121
Sieve is simpler than LRU
SIEVE is Simpler than LRU: an Efficient Turn-Key Eviction Algorithm for Web Caches
122
Building a highly-available web service without a database
In this viral post, we explain a new paradigm of building simple clean apps: from prototyping to scaling up.
123
FlashAttention-3: Fast and Accurate Attention with Asynchrony and Low-Precision
FlashAttention-3 achieves up to 75% GPU utilization on H100s, making AI models up to 2x faster and enabling efficient processing of longer text inputs. It allows for faster training and inference of LLMs, supports lower precision operations for improved efficiency.
124
Sourcehut network outage post-mortem
sourcehut is a network of useful open source tools for software project maintainers and collaborators, including git repos, bug tracking, continuous integration, and mailing lists.
125
Zero-latency SQLite storage in every Durable Object
Kenton Varda introduces the next iteration of Cloudflare's Durable Object platform, which recently upgraded from a key/value store to a full relational system based on SQLite. For useful background on …
126
Erasure Coding for Distributed Systems
An overview of erasure coding, its trade-offs, and applications in distributed storage systems.
127
The simple beauty of XOR floating point compression
I recently implemented a small program to visualize the inner workings of a scheme that compresses floating point timeseries by XORing subsequent values. The resulting visualizations are quite neat…
128
A BSD person tries Alpine Linux
By Ruben Schade in Eora/Sydney, Australia.
129
We switched to Java 21 virtual threads and got a deadlock in TPC-C for Postgres
We revise some fundamental background and highlight an important concern behind Java 21 virtual threads: deadlocks might be unpredictable
130
The V8 Sandbox
V8 features a lightweight, in-process sandbox to limit the impact of memory corruption bugs
131
Serialization Is the Secret
If a value mutates in the forest with no one to see it, does it really mutate?
132
tolower() with AVX-512
A couple of years ago I wrote about tolower() in bulk at speed using SWAR tricks . A couple of days ago I was interested by Olivier Giniaux’s article about unsafe read beyond of death , an optimization for handling small strings with SIMD instructions, for a fast hash function written in Rust.
133
Making EC2 boot time faster
It's possible to cut EC2 boot time from 40 seconds to 5 seconds by combining several optimizations like root volume streaming, instance warm pools, and instance resizing.
134
The Billion Row Challenge (1BRC) – Step-by-Step from 71s to 1.7s
A step-by-step walkthrough of the Billion Row Challenge, taking a Parallel Java Streams aggregation from 71 seconds down to 1.7 with SIMD and tuning.
135
136
How to get the most out of Postgres memory settings
Engineering deep-dives, product updates, and company news from Tembo — the platform for running coding agents across your stack.
137
Ethereum has blobs. Where do we go from here?
On March 13, the Dencun hard fork activated, enabling one of the long-awaited features of Ethereum: proto-danksharding (aka EIP-4844 , aka blobs). Initially, the fork reduced the transaction fees of rollups by a factor of over 100, as blobs were nearly free. In the last day, we finally saw blobs spike up in volume and…
138
Fast B-Trees
( This is part of a series on the design of zest . See the list of posts here .)
139
An illustrated proof of the CAP theorem (2018)
The CAP Theorem is a fundamental theorem in distributed systems that states any distributed system can have at most two of the following three properties.
140
Three Laws of Software Complexity
I posit that most software engineers (particularly those working on infrastructural systems) are destined to wallow in unnecessary complexity due to three fundamental laws.
141
How we found and fixed an eBPF Linux kernel vulnerability
This blog post takes a look at the years where eBPF was one of the kernel subsystems that grabbed the attention of a lot of security researchers. We will tell the story of how we discovered CVE-2023-2163, what our root-cause analysis process looked like, and what we did to ultimately fix the issue.
142
Using Erlang hot code updates
One of the Erlang ecosystem's spiciest nerd snipes are hot code updates. Because it can do it. In ways that almost no other runtime can. I use Elixir which builds on Erlang and has the same capabilities.
143
The six dumbest ideas in computer security (2005)
There's lots of innovation going on in security - we're inundated with a steady stream of new stuff and it all sounds like it works just great. Every couple of months I'm invited to a new computer security conference, or I'm asked to write a foreword for a new computer security book. And, thanks to the fact that it's…
144
Antenna Diodes in the Pentium Processor
I was studying the silicon die of the Pentium processor and noticed some puzzling structures where signal lines were connected to the silico...
145
Why does everyone run ancient Postgres versions?
Many companies are running old Postgres versions and missing out on many improvements. The reasons: Postgres works and upgrades suck.
146
Opening Windows in Linux with sockets, bare hands and 200 lines of C
In this post I want to create a single file C file program to open a windows inside Linux without using xlib or any similar libraries. The idea is to explore X11 protocol and see how it is used to interact with X server to create windows.
147
Sysadmin friendly high speed Ethernet switching
I’ve been on the lookout for a ethernet switch that I don’t hate, the problem with a lot of higher speed (10G and above) ethernet switches is that they are quite expensive new and if you buy the used then they rarely have many years left ( or none at all ) of software patches.
148
149
How much memory do you need in 2024 to run 1M concurrent tasks?
Please refer to How Much Memory Do You Need in 2024 to Run 1 Million Concurrent Tasks? - Take 2 to check the latest result.
150
A Day in the Life: The Global BGP Table
Update: This article was discussed on Hackernews Much has been written and a lot of analysis performed on the global BGP table over the years, a significant portion by the inimitable Geoff Huston.