We discussed the upcoming Einstein Toolkit release and the Llama infrastructure in yesterday's telecon. Unfortunately, my wireless connection was not well enough to support much of a discussion with me.
When we (we = main Llama developers) released Llama earlier this year, we did so in the knowledge that the infrastructure supports high-quality production simulations, but lacks documentation and tutorials for newcomers. Our plan was (and is) to attract users who can help us add these tutorials, examples, documentation, and also help clean up the code. As you may imagine, Llama grew organically, and there are also some left-over pieces of unsuccessful sub-projects. Some pieces have also by now migrated into Cactus, and should now either be removed or rewritten and simplified.
We intentionally did not submit Llama for inclusion into the Einstein Toolkit yet: Llama should not come "for free" with the Einstein Toolkit, but people should realize and acknowledge that Llama represents an independent project and significant amount of work. We are happy it if people find Llama useful and appreciate any contributions towards making it even more useful. Personally, I expect Llama to become part of the ET, but the code is not ready yet.
In summary:
- We (we = main Llama deveopers or users) should give a Llama tutorial soon, helping others to get started.
- We are looking for a tutorial write-up and examples, in particular discussing some of the unexpected features (memory requirements, typical grid structures for binary systems, time step limitations, etc.)
- A code review with the intent to better integrate Llama with the core of Cactus, to remove outdated parts, and consolidate the existing code would be in order.
-erik