Fixed some hlsl functions

[Przemog1]

• lp_norm partial spec is wrong/wont get called Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment) & Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• more requirements on lp_norm Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• dot should not constrain the dimensions or types
• cross_helper should allow non-float type vectorials
• make sure correct tgmath sqrt is being called throughout functions.hlsl Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• call nabla dot or dot_helper in lp_norm::__sum Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• nitpicks in lp_norm Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• beziers.hlsl should be using tgmath exp,exp2,log2 and friends Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment) also NBL_FLOAT64_LITERAL and stuffz
• loosen the requirements on reflect and refract also make them compile Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• use flipSign for sincos Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• use better concepts Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• condAbsOrMax Maths for HLSL BxDFs (template cmath, tgmath) #803 (comment)
• ideally applyChainRule should allow any Matricial, and call our templated mul (does it exist btw? or did it clash with HLSL mul )

[devshgraphicsprogramming]

beziers.hlsl should be using tgmath exp,exp2,log2 and friends #803 (comment) also NBL_FLOAT64_LITERAL and stuffz

@Przemog1 can you at least test example 62 ?