FMOD LLK implementation

[KalaivaniMCW]

Create a binary_sfpu LLK for FMOD operation

• The accuracy with the composite FMOD is not very good.
• This maybe due to the involvement of div along with data-loss at every intermediate level due to rounding.
• We need DST_ACCUM_MODE flag passing in as well to handle result rounding for non-fp32 float datatypes

Goal:
To implement a native LLK kernel running on SFPU and test accuracy.

[VirdhatchaniKN]

Update
#37050 - PR Ready with approvals. Will be merged once pipelines are clear

[VirdhatchaniKN]

Update
#37050 - PR Ready with approvals. Ready to merge

• ULP Testing : Link
• Perf Comparison for L1 Interleaved config (Device Kernel Duration)

Shape	Main branch DKD (ns)	Proposed branch DKD (ns)	Kernel Duration down by
Single tile (32, 32)	19682	4614	326.57%
8 tile (64, 128)	27016	4811	461.55%