Move semantics for Nim

[@ThePrimeyuri]

programing in nim is pure joy, i am beginner though

[@TinBryn]

On that question about nilability and you answered that Nim style moves can't use a variable after it is used and so the compiler can pretend that it's ok to set it to nil even when it is non-nilable. I think it a better solution is to have destructive moves where it doesn't call the destructor and lets the moved from variable just fall off the stack.

[@andreasrumpf9012]

Yes, in the spec that is actually covered as optimizing away so called `wasMoved(x); destroy(x)` pairs. We're working on that.

I am still wondering cuda with nim. There is nimcuda but it is only supports cuda 8. Cuda is very important to grow nim at hpc area.
I am using Python because modules and good gpu support.

[@mamyratsimbazafy3328]

Nimcuda works with 10.2. I develop Arraymancer (github.com/mratsim/Arraymancer) and the Cuda tests pass on my machine.

[@muhammadreihanhaekal5117]

hope you doing well :)

[@uncannydeduction]

Would be great if you didn't prioritize compiler speed in my opinion. To make the compiler smarter so that we can write simpler code. Computers will get faster in time.

[@ctrlaltdebug]

I disagree. C++ compile times are still slow after 35 years!

[@itellyouforfree7238]

@@ctrlaltdebug because C++ is like 35 diffent languages put together lazily without thinking about how they interact