free performance: autobatching in my SFML fork -- Vittorio Romeo

By Vittorio Romeo | May 9, 2025 08:33 AM | Tags: None

This article shows how a very simple automatic batching strategy can be applied on top of SFML without affecting the library API, resulting in free performance for the end user!

free performance: autobatching in my SFML fork

by Vittorio Romeo

From the article:

In one of my previous articles, I discussed the design and implementation of the batching system in my fork of SFML. Wouldn’t it be nice if drawables were automatically batched, whenever possible?

AoS vs SoA in practice: particle simulation -- Vittorio Romeo

By Vittorio Romeo | May 9, 2025 08:32 AM | Tags: None

This article presents a practical benchmark of a particle simulation using both the AoS (Array of Structures) and SoA (Structure of Arrays) data layouts. How much performance can we gain by merely switching up the way our data is stored?

AoS vs SoA in practice: particle simulation

by Vittorio Romeo

From the article:

The benchmark simulates a large number of 2D particles that continuously change position, scale, opacity, and rotation. Through an ImGui-based UI1, you can choose the number of particles, toggle multithreading, and switch between AoS and SoA on the fly.

A demo is worth a thousand words, and since my fork of SFML supports Emscripten, you can try the benchmark directly in your browser. Play around with all the options – I’m curious to hear what results you get! [...]

Speeding up C++ Code with Template Lambdas -- Daniel Lemire

By Blog Staff | May 5, 2025 01:51 PM | Tags: None

Integer division is one of the most expensive operations in C++, but when the divisor is known at compile time, the compiler can optimize it significantly. This post explores different approaches—using templates, lambda expressions, and template metaprogramming—to speed up division while maintaining clean and efficient code.

Speeding up C++ Code with Template Lambdas

by Daniel Lemire

From the article:

Let us consider a simple C++ function which divides all values in a range of integers:

void divide(std::span<int> i, int d) {
 for (auto& value : i) {
 value /= d;
 }
}

A division between two integers is one of the most expensive operations you can do over integers: it is much slower than a multiplication which is, in turn, more expensive than an addition. If the divisor d is known at compile-time, this function can be much faster. E.g., if d is 2, the compiler might optimize away the division and use a shift and a few cheap instructions instead. The same is true with all compile-time constant: the compiler can often do better knowing the constant. (See Lemire et al., Integer Division by Constants: Optimal Bounds, 2021)

C++26: Removing Language Features -- Sandor Dargo

By Blog Staff | May 1, 2025 01:48 PM | Tags: None

C++ is often seen as an ever-growing language, with each new standard introducing powerful features while maintaining backward compatibility. However, C++26 takes a step toward simplification by officially removing deprecated features, including implicit arithmetic conversions for enumerations and direct comparisons of C-style arrays, both of which previously led to unintended behavior.

C++26: Removing Language Features

by Sandor Dargo

From the article:

Probably you all heard that C++ is an ever-growing language - I wrote so many times as well. Each standard indeed comes with a great bunch of highly-anticipated features. At the same time, due to binary compatibility considerations, very few old features are removed. This has several implications:

• we have probably more than one way to do something
• the standard keeps growing

This is true, but let’s not forget that each new standard removes some features. In this post, let’s review what are the language features that are removed in C++26 and in a later post, we’ll have a look at the removed library features.

At this point, it’s worth mentioning that a removal from the language usually happens in two steps. First a feature gets deprecated, meaning that its users would face compiler warnings for using deprecated features. As a next step, which in some cases never comes, the compiler support is finally removed.

 

Bjarne Stroustrup on How He Sees C++ Evolving -- David Cassel

By Blog Staff | Apr 29, 2025 01:44 PM | Tags: None

Bjarne Stroustrup, creator of C++, is advocating for the adoption of guideline-enforcing profiles to enhance the language's safety and security.

Bjarne Stroustrup on How He Sees C++ Evolving

by David Cassel

From the article:

“I wanted to educate the community at large, and members of WG21 in particular — on my views of C++’s intended direction of evolution,” Bjarne Stroustrup told TNS.

The 74-year-old creator of C++ has spent 40 years of his life watching the growth of the language he designed back in 1985.

