Best of Performance — September 2025

  1. 1
    Article
    Avatar of uiwf7xowejrt4dhwsv7xiPaul Vasile·31w

    RAM is for the entire computer not just for your shitty app

    A critique of developers who justify excessive RAM usage in applications by claiming "RAM is meant to be used" or "RAM is cheap." The author argues that RAM is a shared system resource, not exclusively for individual applications, and calls for common-sense optimization rather than resource-heavy solutions. The piece criticizes the trend of unoptimized code and poor user experience driven by cost-cutting development practices.

  2. 2
    Article
    Avatar of logrocketLogRocket·28w

    Yes, you should upgrade to TypeScript 5.9 — here’s why

    TypeScript 5.9 introduces significant improvements including a redesigned tsc --init command with sensible defaults, import defer syntax for lazy module loading, expandable hovers for better type exploration, and performance optimizations that deliver 11% faster type checking through caching intermediate type instantiations.

  3. 3
    Video
    Avatar of bigboxswebigboxSWE·29w

    RUST IS WINNING

    Rust programming language adoption is growing significantly, jumping from 9% to 15% developer usage according to Stack Overflow surveys. Major tech companies like Amazon, Microsoft, Google, Meta, and Twitter are implementing Rust in critical infrastructure for performance and reliability benefits. The language is also gaining traction in system utilities, with Ubuntu shipping core utilities in Rust and Python tooling like UV package manager being built with it. Despite the learning curve challenges, Rust offers compelling advantages for developers willing to invest in low-level programming skills.

  4. 4
    Article
    Avatar of freekFREEK.DEV·30w

    SQL performance improvements: finding the right queries to fix

    A case study from Oh Dear showing how Mattias significantly improved service performance through SQL query optimization. The post focuses on identifying and fixing the right queries to achieve substantial performance gains.

  5. 5
    Article
    Avatar of bytebytegoByteByteGo·28w

    How Grab’s Migration from Go to Rust Cut Costs by 70%

    Grab successfully migrated their Counter Service from Go to Rust, achieving a 70% cost reduction through improved resource efficiency. While latency remained similar, the Rust version required only 4.5 CPU cores compared to Go's 20 cores for handling 1,000 requests per second. The migration involved careful library selection, rebuilding internal tools like configuration systems, and overcoming challenges with Rust's borrow checker and async concurrency model.

  6. 6
    Article
    Avatar of chromeChrome Developers·29w

    Chrome DevTools (MCP) for your AI agent

    Chrome DevTools now integrates with AI coding assistants through the Model Context Protocol (MCP), allowing AI agents to debug web pages directly in Chrome. This enables AI assistants to see what their generated code actually does in the browser, analyze performance traces, diagnose network errors, simulate user behavior, and debug styling issues in real-time. The public preview is available via npm and aims to improve AI coding accuracy by providing browser debugging capabilities.

  7. 7
    Article
    Avatar of awegoAwesome Go·32w

    Preserving Order in Concurrent Go Apps: Three Approaches Compared

    Explores three approaches to preserve order in concurrent Go applications while maintaining performance benefits. Compares ReplyTo channels, sync.Cond turn-taking, and permission passing chain patterns through detailed benchmarks. The permission passing approach emerges as the winner, offering good performance with zero allocations and clean abstractions for building reusable concurrent operations.

  8. 8
    Article
    Avatar of baeldungBaeldung·32w

    How to Reduce Spring Boot Memory Usage?

    Spring Boot applications typically consume significant memory (150+ MB) due to JVM architecture, embedded server threads, and framework overhead. Memory usage can be reduced through JVM tuning (serial garbage collector, reduced thread stack size, RAM limits), configuring web server thread pools, container-aware deployment practices, and removing unused dependencies. Key techniques include using -XX:+UseSerialGC, -Xss512k for thread stacks, setting MaxRAM limits, reducing Tomcat threads to 20, and matching JVM flags to container limits.

