{"id":10026,"date":"2024-09-16T09:29:02","date_gmt":"2024-09-16T09:29:02","guid":{"rendered":"https:\/\/www.modernescpp.com\/?p=10026"},"modified":"2025-07-03T15:36:45","modified_gmt":"2025-07-03T15:36:45","slug":"an-overview-of-c26-concurrency","status":"publish","type":"post","link":"https:\/\/www.modernescpp.com\/index.php\/an-overview-of-c26-concurrency\/","title":{"rendered":"An Overview of C++26: Concurrency"},"content":{"rendered":"\n

Today, I will finish my overview of C++26 and write about concurrency.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

There are still two library features left before I jump into the concurrency: saturation arithmetic and debugging support. <\/p>\n\n\n\n

Saturation Arithmetic <\/h2>\n\n\n\n

Saturation arithmetic<\/strong> is a version of arithmetic<\/a> in which all operations, such as addition<\/a> and multiplication<\/a>, are limited to a fixed range between a minimum and maximum value.<\/em><\/p>\n\n\n\n

If the result of an operation is greater than the maximum, it is set (“clamped<\/a>“) to the maximum; if it is below the minimum, it is clamped to the minimum. The name comes from how the value becomes “saturated” once it reaches the extreme values; further additions to a maximum or subtractions from a minimum will not change the result.<\/em> (https:\/\/en.wikipedia.org\/wiki\/Saturation_arithmetic)<\/a><\/p>\n\n\n\n

C++26 introduced a set of saturating arithmetic operations: addition, subtraction, multiplication, division, and saturate cast<\/strong>. If the specified integral type T <\/code>cannot represent the result of the operation, the result is instead std::numeric_limits::min<T>()<\/code> or std::numeric_limits::max<T>() <\/code>(whichever is closer).<\/p>\n\n\n\n

cppreference.com<\/a> has a nice example with an explanation to std::add_sat<\/code>.<\/p>\n\n\n\n

#include <climits><\/span>\n#include <limits><\/span>\n#include <numeric><\/span>\n \nstatic_assert(CHAR_BIT ==<\/span> 8<\/span>);\nstatic_assert(UCHAR_MAX ==<\/span> 255<\/span>);\n \nint<\/span> main<\/span>()\n{\n    constexpr int<\/span> a =<\/span> std::<\/span>add_sat(3<\/span>, 4<\/span>); \/\/ no saturation occurs, T = int<\/span>\n    static_assert(a ==<\/span> 7<\/span>);\n \n    constexpr unsigned<\/span> char<\/span> b =<\/span> std::<\/span>add_sat<<\/span>unsigned<\/span> char<\/span>><\/span>(UCHAR_MAX, 4<\/span>); \/\/ saturated<\/span>\n    static_assert(b ==<\/span> UCHAR_MAX);\n \n    constexpr unsigned<\/span> char<\/span> c =<\/span> std::<\/span>add_sat(UCHAR_MAX, 4<\/span>); \/\/ not saturated, T = int<\/span>\n        \/\/ add_sat(int, int) returns int tmp == 259,<\/span>\n        \/\/ then assignment truncates 259 % 256 == 3<\/span>\n    static_assert(c ==<\/span> 3<\/span>);\n \n\/\/  unsigned char d = std::add_sat(252, c); \/\/ Error: inconsistent deductions for T<\/span>\n \n    constexpr unsigned<\/span> char<\/span> e =<\/span> std::<\/span>add_sat<<\/span>unsigned<\/span> char<\/span>><\/span>(251<\/span>, a); \/\/ saturated<\/span>\n    static_assert(e ==<\/span> UCHAR_MAX);\n        \/\/ 251 is of type T = unsigned char, `a` is converted to unsigned char value;<\/span>\n        \/\/ might yield an int -> unsigned char conversion warning for `a`<\/span>\n \n    constexpr signed<\/span> char<\/span> f =<\/span> std::<\/span>add_sat<<\/span>signed<\/span> char<\/span>><\/span>(-<\/span>123<\/span>, -<\/span>3<\/span>); \/\/ not saturated<\/span>\n    static_assert(f ==<\/span> -<\/span>126<\/span>);\n \n    constexpr signed<\/span> char<\/span> g =<\/span> std::<\/span>add_sat<<\/span>signed<\/span> char<\/span>><\/span>(-<\/span>123<\/span>, -<\/span>13<\/span>); \/\/ saturated<\/span>\n    static_assert(g ==<\/span> std::<\/span>numeric_limits<<\/span>signed<\/span> char<\/span>>::<\/span>min()); \/\/ g == -128<\/span>\n}\n<\/pre><\/div>\n\n\n\n
T is the type of both function arguments:   <\/p>\n\n\n\n
template<\/span><<\/span> class<\/span> T<\/span> ><\/span>\nconstexpr T add_sat( T x, T y ) noexcept;\n<\/pre><\/div>\n\n\n\n
 debugging Support <\/h2>\n\n\n\n
C++26 has three functions for debugging. <\/p>\n\n\n\n