C++ code for uint128_t divide_uint128_by_uint64(...)

richter__dan — Sat, 04 Jan 2025 00:21:17 GMT

How to program function in C++ in Visual Studio 2022

uint128_t divide_uint128_by_uint64(uint128_t dividend, uint64_t divisor, uint64_t* remainder)

without using big number libraries? Note the result can exceed 64-bit i.e. I can't use _udiv128 intrinsic.

Do you know of a link to the Intel sample code?

Re: C++ code for uint128_t divide_uint128_by_uint64(...)

richter__dan — Sat, 04 Jan 2025 10:03:08 GMT

What do you think about this code:

struct uint128_t { uint64_t hi; uint64_t lo; };

uint128_t divide_uint128_by_uint64(uint128_t dividend, uint64_t divisor, uint64_t* remainder) { if (divisor == 0) { throw std::runtime_error("Division by zero"); } uint128_t quotient = { 0, 0 }; // Perform the division manually for uint128_t uint64_t temp_rem = 0; for (int i = 127; i >= 0; --i) { temp_rem = (temp_rem << 1) | ((dividend.hi >> (i - 64)) & 1); if (temp_rem >= divisor) { quotient.hi |= (1ULL << (i - 64)); temp_rem -= divisor; } } for (int i = 63; i >= 0; --i) { temp_rem = (temp_rem << 1) | ((dividend.lo >> i) & 1); if (temp_rem >= divisor) { quotient.lo |= (1ULL << i); temp_rem -= divisor; } } *remainder = temp_rem; return quotient; }

topic C++ code for uint128_t divide_uint128_by_uint64(...) in Intel® oneAPI Math Kernel Library

C++ code for uint128_t divide_uint128_by_uint64(...)

Re: C++ code for uint128_t divide_uint128_by_uint64(...)