#parallelcomputing — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #parallelcomputing, aggregated by home.social.
-
👾🎨 Ah, yes, the "Zen" of parallel programming — because nothing screams tranquility like the chaotic symphony of GPU kernels and AMD jargon. 🧘♂️✨ Smolnero's attempt to turn technical mumbo jumbo into spiritual enlightenment is as effective as teaching a cat to do calculus. 😹
https://smolnero.com/posts/the-zen-of-parallel-programming-the-posture-of-a-kernel #ZenProgramming #ParallelComputing #GPUChaos #AMDJargon #TechHumor #HackerNews #ngated -
👾🎨 Ah, yes, the "Zen" of parallel programming — because nothing screams tranquility like the chaotic symphony of GPU kernels and AMD jargon. 🧘♂️✨ Smolnero's attempt to turn technical mumbo jumbo into spiritual enlightenment is as effective as teaching a cat to do calculus. 😹
https://smolnero.com/posts/the-zen-of-parallel-programming-the-posture-of-a-kernel #ZenProgramming #ParallelComputing #GPUChaos #AMDJargon #TechHumor #HackerNews #ngated -
Learn how Mojo combines SIMD, multi-core parallelism, and Python interoperability to accelerate AI inference and data science workloads. https://hackernoon.com/mojo-lets-you-parallelize-ai-code-without-leaving-python-behind #parallelcomputing
-
Learn how Mojo combines SIMD, multi-core parallelism, and Python interoperability to accelerate AI inference and data science workloads. https://hackernoon.com/mojo-lets-you-parallelize-ai-code-without-leaving-python-behind #parallelcomputing
-
China's LineShine Supercomputer: A Sense of Scale ⚡🖥️
China's LineShine 灵晟 supercomputer is the most powerful supercomputer in the world 🌍, delivering a verified 2.198 exaFLOPS on the TOP500 benchmark. That means it can perform more than 2 quintillion (2×10¹⁸) calculations every second 🚀. The system consumes 42.2 megawatts of electricity ⚡ and is powered by more than 13.7 million conventional CPU cores 🧠.
The Scale of Its Computing Power 📊
Compared to Humanity 👥: If every person on Earth performed one mathematical calculation every second, without stopping ⏱️, it would take the entire global population about 4 years 📅 to equal the amount of computation LineShine completes in a single second.
Consumer Hardware 💻: Matching LineShine's sustained performance would require more than 20 million of today's fastest consumer graphics cards 🖥️, such as an RTX 5090, working together in perfect synchronization 🔗.
LineShine demonstrates the extraordinary scale of modern high-performance computing (HPC), enabling scientific calculations that would be practically impossible using conventional computers 🚀.
#Supercomputer #LineShine #Exascale #Exaflop #HPC #HighPerformanceComputing #Technology #Innovation #Science #Engineering #Computing #CPU #ParallelComputing #ClimateScience #WeatherForecasting #Neuroscience #QuantumChemistry #Physics #MolecularScience #DataScience #ArtificialIntelligence #AI #Research #STEM #FutureTech #DigitalEarth #ComputerScience #TechFacts #NextGenComputing #TOP500
-
China's LineShine Supercomputer: A Sense of Scale ⚡🖥️
China's LineShine 灵晟 supercomputer is the most powerful supercomputer in the world 🌍, delivering a verified 2.198 exaFLOPS on the TOP500 benchmark. That means it can perform more than 2 quintillion (2×10¹⁸) calculations every second 🚀. The system consumes 42.2 megawatts of electricity ⚡ and is powered by more than 13.7 million conventional CPU cores 🧠.
The Scale of Its Computing Power 📊
Compared to Humanity 👥: If every person on Earth performed one mathematical calculation every second, without stopping ⏱️, it would take the entire global population about 4 years 📅 to equal the amount of computation LineShine completes in a single second.
Consumer Hardware 💻: Matching LineShine's sustained performance would require more than 20 million of today's fastest consumer graphics cards 🖥️, such as an RTX 5090, working together in perfect synchronization 🔗.
