TU Wien in Vienna, Austria. The theme this year is "OpenMP: Adaptability for Heterogeneous Multi-Device Systems."

Learn more and submit: https://www.iwomp.org/call-for-papers/
#IWOMP2026 #OpenMP #HPC #concurrency

Show thread

Christos Argyropoulos MD PhD May 24

@necrophcodr @openmp_arb Vulcan requires a lot of boilerplate (last time I checked). If one can get Openmp to deliver close to native framework performance, it is a lot cleaner to use #openmp . But if one's card is not supported by CUDA/ROCm (old cards are very useful for things like predictions from stat models , as one only needs fp32 compute and the calculation paths are short) then offload becomes difficult unless one can get the RDNA1 to cosplay as a RDNA2 in clang as I do

Christos Argyropoulos MD PhD May 24

Kind of nuts that one can make all of these cards for offloading with clang (AMD requires cosplaying as a different architecture) via @openmp_arb

True performance portability
#GPU #openmp
Time to try slorado now #RNAseq!
https://www.amd.com/en/blogs/2025/breaking-barriers-in-genomics.html
https://www.biorxiv.org/content/10.64898/2026.03.25.714356v1

Christos Argyropoulos MD, PhD, FASN 🇺🇸 0kale/acc May 24

Kind of nuts that one can make all of these cards for offloading with clang (AMD requires cosplaying as a different architecture) via @[email protected] . True performance portability #GPU #openmp Time to try slorado now #RNAseq! www.amd.com/en/blogs/202... www.biorxiv.org/content/10.6...

OpenMP ARB May 13

We had a very productive F2F meeting last week at the Argonne Leadership Computing Facility, with many thanks to our great hosts at the Argonne National Lab. The main objective was to feature-freeze OpenMP API version 6.1 and we accomplished that mission!

#OpenMP #ParallelComputing #HPC

Christos Argyropoulos MD, PhD, FASN 🇺🇸 0kale/acc May 6

Thinking of going #AMD for my GPU (mostly non AI) HPC workloads . Solid fp64 performance, play nicely with #OpenMP , cheaper on secondary and primary markets. Missing anything?

OpenMP ARB Apr 6

The OpenMP Architecture Review Board has formed a #Python Language Subcommittee — a significant step toward bringing standardized shared-memory parallelism to the world's most widely used programming language.

The subcommittee's goal is to define #OpenMP directive support for Python and include it in the OpenMP API 7.0 specification, targeted for 2029.

https://www.openmp.org/2026/python-subcomittee/
#HPC #parallelcomputing

Python Language Subcommittee - OpenMP

The OpenMP ARB has formed a Python Language Subcommittee, to bringing the most used parallel programming API to the most popular programming language.

OpenMP

HGPU group Mar 25

High-level Programming of Vulkan-based GPUs Through OpenMP

#OpenMP #Vulkan

https://hgpu.org/?p=30705

High-level Programming of Vulkan-based GPUs Through OpenMP

Modern applications often involve complex, structured or data-parallel computations on large datasets. Traditionally, GPUs have served as the primary accelerators for such tasks, mostly through com…

hgpu.org