  9. 9
    Article
    Avatar of crunchydataCrunchy Data·31w

    Get Excited About Postgres 18

    PostgreSQL 18 introduces significant performance improvements including asynchronous I/O for faster reads, UUID v7 support for better indexing, B-tree skip scans for multi-column indexes, virtual generated columns as the default, and OAuth 2.0 authentication support. The async I/O feature allows workers to batch reads and optimize idle time, while UUID v7 provides better locality through timestamp-based prefixes. Skip scans enable using multi-column indexes even when leading columns aren't in query conditions, and virtual generated columns compute values on-the-fly without storage overhead.

  10. 10
    Article
    Avatar of collectionsCollections·29w

    Key Features and Enhancements in PostgreSQL 18

    PostgreSQL 18 introduces major performance improvements including asynchronous I/O that delivers up to 3x faster storage reads, NUMA support, and IO_uring on Linux. The release adds enterprise features like OAuth 2.0 authentication, temporal constraints, and virtual computed columns. Key enhancements include better extension loading for Kubernetes deployments, improved logical replication conflict handling, UUIDv7 support, and advanced vacuum operations for large tables. The update also brings query planning improvements with skip scan lookups and enhanced Unicode collation support.

  11. 11
    Article
    Avatar of opensoulsOpenSouls·29w

    UUIDv47

    UUIDv47 is a library that enables reversible transformation between UUIDv7 (time-ordered) and UUIDv4 (random-appearing) formats. It allows developers to hide time-ordering from external APIs while maintaining internal chronological benefits, using cryptographically secure SipHash-2-4 algorithm with high performance and zero dependencies.

  12. 12
    Article
    Avatar of codeconfessionsConfessions of a Code Addict·29w

    Compiling Python to Run Anywhere

    Muna's founders built a Python compiler that generates optimized C++ code from unmodified Python functions, enabling cross-platform deployment without interpreters. The system uses symbolic tracing to create intermediate representations, type propagation to bridge Python's dynamic typing with C++'s static typing, and AI-powered code generation to implement thousands of library functions. Performance optimization happens through exhaustive testing of multiple implementation variants across different hardware, with telemetry data driving automatic selection of the fastest approaches.

  13. 13
    Article
    Avatar of webdevWebDev·30w

    UUIDv47

    UUIDv47 is a library that enables reversible transformation between UUIDv7 (time-ordered) and UUIDv4 (random-appearing) formats. It allows developers to hide time-ordering from external APIs while maintaining internal chronological benefits, using cryptographically secure SipHash-2-4 algorithm with high performance and zero dependencies.

  14. 14
    Article
    Avatar of endlerMatthias Endler·29w

    On Choosing Rust

    Addresses common criticisms and misconceptions about adopting Rust in established software projects like git and coreutils. Debunks claims about Rust's performance limitations, security benefits, and motivations behind language adoption. Argues that Rust adoption is driven by practical concerns like memory safety and maintainability rather than hype, citing real-world examples and production use cases.

  15. 15
    Article
    Avatar of nestjsdevsNestjs Developers·29w

    UUIDv47

    UUIDv47 is a library that enables reversible transformation between UUIDv7 (time-ordered) and UUIDv4 (random-appearing) formats. It allows developers to hide time-ordering from external APIs while maintaining internal chronological benefits, using cryptographically secure SipHash-2-4 algorithm with high performance and zero dependencies.

  16. 16
    Article
    Avatar of tigrisTigris·32w

    How LogSeam Searches 500 Million Logs per second

    LogSeam demonstrates how to build a high-performance security data lakehouse that searches 500 million logs per second with 1.5 TB/s throughput while reducing costs by 40-80%. The architecture converts JSON logs to compressed Parquet files, achieving 100x storage reduction, and uses globally distributed object storage with elastic compute for sub-second query responses across continents.

  17. 17
    Video
    Avatar of mentaloutlawMental Outlaw·31w

    CachyOS - The Blazingly Fast Linux Distro

    CachyOS is a Linux distribution based on Arch Linux that focuses on performance optimization through compiler flags and kernel configurations, similar to Intel's discontinued Clear Linux. Benchmarks show modest performance improvements over standard distributions, with better results in some applications but marginal gains in others. The distro includes gaming-focused features and handheld device support, making it appealing for gaming and performance-oriented users who want Arch benefits without manual optimization work.

