together, and it’s a major bug source, which is why I’m saying stay as simple as you can, use {signed} integers till you really need something else.” ~Bjarne StroustrupAlex Dathskovsky | alex.dathskovsky@speedata linkedin.com/in/alexdathskovsky “The need for signed integer arithmetic is often misplaced as most integers never represent negative values within a program. The indexing of arrays and iteration count of well. There should be a propensity to use unsigned integers more often than signed, yet despite this, most coders incorrectly choses to use signed integers almost exclusively.” ~ Dale Weiler https://graphitemaster

C++ CppCon 2024-09-19 9Example: Biggest odd magnitude Problem statement Input Non-empty list of integers Output The biggest magnitude of an odd integer in the list Jonathan Müller — @foonathan Back to C++ CppCon 2024-09-19 10Example: Biggest odd magnitude Problem statement Input Non-empty list of integers Output The biggest magnitude of an odd integer in the list Input [3, 0, 2, -1, 5, -7, 8] Output Programming in C++ CppCon 2024-09-19 25Example: Find sum of two integers Problem statement Input Two lists of integers, a number x Output The two integers from the list whose sum is x Jonathan Müller — @foonathan

discrete_distribution( InputIt first, InputIt last ); cppreference.com Constructs the distribution on integers with weights in the range [first, last)Sample a position in a matrix randomly 12 std::pairintegers with weights in the range [first, last)Sample a position in a matrix randomly 12 std::pairintegers with weights in the range [first, last) Eigen: STL iterators are intrinsically designed to iterate

Detecting ties becomes a divisibility check by using the minverse algorithm 5F Optimisation for integers Optimisation for c The dragon is in the detailsa = ⌊m × ′ 10F 2E−1 ⌋ can be precomputed Detecting ties becomes a divisibility check by using the minverse algorithm 5F Optimisation for integers Optimisation for c Other tricks The dragon is in the detailsIntel i7 (10510U) - gcc 13.2.1 Benchmarks 1Benchmarks - integers (a) Intel i7 (10510U) - gcc 13.2.1Benchmarks - integers (b) teju x dragonbox wins ties losses 100.0% 0.0% 0.0% Intel i7 (10510U) - gcc 13.2.1Benchmarks - integers (b) teju

PROBLEM CONSTRAINTS 64-bit integers ⇒ we can't just do "largest + 1" or similar ⇒ totally ordered 9 . 1RESTATE THE PROBLEM CONSTRAINTS RESTATE THE PROBLEM CONSTRAINTS 64-bit integers ⇒ we can't just do "largest ordered unique elements 9 . 1RESTATE THE PROBLEM CONSTRAINTS RESTATE THE PROBLEM CONSTRAINTS 64-bit integers ⇒ we can't just do "largest + 1" or similar ⇒ totally ordered unique elements random order ⇒ can do this in place 9 . 1RESTATE THE PROBLEM CONSTRAINTS RESTATE THE PROBLEM CONSTRAINTS 64-bit integers ⇒ we can't just do "largest + 1" or similar ⇒ totally ordered unique elements random order ⇒

specific object type ● 14 ● px is a pointer ● px’s type is int* ○ px can stores addresses of integers.What is a Pointer? (3/8) ● A pointer is a variable that stores the memory address of a specific specific object type ● 15 ● px is a pointer ● px’s type is int* ○ px can stores addresses of integers. ● I retrieve the address of a variable using the ampersand operator (&) ○ You could also use &(x) specific object type 16 ● px is a pointer ● px’s type is int* ○ px can stores addresses of integers. ● I retrieve the address of a variable using the ampersand operator (&) ○ You could also use

separating the higher half from the lower half of a data sample • If I have an array a of seven integers, and they are sorted, the median is a[3] • Median filters are good at • Preserving edge features -march=skylake • Element counts of 1E02 through 1E07 (by 10s) • Pre-sorted integers increasing monotonically • Randomly generated signed integers [-50, +50] • Always generated from the same seed • Collect timings

vectorizable types • add user-defined numeric types to vectorizable types (fixed-point, saturating integers, etc.) Matthias Kretz CppCon ’23 46 GSI Helmholtz Center for Heavy Ion ResearchMotivation std::simd overhead 1 mov eax, DWORD PTR [rdi] 2 imul eax, eax 3 mov DWORD PTR [rdi], eax • src/dst: array of integers • throughput: 0.5/1/1 cycles (Intel) • integer multiplications: 1 1 vmovdqu32 zmm0, ZMMWORD PTR PTR [rdi] 2 vpmulld zmm0, zmm0, zmm0 3 vmovdqu32 ZMMWORD PTR [rdi], zmm0 • src/dst: array of integers • throughput: 0.5/1/1 cycles (Intel) • integer multiplications: 16 Take-Away #8 SIMD is relevant

enumerations In traditional C++, enumerated types are not type-safe, and enumerated types are treated as integers, which allows two completely different enumerated types to be directly compared (although the compiler foooooo. o{1,} is equivalent to o+. o{0,} is equivalent to o*. {n,m} m and n are non-negative integers, where n is less than or equal to m. Matches at least n times and matches up to m times. For E.g: std::atomic counter; And provides basic numeric member functions for atomic types of integers or floating-point numbers, for example, Including fetch_add, fetch_sub, etc., and the corresponding

hard; refers to repo only 4C++20 Features: 60 -> 23 Left 5C++20 STL Features Part 1 6Comparing Integers 7Usual Arithmetic Conversions #include using namespace std; int main() { cout << boolalpha; • What does this print? • warning C4018: '<': signed/unsigned mismatch • warning: comparison of integers of different signs: 'short' and 'unsigned int' [-Wsign-compare] • This prints false 🙀 8Integer