gpu

[behinger]

• Update typings to be more type-stable
• Reduce memory allocation and implement other performance improvements
• add array dimensions
• re-enable multi threading
• make Calculate_Q type stable
• use AppleAccelerate in certain places
• add intelvectormath
• minor
• more optimized methods
• minor
• log
• fixes, make fp local
• update project.toml
• precalculate y^rho
• things are faster now
• optimize more sums
• fix sphering LL
• speed improvement from 272µs to 640ns
• insane additional 100x speed improvement in calculate_y
• this should work, but havent tested -oops
• added missing broadcast to first minus, 340 to 230 µs
• actual 10% slower, but less allocs. try later again to use?
• 15% or so improvement, much better on allocations
• tried to pull out generation of Q
• improved y_rho, removed experimental exp/pow
• 1.46s/it vs. 0.6s/it
• finally works again....
• rearrange indizes from [n, N, m] to [m, n, N] and implement intel powx
• powx works
• handling for pcs without intel vector math (apple silicon)
• unsure
• Move some things to the Amica object, format all files
• 0,81 s / iter
• minor
• 0,57s/iter
• minor
• ivm.abs
• replace pinv(x)*x with ldiv!()
• sum -> loop
• one more sum to loop
• fix float32
• fix float32
• supposedly now whitening does something?
• better testcases maybe
• added GPU support