LineShine demonstrates the extraordinary scale of modern high-performance computing (HPC), enabling scientific calculations that would be practically impossible using conventional computers 🚀.
#Supercomputer #LineShine #Exascale #Exaflop #HPC #HighPerformanceComputing #Technology #Innovation #Science #Engineering #Computing #CPU #ParallelComputing #ClimateScience #WeatherForecasting #Neuroscience #QuantumChemistry #Physics #MolecularScience #DataScience #ArtificialIntelligence #AI #Research #STEM #FutureTech #DigitalEarth #ComputerScience #TechFacts #NextGenComputing #TOP500
-
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!
-
Sharing big R objects across processes shouldn’t mean copying them over and over.
mori uses shared memory + ALTREP to give you zero-copy access—multiple processes, one underlying object.
Fast, memory-efficient, and built for modern parallel workflows.
-
Sharing big R objects across processes shouldn’t mean copying them over and over.
mori uses shared memory + ALTREP to give you zero-copy access—multiple processes, one underlying object.
Fast, memory-efficient, and built for modern parallel workflows.
-
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 -
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 -
🎉 Wow, #groundbreaking revelation: when you give a hungry #AI agent a buffet of GPUs, it eats faster! 🚀 Who knew? Apparently, parallel computing is a thing now. 🙄 Thanks for the 12-minute read on how technology works, we were all clueless. 😂
https://blog.skypilot.co/scaling-autoresearch/ #parallelcomputing #technews #GPUbuffet #humor #HackerNews #ngated -
🎉 Wow, #groundbreaking revelation: when you give a hungry #AI agent a buffet of GPUs, it eats faster! 🚀 Who knew? Apparently, parallel computing is a thing now. 🙄 Thanks for the 12-minute read on how technology works, we were all clueless. 😂
https://blog.skypilot.co/scaling-autoresearch/ #parallelcomputing #technews #GPUbuffet #humor #HackerNews #ngated -
Setting Up a Cluster of Tiny PCs for Parallel Computing
https://www.kenkoonwong.com/blog/parallel-computing/
#HackerNews #ParallelComputing #TinyPCs #ClusterSetup #TechInnovations #HPC
-
Setting Up a Cluster of Tiny PCs for Parallel Computing
https://www.kenkoonwong.com/blog/parallel-computing/
#HackerNews #ParallelComputing #TinyPCs #ClusterSetup #TechInnovations #HPC
-
David Lattimore delves into the complexities of parallelizing dynamic graph traversals with Rust's Rayon. His exploration moves beyond fixed workloads, examining iterative approaches: custom work-sharing, scoped spawning, & channel-based solutions. Key insights reveal significant trade-offs involving heap allocations, deadlock risks, and compositional limitations inherent in parallel paradigms. Thoughtful work for those navigating concurrent systems. #RustLang #ParallelComputing #TechEthics
-
Questa settimana ho fatto la-due-giorni-a-Bologna 🚀🚀
Intensa di #talk ispiranti, #gadget bellissimi … e le persone hanno reso tutto davvero indimenticabile. ✨
Conosco i retroscena dell’organizzazione e i ragazzi del @grusp hanno resto tutto perfetto, leggero e spensierato .. sebbene non fosse per nulla facile 💪
E’ sempre un piacere essere accettati come #speaker ai loro eventi ❤️
Alla prossima !
