Optimize Decimal::sqrt using ilog10

[mk0walsk]

Resolves the TODO in Decimal::sqrt by replacing the iterative precision search with a direct calculation using ilog10.

Changes:

• Implements Uint128::ilog10 (delegating to u128::checked_ilog10).
• Updates Decimal::sqrt to calculate optimal precision in O(1) time using the derived formula: min(9, (38 - log10(atomics)) / 2).

Impact:

• Significantly reduces gas costs for large numbers by eliminating the loop that performed expensive isqrt and multiplication checks.
• Maintains correctness and robustness with a fallback mechanism for edge-case boundary overflows.