To encourage some long-desired features, last month in Communications of the ACM Stroustrup published “21st Century C++“, a 6,300-word article promising to show “key concepts” for modern, type-safe “21st-century C++” to create “C++ on steroids”. For example, in the article Stroustrup highlighted long-standing experiments in approaches like writing safer code with guideline-enforcing profiles. To maintain compatibility with decades of already-written C++ code, “We can’t change the language,” Stroustrup writes. “But we can change the way it is used…”

Yet that evolution isn’t entirely up to him. In a section towards the end, Stroustrup acknowledges WG21, the standardization working group, and how it will inevitably play a role in how much the language can change. 

Details of std::mdspan from C++23 -- Bartlomiej Filipek

By Blog Staff | Apr 25, 2025 01:34 PM | Tags: None

In this article, we’ll see details of std::mdspan, a new view type tailored to multidimensional data. We’ll go through type declaration, creation techniques, and options to customize the internal functionality.

Details of std::mdspan from C++23

by Bartlomiej Filipek

From the article:

In this article, we’ll see details of std::mdspan, a new view type tailored to multidimensional data. We’ll go through type declaration, creation techniques, and options to customize the internal functionality.

Type declaration 

The type is declared in the following way:

template<  
     class T,  
     class Extents,  
     class LayoutPolicy = std::layout_right,  
     class AccessorPolicy = std::default_accessor<T> 
> class mdspan; 

And it has its own header <mdspan>.

The main proposal for this feature can be found at https://wg21.link/P0009

Following the pattern from std::span, we have a few options to create mdspan: with dynamic or static extent.

The key difference is that rather than just one dimension, we can specify multiple. The declaration also gives more options in the form of LayoutPolicy and AccessorPolicy. More on that later.

C++26: No More UB in Lexing -- Sandor Dargo

By Blog Staff | Apr 23, 2025 01:34 PM | Tags: None

Undefined behavior in C++ is a well-known source of headaches for developers, but surprisingly, even the lexing process contained cases of it—until now. Thanks to P2621R3 by Corentin Jabot, unterminated strings, macro-generated universal character names, and spliced UCNs are now formally defined, aligning the standard with real-world compiler behavior.

C++26: No More UB in Lexing

by Sandor Dargo

From the article:

If you ever used C++, for sure you had to face undefined behaviour. Even though it gives extra freedom for implementers, it’s dreaded by developers as it may cause havoc in your systems and it’s better to avoid it if possible.

Surprisingly, even the lexing process in C++ can result in undefined behaviour. Thanks to Corentin Jabot’s work and his P2621R3 that won’t be the case anymore. As it was accepted as a defect report starting from C++98, in fact, you benefit from this already if you use a new enough compiler.

Truth be told, compilers didn’t do any dangerous. They handled the below cases safely and deterministically. So this change is really about updating the standard and matching implementers’ work.

Let’s quickly see the three cases.

Unterminated strings

     // unterminated string used to be UB
     const char * foo = "

Who would have thought that an unterminated string or a character was UB?! Despite the permissive standard, all major compilers identified it as ill-formed. From now on, even the standard says so.

 

On Trying to Log an Exception as it Leaves your Scope -- Raymond Chen

By Blog Staff | Apr 21, 2025 01:22 PM | Tags: None

A customer attempted to log exceptions using a scope_exit handler, expecting it to capture and process exceptions during unwinding. However, they encountered crashes because ResultFromCaughtException requires an actively caught exception, which isn’t available during unwinding—leading to an unexpected termination.

On Trying to Log an Exception as it Leaves your Scope

by Raymond Chen

From the article:

A customer wanted to log exceptions that emerged from a function, so they used the WIL scope_exit object to specify a block of code to run during exception unwinding.

void DoSomething()
{
    auto logException = wil::scope_exit([&] {
        Log("DoSomething failed",
            wil::ResultFromCaughtException());
    });

    ⟦ do stuff that might throw exceptions ⟧

    // made it to the end - cancel the logging
    logException.release();
}

They found, however, that instead of logging the exception, the code in the scope_exit was crashing.

They debugged into the Result­From­Caught­Exception function, which eventually reaches something like this:

try
{
    throw;
}
catch (⟦ blah blah ⟧)
{
    ⟦ blah blah ⟧
}
catch (⟦ blah blah ⟧)
{
    ⟦ blah blah ⟧
}
catch (...)
{
    ⟦ blah blah ⟧
}

The idea is that the code rethrows the exception, then tries to catch it in various ways, and when it is successful, it uses the caught object to calculate a result code.

