#ScreenshotSaturday The 424 is undergoing a major redesign.

https://perrinn.com/chat/FZn4icU4Xuw9P7Uufz0H

Getting a bit further with my exploration of Computational Fluid Dynamics (CFD) this evening. Using the mighty @FreeCAD I've modelled this "Ballute" style parachute and simulated it descending at 6m/s in air. The air velocity is plotted in the images. Ace to see the turbulence in the point cloud/steam chaser view. If any #CFD experts fancy a chat.... I HAVE QUESTIONS! #OpenFOAM #FreeCAD

Oft wird angenommen, dass künftig die Förderung von #Strom aus Erneuerbaren nach dem #EEG 2023 pauschal auf #CfD umgestellt werden muss. Wie eine neue Studie der Stiftung #Umweltenergierecht zeigt, bestehen aber mehr Spielräume als gedacht.
https://nachrichten.idw-online.de/2025/04/15/zweiseitige-differenzvertraege-cfd-und-eu-recht-spielraeume-bei-der-reform-des-eeg-2023

Following on from my earlier stuff about being inspired about parachute/recovery systems and Ballute design in @FreeCAD I've dived into setting up the CfDoF workbench and bashing through a simple tutorial pushing water through a stepped tube. I think I borked the glyph a little, but I feel like I've learnt a little... and seen a LOT to learn. #OpenFOAM #CFD #Paraview

What an honor to start the #IWOCL conference with my keynote talk! Nowhere else you get to talk to so many #OpenCL and #SYCL experts in one room! I shared some updates on my #FluidX3D #CFD solver, how I optimized it at the smallest level of a single grid cell, to scale it up on the largest #Intel #Xeon6 #HPC systems that provide more memory capacity than any #GPU server.

#ScreenshotSaturday The 424 is undergoing a major redesign, integrating a new front powertrain concept. Key updates include a narrower chassis for a central driving position, an extended nose, refined aerodynamics, and a roof opening. The development continues, with CFD testing planned soon.

I made this #FluidX3D #CFD simulation run on a frankenstein zoo of AMD + Nvidia + Intel #GPU​s!
https://www.youtube.com/watch?v=_8Ed8ET9gBU

The ultimate SLI abomination setup:
- 1x Nvidia A100 40GB
- 1x Nvidia Tesla P100 16GB
- 2x Nvidia A2 15GB
- 3x AMD Instinct MI50
- 1x Intel Arc A770 16GB

I split the 2.5B cells in 9 domains of 15GB - A100 takes 2 domains, the other GPUs 1 domain each. The GPUs communicate over PCIe via #OpenCL.

Huge thanks to Tobias Ribizel from TUM for the hardware!

I got access to @LRZ_DE's new coma-cluster for #OpenCL benchmarking and experimentation
I've added a ton of new #FluidX3D #CFD #GPU​/​#CPU benchmarks:
https://github.com/ProjectPhysX/FluidX3D?tab=readme-ov-file#single-gpucpu-benchmarks

Notable hardware configurations include:
- 4x H100 NVL 94GB
- 2x Nvidia L40S 48GB
- 2x Nvidia A2 15GB datacenter toaster
- 2x Intel Arc A770 16GB
- AMD+Nvidia SLI abomination consisting of 3x Instinct MI50 32GB + 1x A100 40GB
- AMD Radeon 8060S (chonky Ryzen AI Max+ 395 iGPU with quad-channel RAM) thanks to @cheese

Proud to be part of the PERRINN 424 project alongside these amazing companies!

https://discover.perrinn.com/perrinn-com/perrinn-early-days

This thread about writing games vs writing game engines <https://peoplemaking.games/@eniko/114137694102639793> by @eniko, and the comments within by many other people, is a fascinating read for me, particularly in relation to the similarities and the differences with our experience in the development of what is, for all intents and purposes, a #CFD engine, but also many of the test cases it has been used for.

For many #ComputationalFluidDynamics methods, it's actually pretty simple to write an implementation for a “trivial” test case (straight walls, right angles if any at all, periodic boundary conditions, etc). We did that for example in a couple of hours during the MODCLIM 2016 training school https://modclim.ulpgc.es/index.php/events/8-modclim-events

Things become quickly non-trivial as soon as you start needing

1. non-trivial geometries

and

2. more complex physics and/or more sophisticated methods.

