C++ Basics: Choose the Accumulator Type Explicitly

Use the initial value to control numeric accumulation instead of accidentally truncating fractional values.

Prepare a scratch directory

Tested on Linux x86_64 with GCC 14.2.0. Install a C++20-capable compiler and standard-library development headers, with g++ on PATH. You also need Bash, a text editor, and the mktemp, rm, and rmdir utilities. No third-party library is required.

Keep the same terminal open. Stop if setup fails, before saving the source:

start_directory="$PWD"
work_directory="$(mktemp -d)" || exit 1
cd -- "$work_directory" || exit 1

Run the example

Save this as accumulate-initial-type-demo.cpp:

#include <array>
#include <iomanip>
#include <iostream>
#include <numeric>

int main() {
    const std::array<double, 3> readings{1.5, 2.5, 3.5};
    const int truncated_total = std::accumulate(readings.begin(), readings.end(), 0);
    const double precise_total = std::accumulate(readings.begin(), readings.end(), 0.0);
    std::cout << "Integer initial value: " << truncated_total << '\n';
    std::cout << std::fixed << std::setprecision(1)
              << "Double initial value: " << precise_total << '\n';
}

Compile and run it. The second command runs only after compilation succeeds:

g++ -std=c++20 -Wall -Wextra -Werror -pedantic accumulate-initial-type-demo.cpp -o accumulate-initial-type-demo &&
./accumulate-initial-type-demo

Observed output:

Integer initial value: 6
Double initial value: 7.5

Read the result

The initial literal 0 makes the accumulator an int. Each intermediate result is converted back to that type, producing six. The literal 0.0 selects double and retains the half-unit fractions, producing 7.5.

Apply it carefully

The return variable does not choose the internal accumulator type; the initial argument does. Pick a type with enough range as well as suitable precision. This small example keeps all conversions representable. A floating accumulator does not guarantee exact arithmetic for arbitrary decimal data, and an integer accumulator can overflow if its range is exceeded.

The C++ working draft describes the relevant library rules. This example uses only C++20 facilities, even though the linked draft also documents newer additions.

Clean up

Remove only the source and executable created here, then remove the empty practice directory:

rm -- accumulate-initial-type-demo.cpp accumulate-initial-type-demo
cd -- "$start_directory" || exit 1
rmdir -- "$work_directory"
unset start_directory work_directory

If the directory is not empty, inspect its remaining contents before removing anything else.

Acknowledgements

This post was created with assistance from AI.