Explore whether adding the K8S CPU limit will degrade service performance.

Even though we often hear terms like L1, L2, cache block size, etc., most programmers have a limited understanding of what cache really is. This is a beginner-friendly primer on how cache works.

This blog introduces three mechanisms to investigate the execution time of a Maven build. Having a reliable way to measure build execution time can help identify bottlenecks. This in turn helps making effective improvements, thereby contributing to higher developer productivity. Find out how to effe...

Explore how our AI-powered tool is revolutionizing flamegraph interpretation

Find out how to use perf, a built-in Linux profiler, to analyze the performance of Java applications.

Updating Kubernetes pod resources in-place is a new way to optimize your k8s cluster without disruptions or restarts!

By Benoit Rostykus, Gabriel Hartmann

I wrote 84 new matmul kernels to improve llamafile CPU performance.

Back in 2019, Nick Fitzgerald published always bump downwards, an article making the case that for bump allocators, bumping “down” (towards lower addresses) is better than bumping up. The biggest reasons for this are bumping up requires 3 branches vs 2 for bumping down and rounding down requires few...

New features and improvements to execution traces from the last year.

We’re building an open source alternative to AWS. For IPv4 assignment and firewall rules, we use Linux’s Netfilter / Nftables. This subsystem provides a powerful way to handle packets addressed to the host. We recently came across flowtables - a network acceleration feature in the Linux kernel that ...

eBPF Documentary

Are database benchmarks still relevant ? Let's understand why it's a poor way to choose a database.

Grafana Beyla 1.2 offers improved Kubernetes support, including the ability to decorate metrics and traces with the metadata of Kubernetes pods and nodes.

This month’s PGSQL Phriday #015 topic is about UUIDs, hosted by Lætitia Avrot. Lætitia has called for a debate. No, no, no. I say let’s have an all-out war. A benchmark war. I have deci…

Benchmarks for JDK HTTP Server running on Java 21 with Virtual Threads - GitHub - ebarlas/java-httpserver-vthreads: Benchmarks for JDK HTTP Server running on Java 21 with Virtual Threads

Calculate the optimal size for your bloom filter, see how many items a given filter can hold, or just admire the curvy graphs. Also borrow my MIT licensed Javascript for your own programs.

A collection of bad practices in HTML, copied from real websites.

Background Async-profiler has two options for CPU profiling: -e cpu and -e itimer. -e cpu (default) relies on perf_events, while -e timer is based on setitimer. Perf events typically yield more acc...

Grafana k6 v0.47.0 is here, introducing binary metadata support in the k6 gRPC module, Docker images for ARM64 architecture, and other exciting new features.

Learn how the PostgreSQL JDBC Driver provides Statement Caching and what settings you need to configure in order to optimize performance.

In this talk, we’ll discuss a severe JVM performance issue, the methodology leading to its discovery, how it affects the whole Java ecosystem, and what you could do to avoid it in your code. This scalability bottleneck has existed in OpenJDK for decades; the most surprising aspect is how such an im...

Having a fast and responsive app is orthogonal to “knowing your big Os”. Unfortunately, most tech companies over-emphasize algorithms in interviews and downplay systems knowledge, and I believe that’s one reason behind sluggish apps and bloated systems. I’ve seen this play out repeatedly. Interviewe...

A Java version of simdjson. Contribute to simdjson/simdjson-java development by creating an account on GitHub.

A collection of notes on topics of interest

At Flipkart, over a billion recommendations are served to users every day. These recommendations help millions of users discover products…

Compare the pros and cons of various caching strategies to choose the best one for your use case.

As a computer engineer who has spent half a decade working with caches at Intel and Sun, I’ve learnt a thing or two about cache-coherency. This was one of the hardest concepts to learn back in coll…

Developers often believe that software performance follows a Pareto distribution: 80% of the running time is spent in 20% of the code. Using this model, you can write most of your code without any care for performance and focus on the narrow pieces of code that are performance sensitive. Engineers l