  18. 18
    Article
    Avatar of nuxt_sourceNuxt·32w

    Nuxt 4.1 · Nuxt Blog

    Nuxt 4.1 introduces enhanced build stability through import maps that prevent cascading hash changes, experimental Rolldown support for faster Rust-powered bundling, and improved lazy hydration macros. The release includes new module development features like dependency management, lifecycle hooks, and enhanced file resolution utilities. Performance optimizations include better route rules cache management and TypeScript processing improvements.

  19. 19
    Article
    Avatar of cloudflareCloudflare·29w

    Cloudflare just got faster and more secure, powered by Rust

    Cloudflare has replaced their core NGINX-based proxy system (FL1) with a new Rust-based modular proxy (FL2) built on their Oxy framework. The migration delivers a 25% performance boost with 10ms faster median response times, uses less than half the CPU and memory, and provides better security through Rust's memory safety guarantees. The new modular architecture allows faster product development, graceful restarts without dropping connections, and eliminates the incremental performance cost of adding new features that plagued the legacy system.

  20. 20
    Article
    Avatar of freecodecampfreeCodeCamp·29w

    How to Fix Memory Leaks in React Applications

    Memory leaks in React applications occur when components create resources like event listeners, timers, or subscriptions but fail to clean them up during unmounting. This causes applications to consume increasing amounts of RAM, leading to slowdowns and crashes. The guide demonstrates how to fix common memory leak sources by properly cleaning up event listeners with removeEventListener, clearing timers with clearInterval/clearTimeout, unsubscribing from external data sources, and canceling API requests using AbortController in useEffect cleanup functions.

  21. 21
    Article
    Avatar of logrocketLogRocket·30w

    React Server Components broke my app and I still don’t know why

    React Server Components can introduce unexpected issues including caching bugs and loading problems that break applications, despite their promise of performance improvements. The technology presents hidden pitfalls that developers need to be aware of when implementing server-side rendering solutions.

  22. 22
    Article
    Avatar of telerikTelerik·31w

    React Basics: How to Use React useCallback Effectively

    The useCallback hook prevents unnecessary component rerenders by memoizing function definitions between renders. When callback functions are passed as props to memoized child components, React recreates them on every render, causing performance issues. useCallback maintains stable function references unless dependencies change, optimizing performance in scenarios like passing callbacks to React.memo components, custom hooks, and useEffect dependencies. Best practices include accurate dependency arrays, avoiding overuse, and writing components that minimize the need for memoization.

  23. 23
    Article
    Avatar of distilledCode Distilled·31w

    FrankenPHP vs PHP-FPM

    A performance comparison between FrankenPHP and PHP-FPM examining CPU usage, memory consumption, and operational costs. The analysis reveals that FrankenPHP with workers offers superior performance for high-traffic applications, providing lower latency and higher throughput. The comparison includes recommendations for when to choose each solution based on traffic patterns and infrastructure requirements.

  24. 24
    Article
    Avatar of phoronixPhoronix·30w

    Ubuntu 25.10's Rust Coreutils Transition Has Uncovered Performance Shortcomings

    Ubuntu 25.10's transition from GNU Coreutils to Rust Coreutils has revealed several performance issues, with the Rust version being significantly slower in some cases. Notable problems include cksum being up to 17x slower for large files and sort command failing to complete for large single-line files. While some issues like base64 performance have been quickly resolved and even improved beyond GNU's performance, developers are actively working to address remaining performance gaps before the stable release.

  25. 25
    Article
    Avatar of golangGo·31w

    A new experimental Go API for JSON

    Go 1.25 introduces experimental encoding/json/v2 and encoding/json/jsontext packages to address long-standing issues in the original encoding/json package. The new v2 API fixes behavioral flaws like invalid UTF-8 handling and duplicate JSON keys, improves performance especially for unmarshaling (up to 10x faster), and provides better streaming capabilities. The jsontext package handles JSON syntax while v2 manages semantics, offering new interfaces like MarshalerTo and UnmarshalerFrom for better performance. The original v1 package will be reimplemented using v2 under the hood, allowing gradual migration while maintaining backward compatibility.

See all Performance archives