#DataAnalysis #Dask #Kubernetes #ParallelComputing #Scalability #AWS #DevSecOpsDay #ContainerDay
-
Efficient GPU algorithm converts Bézier paths into renderable geometry, enabling real-time, cross-platform vector graphics rendering. https://hackernoon.com/implementing-data-parallel-stroke-to-fill-conversion-on-modern-gpus #parallelcomputing
-
Efficient GPU algorithm converts Bézier paths into renderable geometry, enabling real-time, cross-platform vector graphics rendering. https://hackernoon.com/implementing-data-parallel-stroke-to-fill-conversion-on-modern-gpus #parallelcomputing
-
Efficiently convert cubic Bézier curves to Euler spirals for smoother GPU rendering and accurate parallel curve computations. https://hackernoon.com/how-to-convert-cubic-bezier-curves-into-euler-spirals-for-gpu-optimization #parallelcomputing
-
Efficiently convert cubic Bézier curves to Euler spirals for smoother GPU rendering and accurate parallel curve computations. https://hackernoon.com/how-to-convert-cubic-bezier-curves-into-euler-spirals-for-gpu-optimization #parallelcomputing
-
Today I introduced a much-needed feature to #GPUSPH.
Our code supports multi-GPU and even multi-node, so in general if you have a large simulation you'll want to distribute it over all your GPUs using our internal support for it.
However, in some cases, you need to run a battery of simulations and your problem size isn't large enough to justify the use of more than a couple of GPUs for each simulation.
In this case, rather than running the simulations in your set serially (one after the other) using all GPUs for each, you'll want to run them in parallel, potentially even each on a single GPUs.
The idea is to find the next avaialble (set of) GPU(s) and launch a simulation on them while there are still available sets, then wait until a “slot” frees up and start the new one(s) as slots get freed.
Until now, we've been doing this manually by partitioning the set of simulations to do and start them in different shells.
There is actually a very powerful tool to achieve this on the command, line, GNU Parallel. As with all powerful tools, however, this is somewhat cumbersome to configure to get the intended result. And after Doing It Right™ one must remember the invocation magic …
So today I found some time to write a wrapper around GNU Parallel that basically (1) enumerates the available GPUs and (2) appends the appropriate --device command-line option to the invocation of GPUSPH, based on the slot number.
#GPGPU #ParallelComputing #DistributedComputing #GNUParallel
-
Today I introduced a much-needed feature to #GPUSPH.
Our code supports multi-GPU and even multi-node, so in general if you have a large simulation you'll want to distribute it over all your GPUs using our internal support for it.
However, in some cases, you need to run a battery of simulations and your problem size isn't large enough to justify the use of more than a couple of GPUs for each simulation.
In this case, rather than running the simulations in your set serially (one after the other) using all GPUs for each, you'll want to run them in parallel, potentially even each on a single GPUs.
The idea is to find the next avaialble (set of) GPU(s) and launch a simulation on them while there are still available sets, then wait until a “slot” frees up and start the new one(s) as slots get freed.
Until now, we've been doing this manually by partitioning the set of simulations to do and start them in different shells.
There is actually a very powerful tool to achieve this on the command, line, GNU Parallel. As with all powerful tools, however, this is somewhat cumbersome to configure to get the intended result. And after Doing It Right™ one must remember the invocation magic …
So today I found some time to write a wrapper around GNU Parallel that basically (1) enumerates the available GPUs and (2) appends the appropriate --device command-line option to the invocation of GPUSPH, based on the slot number.
#GPGPU #ParallelComputing #DistributedComputing #GNUParallel
-
We are excited to return to Supercomputing! Join us on Sunday, November 16th for the OpenMP tutorial, Mastering OpenMP Tasking. This tutorial will provide performance and scalability recipes to improve the performance of OpenMP tasking applications.
Learn more about all of OpenMP's activities at #SC25 at: https://www.openmp.org/events/sc25/
#OpenMP #Tasking #parallelcomputing #hpc #multiprocessor -
Join us at Supercomputing 2025 in St. Louis!
We have a packed agenda at this year's show with BOFs and tutorials, and be sure to join us in booth #911 to meet with OpenMP experts to ask your toughest questions, enter the daily Book Drawing, get your free OpenMP API 6.0 reference guide, and have an afternoon beverage.
