Electronic Frontier FoundationEvery person makes a difference in the online rights movement, and every person counts in our Year-End Challenge. Donate ANY AMOUNT by Dec 31 and help us unlock bonus grants: https://eff.org/yec
Aditya
@slaedr
- 0 Followers
- 10 Following
- 9 Posts
Scientist in applied mathematics and computational science, software engineer, metal music enthusiast, on-and-off amateur musician
Andreas HertenFinally, #JUPITER crossed the 1 ExaFLOP/s threshold today. The list is lying to you, though, it's not like it's 1000 PFLOP/s exactly, it's 1000.184 PFLOP/s; the rest got lost to rounding.
The 184 TFLOP/s are pretty much exactly the same as the previous #JSC supercomputer JUGENE from 2008. Incredible, that this gets lost in the margins only 17 years later.
https://mastodon.social/deck/@fzj_jsc@social.fz-juelich.de/115566616685186256
#HPC #exa_JUPITER
The 184 TFLOP/s are pretty much exactly the same as the previous #JSC supercomputer JUGENE from 2008. Incredible, that this gets lost in the margins only 17 years later.
https://mastodon.social/deck/@fzj_jsc@social.fz-juelich.de/115566616685186256
#HPC #exa_JUPITER
https://openai.com/global-affairs/openais-economic-blueprint/ Seriously? Perhaps if the US did not jump so quickly and completely into cars, perhaps it would have better public transport. In any case, I don't trust OpenAI with policy. #AI #AIPolicy
Eugenia LShutter Encoder is a new free GPL app (donationware optionally) that is a front-end to ffmpeg. It's kind of like Handbrake, or Avidemux, or LosslessCut, but with a lot more ffmpeg options exposed. Available for Mac, Win, and Linux.
#linux #opensource #ffmpeg #foss #video #videos #videoeditor #handbrake
arXiv cs.CE botBDDC Preconditioning on GPUs for Cardiac Simulations
Fritz Goebel, Terry Cojean, Hartwig Anzt
https://arxiv.org/abs/2410.14786 https://arxiv.org/pdf/2410.14786 https://arxiv.org/html/2410.14786
arXiv:2410.14786v1 Announce Type: new
Abstract: In order to understand cardiac arrhythmia, computer models for electrophysiology are essential. In the EuroHPC MicroCARD project, we adapt the current models and leverage modern computing resources to model diseased hearts and their microstructure accurately. Towards this objective, we develop a portable, highly efficient, and performing BDDC preconditioner and solver implementation, demonstrating scalability with over 90% efficiency on up to 100 GPUs.
BDDC Preconditioning on GPUs for Cardiac Simulations
In order to understand cardiac arrhythmia, computer models for electrophysiology are essential. In the EuroHPC MicroCARD project, we adapt the current models and leverage modern computing resources to model diseased hearts and their microstructure accurately. Towards this objective, we develop a portable, highly efficient, and performing BDDC preconditioner and solver implementation, demonstrating scalability with over 90% efficiency on up to 100 GPUs.
I missed the fact that #hpgmp (High-performance GMRES mixed-precision) is now a separate project (finally). https://github.com/hpg-mxp/hpg-mxp #mixedprecision #hpc
drevil@DrGravitas @Craftycat @nixCraft I was thinking about that too. The issue I have is that, if this partnership is done to ease the integration of ChatGPT into StackOverflow, they will include ai generated answers first regardless of if people delete their posts or not.
By removing your answers at least you decrease the value of SO, though given the very, very small portion of their user base that cares about this, I don't think it will affect them much.
I guess if you want to give them the middle finger for this, than go ahead, but truth be told, it's unlikely this will cause any significant changes on their behavior. Most people I'd imagine will actually be thrilled about ChatGPT integration, since for a lot of people all that it means is that they might get automated replies without waiting for a human, regardless of the quality of the answers they might receive (assuming they are able to tell if they are bad or not to begin with).
Either way, it's a very bleak future I'm afraid, and the only saving grace I can think of is that the answers become so repetitive and terrible, people just revert back to forums/alterantives to SO. Or, you know, just reading documentation when possible.