James Schloss

Tried: ``` Base.Cartesian.@nif(4, d->meh == d, d->tuple_thingy[d](tid), d->@print(d,'\t',meh)) ``` It ran, but did not change the input at all Note that I also tried

I tried ``` Base.Cartesian.@nif(4, d->meh == d, d->input[tid] = tuple_thingy[d](tid), d->@print(d,'\t',meh)) ``` ~~It ran, but didn't produce any output on the GPU (worked on CPU)~~ IGNORE: worked after turning Julia...

I tried to roll my own nif function: ``` macro nif(N::Symbol, condition, operation...) # Handle the final "else" ex = esc(Base.Cartesian.inlineanonymous(length(operation) > 1 ? operation[2] : operation[1] , eval(N))) #...

I split off the atomix tests because they are not passing yet. This PR is currently passing all the tests locally and *most* of the tests here (I don't know...

For those that are struggling with Atomix in KA, you can use this PR with: ``` ] add https://github.com/leios/KernelAbstractions.jl#atomic_attempts_3 ] add https://github.com/leios/KernelAbstractions.jl:lib/CUDAKernels#atomic_attempts_3 ``` The first line adds KA, the second...

I have no problem with Atomix being the default API in the long run (though I have a somewhat strong preference to just using `atomic_add(...)` instead of using a macro...

I don't think `@uniform` is even documented, is it?

Haha, some of these look like loads of fun! I'll try a few of these on stream tomorrow, for sure! Thanks again for all your help!

Yeah, I will certainly be checking them out. I never really thought about doing raytracing art before... I also found the formula for the refractive index of the invisible lens,...

Haha. I assume you are Toeger? I am working on it right now. I think there will be too much of a discontinuity using the code for tomorrow, so I...