Learn more: https://www.openmp.org/events/sc25/
#SC25 #OpenMP #parallelcomputing #hpc #gpu #pyomp -
Wir freuen uns, Euch auch in diesem Jahr wieder spannende MATLAB-Kurse im Online-Format in der GWDG Academy anzubieten, welche von MathWorks-Mitarbeitern durchgeführt werden:
💠 Parallel Computing with MATLAB
Termin: 17.11.2025, 10:00 – 13:00 Uhr
💠 Demo Session: Scaling up MATLAB to the GWDG Scientific Compute Cluster
Termin: 19.11.2025, 15:00 – 16:30 Uhr
💠 Introduction to Research Software Development with MATLAB
Termin: 20.11.2025, 09:00 – 12:00 Uhr
💠 Connecting MATLAB with Python and other Open Source Tools
Termin: 20.11.2025 14:00 – 17:00 UhrDie Kurstermine werden ergänzt um eine sogenannte Office Hour (online) am 21.11.2025, 14:00 – 15:00 Uhr, während der Fragen zu den vorgestellten Themen der Kurse ausgiebig gestellt und behandelt werden können, um einen Austausch zwischen den Teilnehmer*innen und den Dozenten zu erreichen.
#gwdg #academy #gwdgacademy #kurs #matlab #parallelcomputing #göttingen #unigöttingen #mathswork
-
Accelerated Game of Life with CUDA / Triton
https://www.boristhebrave.com/2025/09/11/accelerated-game-of-life-with-cuda-triton/
#HackerNews #AcceleratedGameOfLife #CUDA #Triton #GamingTech #ParallelComputing
-
Accelerated Game of Life with CUDA / Triton
https://www.boristhebrave.com/2025/09/11/accelerated-game-of-life-with-cuda-triton/
#HackerNews #AcceleratedGameOfLife #CUDA #Triton #GamingTech #ParallelComputing
-
GPUPrefixSums – state of the art GPU prefix sum algorithms
https://github.com/b0nes164/GPUPrefixSums
#HackerNews #GPUPrefixSums #GPUAlgorithms #PrefixSum #ParallelComputing #DataStructures
-
GPUPrefixSums – state of the art GPU prefix sum algorithms
https://github.com/b0nes164/GPUPrefixSums
#HackerNews #GPUPrefixSums #GPUAlgorithms #PrefixSum #ParallelComputing #DataStructures
-
📢 OpenMP Newsletter – July 2025 Edition
Highlights:
🗓️ IWOMP 2025 preliminary program
👥 3 new members join the OpenMP Architecture Review Board
🛠️ OpenMP support in:* GCC 15.1
* Intel oneAPI HPC Toolkit 2025.2
* NumPy 2.3
Full newsletter: https://mailchi.mp/e82391a1d7b0/thanks-for-your-interest-in-openmp-17461513
#OpenMP #HPC #IWOMP2025 #ParallelComputing #NumPy #GCC #InteloneAPI
-
Going to the (Parallel) Chapel - There is always the promise of using more computing power for a single task. Your ... - https://hackaday.com/2025/07/06/going-to-the-parallel-chapel/ #softwaredevelopment #parallelcomputing
-
Going to the (Parallel) Chapel - There is always the promise of using more computing power for a single task. Your ... - https://hackaday.com/2025/07/06/going-to-the-parallel-chapel/ #softwaredevelopment #parallelcomputing
-
Calculating the Fibonacci numbers on GPU
https://veitner.bearblog.dev/calculating-the-fibonacci-numbers-on-gpu/
#HackerNews #Calculating #Fibonacci #GPU #Performance #ParallelComputing #TechInnovation #Algorithms
-
Calculating the Fibonacci numbers on GPU
https://veitner.bearblog.dev/calculating-the-fibonacci-numbers-on-gpu/
#HackerNews #Calculating #Fibonacci #GPU #Performance #ParallelComputing #TechInnovation #Algorithms
-
📸 Full house at the OpenMP BOF at #ISC25 — over 140 attendees joined us in Hamburg! 🎉
Our session "What to Expect from OpenMP API Version 6.0" covered:
✅ A dive into key features of OpenMP 6.0
✅ A preview of 6.1 and 7.0
✅ Updates from toolchain developers
✅ Lively Q&A to help shape future OpenMP directionsThanks to everyone who contributed — your feedback is powering the future of parallel programming! 💡
#OpenMP #HPC #ISC2025 #OpenMP6 #ParallelComputing #Supercomputing
-
We’re excited to welcome NextSilicon to the OpenMP Architecture Review Board! 🎉
Their Intelligent Compute Architecture blends adaptive computing with self-optimizing hardware/software and open frameworks like OpenMP. Together, we’re shaping a future of performant, portable, shared-memory parallelism. 💻🌐
Read the press release:
https://tinyurl.com/yksfbrah -
Join us at #ISC25 for the tutorial “Advanced OpenMP: Performance and 6.0 Features” on Friday, June 13, 9:00–13:00 CEST in Hall Y12, 2nd Floor, Hamburg Congress Center.
