Topics /
The most popular blog posts about systems & performance on Hacker News. 391 posts.
Our web app allows users can change the length of a song or find loops present in it for infinite listening, remixing, or for their next next video edit or performance. After uploading a song, there is an initial server-side analysis step after which...
New year, new RISC-V Yocto blog post \o/ When I wrote my last post, I did really not expect my brand new VisionFive-2 board to find its way to me so soon… But well, a week ago it was suddenly…
The groundbreaking Intel 386 processor (1985) was the first 32-bit processor in the x86 line. It has numerous internal registers: general-pu...
This post will mainly focus on ROCm HIP wasted opportunities for adoption. I wrote an overview of ROCm earlier at https://threedots.ovh/blog/2021/11/what-is-amd-rocm/ . This post will specifically eschew talking about issues about the API renames in ROCm HIP – breaking source compatibility with CUDA quite needlessly.
It’s been a while since my last blog post, but I believe it’s better to post late than never, so here I am!
If your household uses Xfinity for cable, you have almost definitely heard of Xfinity Stream at some point. If you haven’t, Xfinity Stream is a service that allows you to access Xfinity’s on-demand TV over the internet. This can be useful if you want to watch something on Xfinity’s on-demand TV but do not want to go…
The futex ( f ast u serspace mu tex ) is a Linux kernel feature designed for synchronization primitives (mutexes, condition variables, semaphores, etc.). Like many topics in concurrency, they have a reputation for being tricky (for example, see the paper Futexes Are Tricky ). Despite that, they really are a…
We have had Jails on FreeBSD since 4.0 came out 19 years ago in March 2000. This describes how to setup jails on FreeBSD 12 without any helpers.
We investigate how the data cache influences the performance of your program, talk about ways for you to write faster programs by better leveraging the data cache.
The Bhyve FreeBSD hypervisor (called/spelled ‘beehive’ usually) was created almost 10 years ago. Right now it offers speed and features that other similar solutions provide – such…
The Pico Cray is a cluster of 8 Picos and a controller Pico. All Picos are connected in parallel by an I2C bus. The controller is designated by one GPIO being held low, the processor Pico’s addresses are allocated automatically. The Controller Pico has a 320x240px display connected by SPI and a couple of buttons. ……
Twitter user opdroid1234 remarked that they are getting more performance out of the ARM nodes than out of the Intel nodes on Amazon's cloud (AWS). I found previously that the Graviton 3 processors had less bandwidth than comparable Intel systems. However, I have not done much in terms of raw compute power. The Intel…
We write about Windows Internals, Hypervisors, Linux, and Networks.
There are many legitimate arguments to be had about performance, but there are also plain old excuses. It's time to end the excuses.
Don’t want a big old bloated multi-kilobyte executable with a data segment, but you do need to store some data?
This post will show you how to use D to write a bare-metal “Hello world” program that targets the RISC-V QEMU simulator. In a future blog post ( now available ) we’ll build on this to target actual hardware: the VisionFive 2 SBC. See blog-code for the final code from this post. For a more complex example, see…
This article contains a basic Nvidia CMP 170HX performance overview, a performance lockdown workaround, a hardware teardown, a watercooling installation guide, and a repair log.
Suppose we’ve got a service. We’ll gloss over the details for now, but let’s stipulate that it accepts requests from the outside world, and takes some action in response. Maybe those requests are HTTP requests, or RPCs, or just incoming packets to be routed at the network layer. We can get more specific later. What…
Born secretly on 27 March 2017, it was Apple’s biggest gamble. Introduced in High Sierra, it didn’t work on Fusion Drives and had other problems. Also explains version numbering.
This post explores one of my favorite hacks in SerenityOS. I don’t recommend doing this in your codebase, but it has worked for us so far. :^)
Between the Material You design and move to break apart the Chrome browser from ChromeOS, now more than ever, ChromeOS is Linux with Google's desktop environment.
Tutorial on how to make a stack trace or backtrace manually unwinding the stack using ELF’s .eh_frame and call frame information (CFI)
As regular readers of this blog will now, JDK Flight Recorder (JFR) is one of my favorite tools of the Java platform. This low-overhead event recording engine built into the JVM is invaluable for …
[Updated on 2023-01-24: add a small section on Distillation.] Large transformer models are mainstream nowadays, creating SoTA results for a variety of tasks. They are powerful but very expensive to train and use. The extremely high inference cost, in both time and memory, is a big bottleneck for adopting a powerful…
This is a brief mini-howto on how to build a linux based wireless router out of what amounts to spare parts, using the linux router project's single disk router software. The total cost of each end of the router, including antennas and cabling, is less than 1300 US dollars, which is a considerable cost savings over…
Fiber, a less known concept compared with coroutine, is a pretty powerful addition to cooperative multitasking. As a graphics programmer in game industry, I totally appreciate the great flexibility that fiber brings on the table. As a matter of fact, I feel the tech is a little bit underappreciated due to the lack of…
Service mesh is a blazing hot topic in software engineering right now and rightfully so. I think it’s extremely promising technology and I’d love to see it widely adopted (in cases where it truly makes sense). Yet it remains shrouded in mystery for many people and even those who are familiar with it have trouble…
