751
Topics /
Systems & Performance
The most popular blog posts about systems & performance on Hacker News. 2,560 posts.
752
Beyond 10,000 Lines: Lessons Learned from a Large Phoenix Project
Lessons Learned from a Large Phoenix Project
753
Keeping Netflix Reliable Using Prioritized Load Shedding
How viewers are able to watch their favorite show on Netflix while the infrastructure self-recovers from a system failure
754
Instrumentation: The First Four Things You Measure
Old observability metrics like uptime and MTTR aren't enough anymore. Teams must connect technical signals to business outcomes, like savings and speed.
755
The Merge
Learn about The Merge, when Ethereum Mainnet moved from proof-of-work to proof-of-stake in September 2022, cutting its energy use by about 99.95%.
756
757
Update on Metro
Earlier this week, I asked our engineering leads and release managers to take the Windows Metro version of Firefox off the trains. The team is solid and did good work, ...
758
GitHub's new Git storage architecture
GitHub's new Git storage architecture is built so thousands of agents can push to a single repository at once. Here's how it works.
759
Automatically closing FIN_WAIT_2 is a violation of the TCP specification
I was asked to debug another weird issue on our network. Apparently every now and then a connection going through CloudFlare would time out with 522 HTTP error.
760
Fast Llama 2 on CPUs with Sparse Fine-Tuning and DeepSparse
From live events to hands-on product demos to deep technical research, see what we're doing with AI at Red Hat.
761
GNU Guix 1.2.0
Blog posts about GNU Guix.
762
eBPF will help solve service mesh by getting rid of sidecars
eBPF Service Mesh - How we can build an eBPF-based service mesh in the kernel to replace the complex sidecar model
763
Technical Challenges Developing a Distributed SQL Database
Here is an outline of some of architectural issues addressed as we build YugabyteDB, our high-performance distributed SQL database.
764
Partitioning GitHub’s relational databases to handle scale
In 2019, to meet growth and availability challenges, we set a plan in motion to improve our tooling and ability to partition relational databases.
765
Look Ma, No Bubbles: Designing a Low-Latency Megakernel for Llama-1B
Benjamin Spector* , Jordan Juravsky* , Stuart Sul , Owen Dugan , Dylan Lim , Dan Fu , Simran Arora , Chris Ré
766
Why has CPU frequency ceased to grow? (2014)
A portfolio of HPC tools help data scientists, researchers, and developers efficiently build cross-architecture applications.
767
macOS 14.4 causes JVM crashes
Java users on macOS 14 running on Apple silicon systems should consider delaying the macOS 14.4 update
768
How fast is PDF.js?
Hi, my name is Thorben and I work at Opera Software in Oslo, not at Mozilla. So, how did I end up writing for Mozilla Hacks? Maybe you know that ...
769
Tests show high-temperature superconducting magnets are ready for fusion
A comprehensive study of high-temperature superconducting magnets built by MIT and Commonwealth Fusion Systems confirms they meet requirements for an economic, compact fusion power plant.
770
Things you must know before you can understand Meltdown as a developer
Meltdown is a cache-timing attack on Intel CPUs that allows all memory to be read by any process because of how they do Speculative Execution.
771
772
The State of Atom's Performance
We are archiving Atom and all projects under the Atom organization for an official sunset on December 15, 2022.
773
Making memcpy(NULL, NULL, 0) well-defined
C2y makes memcpy(NULL, NULL, 0) and other zero-length operations on null pointers well defined. This removes a source of bugs that may result in security issues.
774
775
AI models need a virtual machine
Neural networks are more useful when placed in a suitable, specialized environment.
776
Windows NTFS Tricks Collection
In this blog post René Freingruber (@ReneFreingruber) from the SEC Consult Vulnerability Lab shares different filesystem tricks which were collected over the last years from various blog posts or found by himself.
777
Why Reddit was down for 6 hours
Get the latest product news, company announcements, and user insights from the official Reddit blog.
778
779
What Flips Your Bit: Cosmic Ray Errors at Mozilla
(“This Week in Glean'' is a series of blog posts that the Glean Team at Mozilla is using to try to communicate better about our work. They could be release ...
780
Grafana Tempo, a scalable distributed tracing system
Tempo is designed to be a robust trace id lookup store whose only dependency is object storage (GCS/S3).