Closed ax3l closed 5 years ago
How about: PIConGPU - Performance-portable exascale-ready plasma simulations ?
... or use many-core architectures in the subtitle?
plasma simulations for the manycore era ?
Also not bad - I like leaving out an article like "A" or the "The" because one is modest and the other bold but we need neither
update
What about manycore plasma simulations for the exascale era ?
I find the last update sounds somewhat doubled/too verbose with both the specific manycore and the reference exascale (which could also become something different in the end)
... plasma simulation on GPUs and many more many-core architectures
PIConGPU - Inherently parallel plasma simulations.
PIConGPU - An exascale particle-in-cell code for hybrid computing architectures PIConGPU - An exascale ready particle-in-cell code for hybrid clusters
With the 0.4.0 release around the corner - what do we go for? Feel free to add your :+1: to any of the comments you like! :)
Well, this is a though task. We have several strong points we could make, e.g.
But combining all of these contradicts our aim of a one-liner. Also, for non-Experts we should say what the project is about, i.e. particle-in-cell simulations.
With all these in mind, what do you think of
PIConGPU - High performant particle-in-cell simulations
which reduces the one-liner to our key feature but does not exclude everything else. Furthermore, this statement will still be true when we are in the zetta-scale age or run on quantum computers ;-D
All the other outstanding features of picongpu would then be named in the first sentences of the description. Although I would more like to see these as a subtitle but I do not know wheter this is possible with the layout framework given by github.com and readthedocs.io. This suptitle could be something along the lines
Running on CPUs, GPUs and more at Desktops and HPC clusters alike
Let's bash that suggestion ;-D
I don't like performant as an adjective. It is used very rarely and only found its way into English language at all in the last 30 years. Don't the other codes do HPC as well? Although not on all the sexy hardwares. I don't think it's catchy enough.
Well, it's been a while since I used PIConGPU for the last time, but since I've got a notification for this thread I also wanted to propose something, even if it's just to bring some new terms to the table. What about something like:
PIConGPU - A universal, multi-purpose PIC code for the exascale era
Universal because it can now basically run on any architecture of interest, multi-purpose because of all the different physics cases that can be simulated, and well, the rest is kind of self-explanatory. Just an idea ;)
Sounds great as well. I just wonder if we should explain PIC "with PIC" or if we might want to bring in the key applications (plasma-physics, laser-plasma physics, maybe later more on beam-physics as well) in the sub title? Of course, one can use it as well for antenna physics, meta-materials if one just adds modified solvers, etc. ... ;-)
Trying to combine a little Frankenstein from your great suggestions and keeping it as short and striking as possible, I would suggest to change for the 0.4.0 release for now to:
PIConGPU - particle-in-cell simulations for the exascale era
Dear @ComputationalRadiationPhysics/picongpu-maintainers ,
we need a new subtitle for our docs and GitHub that describes PIConGPU in a short one-liner.
Currently, we use PIConGPU - A particle-in-cell code for GPGPUs but this is now a tremendous understatement.
The idea is to keep the one-liner short and striking as before, think of it as a useful addition to the name highlighting more essential characteristics of our capabilities. Since PIConGPU is a well-known brand by now (and GPUs are still the fastest/sexiest general purpose hardware around), the subtitle helps to show that we are also fit for any other hardware (with the upcoming release(s) in a few months) but still keeps our most-successful flavor.
Please make suggestions and/or vote here.
let's start:
PIConGPU - ...
Now you!
(No, PICy McPICface is not an option ;) )
ccing @ComputationalRadiationPhysics/picongpu-developers @BenjaminW3