Communications of the ACM
Refine your search:
How the “Frontier” Became the Slogan of Uncontrolled AI
Artificial intelligence (AI) has been billed as the next frontier of humanity: the newly available expanse whose exploration …
From Daniel Lemire's Blog
Compressing floating-point numbers quickly by converting them to brain floats
We sometimes have to work a large quantity of floating-point numbers. This volume can be detrimental to performance. Thus we often want to compress these numbers...
From Daniel Lemire's Blog
Parsing tiny and very large floating-point values: a programming-language comparison
Most programming languages support floating-point numbers. You typically have the ability to turn a string into a floating-point number. E.g., “3.1416” could be...
From Daniel Lemire's Blog
Faster random integer generation with batching
We often generate random integers. Quite often these numbers must be within an interval: e.g., an integer between 0 and 100. One application is a random shuffle...
From Daniel Lemire's Blog
Reflection-based JSON in C++ at Gigabytes per Second
JSON (JavaScript Object Notation) is a popular format for storing and transmitting data. It uses human-readable text to represent structured data in the form of...
From Daniel Lemire's Blog
Converting ASCII strings to lower case at crazy speeds with AVX-512
AMD Zen 4 and Zen 5, as well as server-side recent Intel processors, support an advanced set of instructions called AVX-512. They are powerful SIMD (Single Instruction...
From Daniel Lemire's Blog
Evolution of iPhone storage capacity
People who should know better often underestimate how fast our storage capacity has grown. We have been able to get 1 TB of storage on iPhones for the last three...
From Daniel Lemire's Blog
Storage costs are plummeting
Storage costs are plummeting like a skydiver in freefall—between 10 and 100 times cheaper with each passing decade. Meanwhile, the programmer population is growing...
From Daniel Lemire's Blog
How big are your docker images?
Docker is a standard to deploy software on the cloud. Developers start with an existing image and add their own code before deploying their systems. How big are...
From Daniel Lemire's Blog
How much of your binary executable is just ASCII text?
We sometimes use binary executable which can span megabytes. I wondered: how much text is contained in these binary files? To find out, I wrote a Python scriptContinue...
From Daniel Lemire's Blog
Safer code in C++ with lifetime bounds
For better performance in software, we avoid unnecessary copies. To do so, we introduce references (or pointers). An example of this ideas in C++ is the std::string_view...
From Daniel Lemire's Blog
Does C++ allow template specialization by concepts?
Recent versions of C++ (C++20) have a new feature: concepts. A concept in C++ is a named set of requirements that a type must satisfy. E.g., ‘act like a string’...
From Daniel Lemire's Blog
Scan HTML even faster with SIMD instructions (C++ and C#)
Earlier this year, both major Web engines (WebKit/Safari and Chromium/Chrome/Edge/Brave) accelerated HTML parsing using SIMD instructions. These ‘SIMD’ instructions...
From Daniel Lemire's Blog
Benchmarking ARM processors: Graviton 4, Graviton 3 and Apple M2
The world of commodity processor is roughly divided in two: x64 chips for servers and PCs, and ARM processors for mobile devices. However, ARM chips increasingly...
From Daniel Lemire's Blog
Scan HTML faster with SIMD instructions: .NET/C# Edition
Recently, the two major Web engines (WebKit and Chromium) adopted fast SIMD routines to scan HTML content. The key insight is to use vectorized classification (Langdale...
From Daniel Lemire's Blog
Performance tip: avoid unnecessary copies
Copying data in software is cheap, but it is not at all free. As you start optimizing your code, you might find that copies become a performance bottleneck. Let...
From Daniel Lemire's Blog
Validating gigabytes of Unicode strings per second… in C#?
We have been working on a fast library to validate and transcode Unicode and other formats such as base64 in C++: simdutf. We wondered: could we achieve the same...
From Daniel Lemire's Blog
Rolling your own fast matrix multiplication: loop order and vectorization
If you must multiply matrices, you should use dedicated libraries. However, we sometimes need to roll our own code. In C++, you can quickly write your own Matrix...
From Daniel Lemire's Blog
Scan HTML faster with SIMD instructions: Chrome edition
Modern processors have instructions to process several bytes at once. Effectively all processors have the capability of processing 16 bytes one once. These instructions...
From Daniel Lemire's Blog
Quickly checking whether a string needs escaping
In software, we often represent strings by surrounding them with quotes ("). What happens if the string itself contains quotes? We then need to escape the string...
From Daniel Lemire's Blog