performance
90 stories and discussions about performance, aggregated from every source we track.
Analysis of Anthropic's Claude Opus 5.5 (Adaptive Reasoning, Max Effort, Default Fallback) and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first…
Tailscale performance updates reduce memory use, increase throughput, and speed startup with multi-queue, writev, and netmap caching.
Inside our performance sprint: the benchmarks Claude built, the loop each Slack thread ran, and the guardrails that let us ship 3,000 changes safely.
Analysis of Xiaomi's MiMo-V2.6-Pro and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.
Open source inference engine for agents that optimizes itself for your exact hardware. Compiles and tunes its kernels on your device, so open models run up to 2x faster than llama.cpp. Works on App...
How we fully migrated github.com away from CSS-in-JS.
My last post on the Rust compiler’s performance was two months ago and a lot has happened since then.
Cloudflare's global network is immense but not limitless. As we look for small ways to trim our resource usage, we sometimes get lucky and we can cut significantly more. Here’s how we reduced one of our Pingora-based…
'egg' is an open-source, high performance e-graph and equality saturation toolkit.
After a user signaled a performance issue in a Matrix client, we have added new tracing timers in the Matrix Rust SDK to spot the problem. Once found, we have fixed an SQL query improving the throughput from 19k to…
This post is a copy of tweets by Zooko Wilcox-O'Hearn . SHA256 was designed by the NSA. BLAKE (the original) and BLAKE3 were designed by Jean-Philippe Aumasson
With Rust, you can do low-level number-crunching and bit-by-bit processing, while enjoying memory safety and concurrency features. With Helix, you can use your Rust code inside of a Rails project.
Do you remember when we first started using AGENTS.md files? You would have a project root file...
In the previous post, I mentioned that buildcache has some unique properties compared to ccache and sccache. One of them is its Lua plugin system, which lets you write custom wrappers for programs that aren’t compilers…
We are pleased to announce the release of pandas 3.0.0, a major release from the pandas 2.x series. This release includes various new features, bug fixes, and performance improvements, as well as p...
A technical review of the live score feature in Google search results, delivered with some surprisingly old-school tech from the 90s.
One of the trickier things in software is gradual degradation. Development that happens in the wrong direction slowly over time which never triggers any alarms or upset users. Then one day you suddenly take a closer…
NOT NULL constraints are a ubiquitous tool for upholding data integrity. However, their performance benefits are less obvious. In this post we'll examine a few of these optimizations that transform query plans…
A writeup of our journey to improve the performance of Zellij
Companies care about cheap computing. Well, the first thing they care about is usability: Is it easy for average programmers to develop so...
How we fully migrated github.com away from CSS-in-JS.
Optimizing code starts with measuring it, and a measurement is only useful if it is repeatable: a 2% improvement is invisible under 5% of noise. Yet on an …
Explore how to tweak TCP performance in FreeBSD. Learn about Initial Window, TCP Segment Offload, and Buffer Tuning in this detailed article.
High-performance canvas trading chart with built-in UI, 66 indicators, real-time streaming, and backtesting.
requestIdleCallback promises to run only when the browser has spare cycles — but on a page that's never truly idle, that promise has no upper bound. Here's why the callback can stall, why the timeout escape hatch has its own trap, and when to reach for it instead of setTimeout or scheduler.yield.
C++ vs Node.js HTTP benchmark with wrk: Crow 5.4k req/s, Node http 12k, uWebSockets 57k. Same route, same laptop, with code and why Crow lost to Node.
Analysis of SpaceXAI's Grok 4.7 (xhigh) and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.
What do you do when your NumPy code isn’t fast enough? We’ll discuss the options, from Numba to JAX to manual optimizations.
The Catchpoint blog offers insights into Internet Performance Monitoring, market news, and the business itself. Don't miss out - sign up for updates today!
Analysis of OpenAI's GPT-6.1 Sol (max) and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.
Analysis of Anthropic's Claude Sonnet 5.5 (Adaptive Reasoning, Max Effort, Default Fallback) and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first…
Agent-first Python packages for editing Word documents, PowerPoint presentations, and Excel workbooks.
Released on Wednesday was a new version of the Dragonfly database that continues to promote itself as 'the world's most efficient in-memory data store' and a modern replacement to the likes of Redis and…
Writing a TSDB from scratch in Go Performance optimizations
A full-stack high-performance Quantum Operating System and simulator engine written in Rust. - strahinjastojanovic826-code/quant
Strands harness is a fully assembled, customizable, state-of-the-art agent you run locally or deploy anywhere.
The GPT-6.1 Sol release offers 5 models, each with different intelligence, performance, and pricing characteristics. Below is a comparison of the key metrics across the 5 models. For intelligence, the top model of…
Notable performance improvements in JDK27
How an inefficient Row-Level Security policy drove Postgres CPU to 80%; and why EXPLAIN failed to reveal it.
Analysis of OpenAI's GPT-6 Sol (max) and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.
Analysis of OpenAI's GPT-6 Luna (max) and comparison to other AI models across key metrics including quality, price, performance (tokens per second & time to first token), context window & more.
Tailscale performance updates reduce memory use, increase throughput, and speed startup with multi-queue, writev, and netmap caching.