Fix abs Operator int accuracy issue

[sgbihu]

Details:

• Current Abs operator's implementation

1. Convert int32 to float
2. Apply sign mask
3. Convert float to int32

Step 3 will result in a precision loss.

Tickets:

• CVS-171704

[Copilot]

Pull Request Overview

This pull request fixes precision loss in the absolute value (abs) operator for integer data types by replacing the DNNL-based implementation with a custom JIT emitter that handles integer precision natively.

• Replaces jit_dnnl_aux_emitter with a dedicated jit_abs_emitter to avoid float conversion precision loss
• Implements direct integer absolute value operations using CPU-specific SIMD instructions (pabsd/vpabsd)
• Adds test coverage for the abs operator with integer precision

Reviewed Changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
activation.cpp	Adds abs operator to integer activation types test list
jit_uni_eltwise_generic.cpp	Updates emitter mapping to use new jit_abs_emitter
jit_eltwise_emitters.hpp	Declares new jit_abs_emitter class
jit_eltwise_emitters.cpp	Implements jit_abs_emitter with ISA-specific optimizations
jit_dnnl_ext_emitters.hpp	Removes old jit_abs_emitter class definition

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[maxnick]

@aobolensk , could you please take a look?