Bit Fields, Byte Order and Serialization -- Wu Yongwei

By Blog Staff | Apr 17, 2025 01:12 PM | Tags: None

Network packets can be represented as bit fields. Wu Yongwei explores some issues to be aware of and offers solutions.

Bit Fields, Byte Order and Serialization

by Wu Yongwei

From the article:

n order to store data most efficiently, the C language has supported bit fields since its early days. While saving a few bytes of memory isn’t as critical today, bit fields remain widely used in scenarios like network packets. Endianness adds complexity to bit field handling – especially since network packets are typically big-endian, while most modern architectures are little-endian. This article explores these problems and their solutions, including my reflection-based serialization project.

Memory layout of bit fields

The memory layout of bit fields is implementation-defined. In a typical little-endian environment, bit fields start from the lower bits of the lower byte and extend toward higher bits and bytes. In a typical big-endian environment, bit fields start from the higher bits of the lower byte and extend toward lower bits and higher bytes.

Let’s consider a practical scenario. Suppose we want to use a 32-bit integer to store a date. How should we achieve this? A simple approach is to store the number of days from a fixed point of time (e.g. 1 January 1900). We can calculate the number of years that can be expressed as follows:

However, with this approach, extracting specific year, month, and day information becomes very cumbersome. A simpler way is to store the year, month, and day as bit fields. We can define the following struct, using only 32 bits:

  struct Date {
    int      year  : 23;
    unsigned month : 4;
    unsigned day   : 5;
  };

Our intention is to use a 23-bit signed integer for the year (ranging from -4,194,304 to 4,194,303), a 4-bit unsigned integer for the month (0–15, covering legal values 1–12), and a 5-bit unsigned integer for the day (0–31, covering legal values 1–31). This representation is similarly compact, with a slightly narrower range, but it’s quite sufficient and much more convenient for many common usages (excepting interval calculation).

Creating a Generic Insertion Iterator, Part 2 -- Raymond Chen

By Blog Staff | Apr 15, 2025 01:11 PM | Tags: None

Last time, our generic insertion iterator failed to meet the requirements of default constructibility and assignability because it stored the lambda directly. To fix this, we now store a pointer to the lambda instead, ensuring the iterator meets standard requirements while still allowing flexible insertion logic.

Creating a Generic Insertion Iterator, Part 2

by Raymond Chen

From the article:

Last time, we tried to create a generic insertion iterator but ran into trouble because our iterator failed to satisfy the iterator requirements of default constructibility and assignability.

We ran into this problem because we stored the lambda as a member of the iterator.

So let’s not do that!

Instead of saving the lambda, we’ll just save a pointer to the lambda.

template<typename Lambda>
struct generic_output_iterator
{
    using iterator_category = std::output_iterator_tag;
    using value_type = void;
    using pointer = void;
    using reference = void;
    using difference_type = void;

    generic_output_iterator(Lambda&& lambda) noexcept :
        insert(std::addressof(lambda)) {}

    generic_output_iterator& operator*() noexcept
        { return *this; }
    generic_output_iterator& operator++() noexcept
        { return *this; }
    generic_output_iterator& operator++(int) noexcept
        { return *this; }

    template<typename Value>
    generic_output_iterator& operator=(
        Value&& value)
    {
        (*insert)(std::forward<Value>(value));
        return *this;
    }

protected:
    Lambda* insert;

};

template<typename Lambda>
generic_output_iterator<Lambda>
generic_output_inserter(Lambda&& lambda) noexcept {
    return generic_output_iterator<Lambda>(
        std::forward<Lambda>(lambda));
}

template<typename Lambda>
generic_output_iterator(Lambda&&) ->
    generic_output_iterator<Lambda>;

This requires that the lambda remain valid for the lifetime of the iterator, but that may not a significant burden. Other iterators also retain references that are expected to remain valid for the lifetime of the iterator. For example, std::back_inserter(v) requires that v remain valid for as long as you use the inserter. And if you use the iterator immediately, then the requirement will be satisfied: