NFX_DATATYPES_ENABLE_SIMD option to control CPU-specific optimizations (useful for WebAssembly builds)target_compile_features for C++20 standard enforcementCALL_GRAPH and CALLER_GRAPH for faster documentation generation-Werror, /WX) for GCC/Clang and MSVC buildsfriend struct to friend class for std::numeric_limits<Decimal> specializationDecimal::fromString() with fast-path parsing for common cases (≤19 digits)Int128::fromString() with fast-path parsing for values fitting in 64 bits (≤20 digits)Int128.cpp, Decimal.cpp)Decimal(const char*) and Int128(const char*) constructors for C-string initializationInt128::toString() with platform-specific fast paths and pre-allocated digit bufferInt128(const Decimal&) constructor using power-of-10 lookup tables instead of iterative multiplicationInt128::operator/ with _udiv128 hardware intrinsic for 128/64-bit division_BitScanReverse64 to reduce loop iterationsDOUBLE_POWERS_OF_10 lookup table for Decimal::toDouble() conversion.Decimal(float) and Decimal(double) constructors causing incorrect decimal displays (e.g., 273.15f showing as “273.149994”, 380.4 showing as “380.399999999999977”)Decimal::toDouble() conversion by replacing iterative division loop with single division operationInt128::operator<< by delegating to toString()Int128: Cross-platform 128-bit signed integer arithmetic
__int128 support on GCC/Clang for maximum performancefromString(), toString())toLow(), toHigh(), toBits())Decimal: High-precision decimal arithmetic
fromString(), toString())abs(), trunc(), floor(), ceil(), round())scale(), decimalPlacesCount())Documentation
Testing & Benchmarking