perf: optimize range checks for small number of small field ops

[ivokub]

Description

Additionally, sets MulNoReduce as default when calling Field.Mul on small field parameter.

Fixes #1698

Type of change

• Bug fix (non-breaking change which fixes an issue)

How has this been tested?

• Test A
• Test B

How has this been benchmarked?

• Benchmark A, on Macbook pro M1, 32GB RAM
• Benchmark B, on x86 Intel xxx, 16GB RAM

Checklist:

• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• I did not modify files generated from templates
• golangci-lint does not output errors locally
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published in downstream modules

---

Note

Medium Risk
Changes Field.Mul semantics for single-limb emulated fields by defaulting to non-reducing multiplication, which can affect callers that assumed Mul always reduced (now must use MulMod for that guarantee). Adds stateful range-check deduplication and a threshold-triggered switch in range-checking behavior, which impacts constraint generation and overflow handling in hot paths.

Overview
Optimizes emulated-field range checks and small-field multiplication. Range checks now go through a new Field.rangeCheck helper that deduplicates checks per limb (tracking the widest enforced bit-width) and counts how many checks have been emitted.

For small-field emulation, range checking switches only after thresholdForInexactOverflow checks to a base-length (16-bit) range checker, and smallAdditionalOverflow() now returns non-zero only after that threshold to keep overflow accounting consistent.

Field.Mul now defaults to MulNoReduce when useSmallFieldOptimization() is enabled, falling back to reducing multiplication only when overflow would exceed limits; callers needing guaranteed modular reduction should use MulMod. Tests/benchmarks are adjusted accordingly (bench compile loop guard + suppress ns/op, and malicious-hint circuit uses MulMod).

^{Written by Cursor Bugbot for commit c994745. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

[YaoJGalteland]

LGTM