1/

#FluidX3D #CFD v3.2 is out! I've implemented the much requested #GPU summation for object force/torque; it's ~20x faster than #CPU #multithreading.
Horizontal sum in #OpenCL was a nice exercise - first local memory reduction and then hardware-supported atomic floating-point add in VRAM, in a single-stage kernel. Hammering atomics isn't too bad as each of the ~10-340 workgroups dispatched at a time does only a single atomic add.
Also improved volumetric #raytracing!
https://github.com/ProjectPhysX/FluidX3D/releases/tag/v3.2

Hot Aisle's 8x AMD #MI300X server is the fastest computer I've ever tested in #FluidX3D #CFD, achieving a peak #LBM performance of 205 GLUPs/s, and a combined VRAM bandwidth of 23 TB/s.
The #RTX 5090 looks like a toy in comparison.

MI300X beats even Nvidia's GH200 94GB. This marks a very fascinating inflection point in #GPGPU: #CUDA is not the performance leader anymore.
You need a cross-vendor language like #OpenCL to leverage its power.

FluidX3D on #GitHub: https://github.com/ProjectPhysX/FluidX3D

I'm going to take advantage of the current #eruption on Mt #Etna to discuss some of the challenges of #modelling #lava flows. Buckle up (or just silence me) because this is going to be a long thread.

First of all, why do we want to model lava flows? The answer most definitely isn't «because we can», since —as I'm going to explain momentarily— we actually cannot. Still having an idea about how lava flows and sets in place is a powerful tool for the assessment (and possibly mitigation) of the associated #hazard and #risk: if we can tell how lava flows, we can tell which areas are going to be reached by the lava, and hopefully also improve the design of tactical and strategic actions that can be taken to minimize the damage.

(Of course, whether or not those actions will then be taken is an entirely different matter, but that's mostly politics, not science.)

1/

The 4x #Nvidia #H100 SXM5 server in the new Festus cluster at Uni Bayreuth is the fastest system I've ever tested in #FluidX3D #CFD, achieving 78 GLUPs/s #LBM performance at ~1650W #GPU power draw.
https://github.com/ProjectPhysX/FluidX3D?tab=readme-ov-file#multi-gpu-benchmarks
https://www.hpc.uni-bayreuth.de/clusters/festus/#__tabbed_1_3

For the first time, the Chapel team at HPE and the CHAMPS team at Polytechnique Montreal will give a joint talk. Join us Thursday, February 20th at 9AM PST in Advanced Modeling & Simulation (AMS) Seminar Series hosted by NASA Ames Research Center!

The talk will be open to public. You can find the meeting link over at https://www.nas.nasa.gov/pubs/ams/2025/02-20-25.html !

#FluidX3D #CFD v3.1 is out! I have updated the #OpenCL headers for better device specs detection via device ID and #Nvidia compute capability, fixed broken voxelization on some #GPU​s and added a workaround for a CPU compiler bug that corrupted rendering. Also #AMD GPUs will now show up with their correct name (no idea why AMD can't report it as CL_DEVICE_NAME like every other sane vendor and instead need CL_DEVICE_BOARD_NAME_AMD extension...)
Have fun!
https://github.com/ProjectPhysX/FluidX3D/releases/tag/v3.1

How CO2 Gets Into the Ocean

Our oceans absorb large amounts of atmospheric carbon dioxide. Liquid water is quite good at dissolving carbon dioxide gas, which is why we have seltzer, beer, sodas, and other carbonated drinks. The larger the surface area between the atmosphere and the ocean, the more quickly carbon dioxide gets dissolved. So breaking waves — which trap lots of bubbles — are a major factor in this carbon exchange.

This video shows off numerical simulations exploring how breaking waves and bubbly turbulence affect carbon getting into the ocean. The visualizations are gorgeous, and you can follow the problem from the large-scale (breaking waves) all the way down to the smallest scales (bubbles coalescing). (Video and image credit: S. Pirozzoli et al.)