Learn how to boost OpenMP code performance on NUMA systems and accelerators, and get hands-on insights into vectorization, data locality, and the latest features in OpenMP 6.0.
Ideal for developers who want to go beyond the basics!
-
Just published the post "Parallel and distributed computing in GNU Health." :gnu: 🏥
https://meanmicio.org/2025/05/27/parallel-and-distributed-computing-in-gnu-health/
#ParallelComputing #GNUHealth #Tryton #OpenScience #GNU -
Just published the post "Parallel and distributed computing in GNU Health." :gnu: 🏥
https://meanmicio.org/2025/05/27/parallel-and-distributed-computing-in-gnu-health/
#ParallelComputing #GNUHealth #Tryton #OpenScience #GNU -
Vediamo le funzionalità del Fortran introdotte nel 2008, con i CoArray, e nel 2018 per scoprire come si possono sfruttare tutti i core delle nostre CPU abbreviando i tempi di calcoli scientifici complessi. #fortran #parallelcomputing #multithreading
https://www.youtube.com/watch?v=78_12a89MWQ -
🏆 Hugo Krawczyk – ACM Paris Kanellakis Theory and Practice Award
For pioneering and lasting contributions to the theoretical foundations of cryptographically secure communications, and to the protocols that form the security foundations of the Internet.
🔗 https://bit.ly/4jBJjHX👏 Congratulations to all the awardees shaping the future of computing!
#ACMTechnicalAwards #Cryptography #ParallelComputing #ComputerScience
-
Hi R people! Could you suggest some guides and links on how to set up parallel processing with `foreach` that can work both on Windows and Linux/Mac? I'm searching for guides on the net, but most of them seem to have become obsolete. Thank you!
-
Musings on a Good Parallel Computer - Until the late 1990s, the concept of a 3D accelerator card was something generally... - https://hackaday.com/2025/03/23/musings-on-a-good-parallel-computer/ #parallelcomputing #cpuarchitecture #computerhacks #gpu
-
👨💻 Oh, the eternal quest for a "good" parallel computer, like a quest for a unicorn that can also do your taxes. 🦄💻 Apparently, GPUs are only good for "predictable" tasks – maybe like predicting the inevitable death of your dreams for a universally versatile chip. 😂 Why not just ask for a toaster that can handle your email while you're at it? 🍞📧
https://raphlinus.github.io/gpu/2025/03/21/good-parallel-computer.html #parallelcomputing #unicornchip #GPU #humor #techdreams #HackerNews #ngated -
So What is a Supercomputer Anyway? - Over the decades there have been many denominations coined to classify computer sy... - https://hackaday.com/2025/03/19/so-what-is-a-supercomputer-anyway/ #parallelcomputing #computerhacks #supercomputer #featured #history #illiac #eniac
-
🚀💻 "I went to an #NVIDIA event, so naturally, my entire existence now revolves around repackaging basic sorting algorithms with #CUDA. Because, really, what else is parallel computing for besides impressing friends at parties? 🤓"
https://ashwanirathee.com/blog/2025/sort2/ #parallelcomputing #sortingalgorithms #techhumor #codinglife #HackerNews #ngated
