Following on fromn the discussion on PIC codes. Is there much activity in using Julia for fusion and plasma simulations? Any pointers gratefully received.
Slightly off topic, here’s a recent paper that is very interesting. They 'black box optimized and tripled yield.
Not to my knowledge–people might be playing with toy programs on single nodes, but the language hasn’t been stable for long enough to get buy-in to build a full radiation hydro code with a lifespan of 15-30 years. The nuts & bolts C++/Fortran approach still dominates. The ability to easily extend kernels to GPUs might be a killer feature considering the hardware of Summit/Sierra, but it’s a lot less work to write a specialized GPU kernel for a hot loop in the same low-level language as the rest of the hydrocode than it is to build a new code from scratch in Julia, fleshing out the language’s HPC bones along the way.
I’m a fusion researcher and I use Julia as my primary programming language. I started the JuliaFusion Organization a while back but, for the moment, none of the packages are registered (I’m waiting for the v1.2 release) and they are only used for my own research.
Thankyou all. I am talking to the UK fusion research centre tomorrow. All responses are very helpful.