tldr: just look at the disasm and try to get the amount of instructions down <a:catnod:9081681287505
tldr: just look at the disasm and try to get the amount of instructions down 
Mare you saying the non-specialized version gets vectorized?
Mthat makes things easier
Ai stopped hyperoptimizing a long time ago, it just isn't worth the effort 90% of the time
if it works and its not stupid slow then its good
if it works and its not stupid slow then its good
Tit does not get vectorized per say. modern software doesn't use x87 
Mi don't optimize anything unless it's for shit like this
Lmath stuff like this you want optimized
Mwhere like yeah, you want this to have as little overhead as regular Vector3
Pwhich you is this
Myou perks
Mforgor to reply
Pi was not saying that
Pi gave you a specialised version
Myeah i'm using that one
PJIT cannot detect SIMD patterns today
Msadge
Mthen again neither can LLVM half the time
Ayeah i get it, for math you want things as fast as you can
Awrite the math library in C and call into it with p/invoke 
Pi must say though with the xplat helpers SIMD is so much easier
Plike have any of you seen the
feature/math-simd branch
Ti'm scared. what did tanner invent this time
Pnono this was the WIP work to vectorize silk.net.maths
Pso much work whereas now we don't have to care about the individual ISA
Toh yeah, basically exactly what net8 does
Tno wait, did we get crossplat in net7?
Ti'm too tired for code rn
Pnet7
Pbut yeah we were essentially writing our own
Vector128 xplat helper until the real thing came alongT"my" software rasterizer got quite a nice boost in net8, although it may be the avx512 doing... something, rather than JIT improvements
Tbut that code is completely destroyed by inlining limits 
Pwhen aggressive isn't aggressive enough
Myou can't do extension operators, can you
Mshame
TJIT literally gives up in the middle of my hot methods 
Mincredibly disgusting
Mi'll make it an instance method at least
Kwhy do you need conversion and division to be together?
Klike an as method should be good enough and then implement division using the same type
Mhm.
Mmy thought is
Vector3 / float is a very common pattern and useless with most integer valuesMalright, yeeted
Mdesign RFC: should static methods be provided
- on Vector3<T> (when doing generics, it'd be Vector3<T>.Lerp<TFloat> for instance)
- on static Vector3 (Vector3.Lerp<T, TFloat>, but for non-generic methods it wouldn't require specifying the <T>)
- on both (pick whichever you want)
- on static Vector3 only if the method is generic, otherwise on Vector3<T> (what i'm currently doing, but may be confusing to some)
Moh no, i've invoked tanner
Pi don't know what's happening
Pis this for silk
Msomeone was talking about generic Vector3, so i made a PoC
Mi think it's for silk?
Mi'm just entertained by things like this, and it's late on a friday
Tthe problem here isn't how to do the computation
yes, it is very much intended to be done as floating-point and then cast to integer
(compute the infinitely precise result and round to the nearest representable result)
the problem comes in from how to implement that support generically and correctly
especially over time using where DIM implementations need to be provided
there also lies a slightly more general issue in that not everything can be represented equivalently
for example,
likewise you end up not being able to represent infinites for trig functions
you end up with many other computations resulting in 0 or 1, regardless of input, due to it requiring fractional results
and so many operations functionally end up "pointless" for integrals
yes, it is very much intended to be done as floating-point and then cast to integer
(compute the infinitely precise result and round to the nearest representable result)
the problem comes in from how to implement that support generically and correctly
especially over time using where DIM implementations need to be provided
there also lies a slightly more general issue in that not everything can be represented equivalently
for example,
Sqrt(-1.0f) produces NaN, but it would have to throw for integerlikewise you end up not being able to represent infinites for trig functions
you end up with many other computations resulting in 0 or 1, regardless of input, due to it requiring fractional results
and so many operations functionally end up "pointless" for integrals
