Add fltflt rounding and fmod functions

[tbensonatl]

Add support for the following float-float (fltflt) functions:

• Round toward nearest, with ties toward even
• Truncate toward zero
• Truncate toward negative infinity
• fmod (floating point remainder)

Also includes are new unit tests and benchmarks for the newly introduced functions.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[greptile-apps]

Greptile Summary

This PR adds four new float-float arithmetic functions: round-to-nearest (with ties to even), truncate toward zero, floor (truncate toward negative infinity), and fmod (floating-point remainder). The implementation includes proper edge case handling with zero-division guards returning NaN, consistent use of fabsf for float operations, and comprehensive test coverage.

Key changes:

• Added fltflt_round_to_nearest(), fltflt_round_toward_zero(), fltflt_floor(), and fltflt_fmod() functions
• Added optimized fltflt_add_same_sign() for same-sign operands
• Made fltflt constructors constexpr for compile-time evaluation
• Comprehensive unit tests covering positive/negative values, zero divisors, and high-precision cases
• Performance benchmarks for all new functions

Previous feedback addressed:
All issues from previous review threads have been resolved - zero-division guards are in place, fabsf is used consistently, and comprehensive tests have been added for fmod.

Confidence Score: 5/5

• This PR is safe to merge with no identified issues
• All previous feedback has been addressed. The implementation includes proper edge case handling, comprehensive test coverage (12+ test cases for fmod alone), performance benchmarks, and follows consistent coding patterns. The code uses appropriate precision functions (fabsf), guards against division by zero, and includes detailed documentation.
• No files require special attention

Important Files Changed

Filename	Overview
include/matx/kernels/fltflt.h	Adds four new rounding/fmod functions with proper zero-division guards and consistent use of fabsf. All previous feedback addressed.
test/00_misc/FloatFloatTests.cu	Comprehensive test coverage for all new functions including edge cases (negative values, zero divisor, high-precision cases).
bench/00_misc/fltflt_arithmetic.cu	Adds performance benchmarks for new rounding and fmod functions with appropriate test cases.
bench/scripts/run_fltflt_benchmarks.py	Updates benchmark list to include newly added functions (round, trunc, floor, fmod, cast operations).

_{Last reviewed commit: 7a37881}

[greptile-apps]

_{4 files reviewed, 4 comments}

_{Edit Code Review Agent Settings | Greptile}

[cliffburdick]

/build

[tbensonatl]

/build

[tbensonatl]

/build