Best of Computing — 2024

  1. 1
    Video
    Avatar of fireshipFireship·2y

    Intel chips can’t possibly be this bad… 100% crash rate?

    Intel's 13th and 14th generation CPUs, also known as Raptor Lake, are experiencing high failure rates, causing frequent crashes in gaming and other high-demand applications. Initial blame was placed on Nvidia drivers, but later investigations pointed to issues within Intel's microcode algorithm. Many users and developers are calling for a recall of these CPUs due to their instability. Intel has acknowledged the problem but has yet to offer a definitive solution.

  2. 2
    Video
    Avatar of bigboxswebigboxSWE·2y

    What Your Linux Distro Says About You

    This post humorously describes various Linux distributions and the types of individuals who use them.

  3. 3
    Article
    Avatar of watercoolerWatercooler·2y

    Linux is free

    Linux is a free operating system that offers numerous benefits to users.

  4. 4
    Video
    Avatar of communityCommunity Picks·2y

    Top 50+ AWS Services Explained in 10 Minutes

    Learn about over 50 different AWS products and services, including AWS services for building robots, options for deploying WordPress sites, and the different database options available.

  5. 5
    Article
    Avatar of hnHacker News·2y

    defrag the game

    Defrag the game offers varying difficulty levels based on the size of the drive chosen for defragmentation. Players can select from Easy (1KB), Normal (128KB), or Hard (1MB) to test their skills.

  6. 6
    Article
    Avatar of hnHacker News·2y

    PyTorch is dead. Long live JAX.

    The post critiques PyTorch's effectiveness in industrial-scale scientific computing, arguing it wasn't designed for large-scale, distributed systems. In contrast, JAX, developed by DeepMind, offers a compiler-centered approach with better scalability and performance, making it more suitable for large-scale AI research. JAX's commitment to functional programming and reproducibility further enhances its utility, while PyTorch's attempts to integrate multiple backends lead to fragmentation and inefficiency. The post urges the adoption of JAX for improved research productivity and reliability.

  7. 7
    Article
    Avatar of communityCommunity Picks·2y

    How to Run an LLM Locally with Pieces

    The post provides information on running Local Large Language Models (LLLMs) locally within Pieces for Developers. It discusses the demand for secure and efficient machine learning solutions, hardware requirements for running LLMs, the difference between GPU and CPU, the best GPUs for local LLMs, troubleshooting common issues, and future-proofing the setup.

  8. 8
    Article
    Avatar of substackSubstack·2y

    How to Build an AI Data Center

    The rise of AI will accelerate the trend of building larger and more power-intensive data centers. Finding enough power for these data centers will become increasingly challenging.

  9. 9
    Article
    Avatar of hnHacker News·2y

    I don't know how CPUs work so I simulated one in code

    To understand basic computer operations, a developer simulated an 8-bit CPU in Go, inspired by J. Clark Scott's book. The project involved creating a simple computer that handles keyboard input and renders text using custom fonts. The developer wrote a crude assembler and implemented various hardware components via Go channels. The enriching experience emphasized core computing concepts like bit manipulation, ALUs, registers, and basic assembly language programming.

  10. 10
    Article
    Avatar of hnHacker News·2y

    Working Turing Machine

    The Turing machine is either an abstract model of an algorithmic machine or an esoteric programming language named after Alan Turing. It comprises an infinitely long tape, a head that reads and writes symbols, registers for machine state, and a table for state-symbol instructions. The machine operates by reading symbols, updating states and symbols based on instructions, and moving the tape. The model described can handle 32 symbol-state combinations and requires no electric motor, though it uses ~2900 Lego parts. Despite its size, it remains functional and educational, allowing users to play, program, and understand its mechanisms.

  11. 11
    Article
    Avatar of devtoDEV·2y

    Comparing SQL engines by CPU instructions for DML

    This post compares the CPU usage of different SQL engines, such as PostgreSQL, Oracle, SQL Server, MySQL, TiDB, YugabyteDB, and CockroachDB, for simple DML operations. The results show that despite their different implementations, PostgreSQL, Oracle, and SQL Server perform well in terms of CPU usage. TiDB, YugabyteDB, and CockroachDB exhibit higher CPU utilization due to their distributed and scalable architectures. Overall, the performance varies depending on the database engine and workload.

  12. 12
    Article
    Avatar of hnHacker News·2y

    Implementing a tiny CPU rasterizer

    Learn how to implement a basic CPU rasterization engine in C++ through a 12-part tutorial series. The project, which is still a work-in-progress, covers everything from drawing the first pixels to advanced techniques and optimizations. All code is available on GitHub, with a single commit for each article.

  13. 13
    Article
    Avatar of hnHacker News·2y

    Pikuma: Exploring How Cache Memory Really Works

    Understanding CPU cache is essential for optimizing programming performance. Cache memory, located within the CPU, is significantly faster than RAM and helps bridge the speed gap between the CPU and main memory. There are different levels of cache (L1, L2, L3), each with varying speeds and sizes. Cache efficiency can be enhanced by techniques such as maintaining data locality, avoiding complex loops, aligning structures, and using appropriate compiler flags. Cache hits and misses play a crucial role in performance, and different cache placement policies can affect how data is retrieved and stored.

  14. 14
    Article
    Avatar of communityCommunity Picks·2y

    Nvidia: An Overnight Success Story 30 Years in the Making

    Nvidia, a technology company founded in 1993, experienced crucible moments throughout its history. The company originally focused on creating 3D graphics cards for gamers but eventually shifted towards AI computing. This pivot proved to be a game-changer for Nvidia and positioned the company as a global leader in AI computing. One of the most significant crucible moments was the decision to abandon their initial chip architecture and start from scratch to align with industry trends. Nvidia's commitment to AI computing has propelled the company's growth and positioned them at the forefront of technological advancements.

  15. 15
    Video
    Avatar of primeagenThePrimeTime·2y

    Zen 5 And AI Doom w/ Casey Muratori

    The discussion between the author and Casey Muratori delves deep into the intricacies of CPUs, specifically focusing on the Zen 5 architecture. They explore how L1 cache functions and its significance, the impact of AVX 512 instruction sets, and how these elements contribute to overall CPU performance. They also touch upon the future implications of AI in the gaming industry and how advancements like the Zen 5 could shape it. The conversation highlights the challenges programmers face in fully utilizing newer CPU features.

  16. 16
    Article
    Avatar of singularityhubSingularity Hub·2y

    What Is a GPU? The Chips Powering the AI Boom, and Why They’re Worth Trillions

    Discover the power and importance of GPUs in the AI boom, their differences from CPUs, and the potential of specialized chips in the future.

See all Computing archives