Vectorize TensorPrimitives operations that are currently scalar only

[stephentoub]

#97192 adds many more operations onto TensorPrimitives. Some are fully vectorized, others currently only have a scalar implementation. The following operations need vectorization:

• Acos _(amd lib has vectorized float impl, possible approach in
  Vectorize TensorPrimitives.Acosh/sinh/tanh in terms of Log #98302) _
• Asin _(amd lib has vectorized float impl, possible approach in
  Vectorize TensorPrimitives.Acosh/sinh/tanh in terms of Log #98302) _
• Atan _(amd lib has vectorized float impl, possible approach in
  Vectorize TensorPrimitives.Acosh/sinh/tanh in terms of Log #98302) _
• SinCos (amd lib has vectorized double impl)
• Acosh
• Asinh
• Atanh
• Atan2
• ILogB
• SinCosPi
• TanPi
• ... a bunch more

[ghost]

Tagging subscribers to this area: @dotnet/area-system-numerics-tensors
See info in area-owners.md if you want to be subscribed.

Issue Details

---

#97192 adds many more operations onto TensorPrimitives. Some are fully vectorized, others currently only have a scalar implementation. The following operations need vectorization:

• Exp2
• Exp10
• Pow
• Cbrt
• Hypot
• Acos
• Acosh
• Asin
• Asinh
• Atan
• Atanh
• Atan2
• Cos
• Sin
• Tan
• Log10
• Log(x,y)
• PopCount
• LeadingZeroCount
• TrailingZeroCount
• ILogB
• RootN
• RotateLeft
• RotateRight
• Round
• ScaleB
• ShiftLeft
• ShiftRightArithmetic
• ShiftRightLogical
• SinCos
• SinCosPi

Author:	stephentoub
Assignees:	-
Labels:	area-System.Numerics.Tensors, tenet-performance
Milestone:	9.0.0