Topics /
The most popular blog posts about systems & performance on Hacker News. 6,943 posts.
An end-to-end engineering practice for the MiMo-V2.5 series inference system, covering KVCache management, tiered caching, SWA-aware prefix cache trees, scheduling, Prefill/Decode pipelines, and multimodal optimizations.
The start of September brings a close to the Go 1.4 merge window . From now until the release in December, only bug fixes and tweaks will be accepted.
In the 1980s, if you wanted your computer to do floating-point calculations faster, you could buy the Intel 8087 floating-point coprocessor ...
I tweeted this the other day, and the internet was not pleased: > HTTPS is slow. No - wait - is it HTTP that's slow?! https://t.co/T49GG7oCaK pic.twitter.com/cfnYOpXMWc [https://t.co/cfnYOpXMWc] — Troy Hunt (@troyhunt) July 8, 2016 [https://twitter.com/troyhunt/status/751317949349130240] In fact, a
Part of a recent assignment for one of my classes involved calculating the Fibonacci sequence both recursively and iteratively and measuring the speed of each method. (BONUS: For a fun diversion, here is a paper I wrote about using the Golden Ratio, which is closely related to the Fibonacci sequence, as a base for a…
As WebGPU specification is approaching a usable shape, we are working on prototype implementations in both Gecko and Servo. If you ever wondered about what i...
Attackers have compromised the software distribution system of BigNox, developer of a popular Android emulator for PCs and Macs.
and how we found out.
I’ve recently been tinkering around with using a Raspberry Pi as a desktop computer. The Pi 4 is fairly powerful; very sufficient for web browsing, audio, video, and terminal use for programming.
Excessive instruction cache misses are the kind of a performance problem that's going to appear only in larger codebases. In this article, I'm describing some ideas on how to deal with this issue.
Exactly 100% of everything on the Internet involves exchanging messages which represent items of interest to humans. These items can be classified into three baskets: One for “media” (images, sound, video), one for text (HTML, PDF, XML), and one for “objects” (chat messages, payments, love poems, order statuses). This…
One of the concerns that often comes up when someone creates a new HTTP header is how much “bloat” it will add on the network. This is especially relevant in requests, when a little bit of extra data can introduce a lot of latency when repeated on every request.
Posted by Ivan Fratric, Project ZeroIntroductionHistorically, DOM engines have been one of the la...
Owen's personal website and portfolio
Spiral journey over kubernetes, cloud and devops world.
Diving into the territory of software sandboxing is diving into mostly uncharted territory. The necessary pieces to implement good sandboxing in your software are scattered all-around and the pioneers haven’t yet gathered enough knowledge into an unified mappa mundi that can guide new sailors through some well…
Let's blink an LED on Pine64 Ox64 BL808 64-bit RISC-V Single-Board Computer... With Nim Programming Language and Apache NuttX Real-Time Operating System
Unit 42 brings together world-renowned threat researchers, incident responders and security consultants to create an intelligence-driven, response-ready organization that's passionate about helping you proactively manage cyber risk.
Following an LDG.E SASS instruction through the hardware units in an RTX 4090.
You're probably reading this page because you've attempted to access some part of my blog (Wandering Thoughts) or CSpace , the wiki thing it's part of. Unfortunately you're using a browser version that my anti-crawler precautions consider suspicious, most often because it's too old (most often this applies to versions…
This is a followup to my previous post, prefix sum on Vulkan. Last year, I got a fancy algorithm for this problem running well on one device. This time, I will dive into the question of how to make it work well across a wide range of devices. The short answer is, it can be made to work pretty portably on Vulkan and…
In the previous year I wrote a program, running on Intel's first microprocessor - 4004, that computes the first 255 digits of π. So let's continue our journey with the successor - the Intel 4040.
Measuring your Python program’s memory usage is not as straightforward as you might think. Learn two techniques, and the tradeoffs between them.
Owing to its historic roots as a derivative of the original Berkeley Systems Distribution (BSD), OpenBSD includes a great deal of old code. Many files bear copyright notices from the year 1980, and in some cases, even older. Although not explicitly stated as a project goal, keeping OpenBSD modern is an important part…
Distributed consensus revised Howard, PhD thesis Welcome back to a new term of The Morning Paper! To kick things off, I’m going to start by taking a look at Dr Howard’s PhD thesis, ‘Distributed con…
