@epickrram

Low-latency systems, high-performance Java, Linux tuning.

epickrram.com
Joined May 2015
  • Tweets463
  • Following24
  • Followers862
31Media
Getting down to 2.5 microseconds scheduling jitter: epickrram.blogspot.co.uk/201…
1
30
1
38
unfortunately, as you probably already know, floating point
1
1
9
3,160
Pretty dense read, but some gems in here scylladb.com/2023/10/02/intr…

Introducing “Database Performance at Scale”: A Free, Open Source Book

Discover new ways to optimize database performance and avoid common mistakes that impact latency and throughput.

scylladb.com
2
260
johnnysswlab.com/faster-hash…

Faster hash maps, binary trees etc. through data layout modification - Johnny's Software Lab

We investigate how to make faster hash maps, trees, linked lists and vector of pointers by changing their data layout.

johnnysswlab.com
5
15
1,801
We've had our heads down building our core product suite @4otcmarkets; now we're expanding our remote-first dev team. Performance and reliability are success metrics. Tech stack C#/Java/Rust/Cloud. Interested? [email protected] linkedin.com/jobs/view/37266…
1
258
It's been *checks watch* a long time since I worked on any OSS. So I'm happy that @4otcmarkets has published DotMetrics, our monitoring sidecar for DotNet applications. Read more here: 4otc.com/tech/dot-metrics-op…
224
Most of the articles I've read on ChatGPT talk about the ability to generate code. Today I discovered a different use-case that feels like it will materially affect my work. 1/7
1
2
4
1,484
more replies
When I saw the first two words on the response, I thought I'd asked too much. Turns out I just hadn't been precise enough... 6/7
1
353
Looks like I won't need to surf around kernel source code so much anymore... 7/7
2
346
1) Always start with a test. Writing the test first shapes your design for the better. Starting with a test will set you up to code in an attempt to pass your test, rather than continuously going back to fix code that's tested too late, and failing. Start as you mean to go on.
2
2
22
Finally able to write complete docs in GitHub wiki pages without embedded images 🎉 github.blog/2022-02-14-inclu…
1
1
blog.janestreet.com/magic-tr…
3
2
Nice detailed write up of page-cache operation biriukov.dev/docs/page-cache…

Linux Page Cache for SRE

SRE deep dive into Linux Page Cache # Last updated: Oct 2025 Contents Prepare environment for experiments Essential Page Cache theory Page Cache and basic file operations Page Cache eviction and page...

biriukov.dev
6
13
Replying to @benrady
@benrady kindly let me ramble on one of my favourite subjects. Go check it out on the site, or wherever you listen to podcasts!
Two's Complement Podcast@twoscp19 Aug 2021
New Episode: CPUs Are Clever. We discuss the kinds of things they do behind the scenes: Branch prediction, hyper-threading, OoO, renaming... Matt gets overly excited, and Ben tries to keep him on track. Listen at twoscomplement.org/#podcast or wherever you get your podcasts

Two's Complement

A Programming Podcast by Ben Rady and Matt Godbolt

twoscomplement.org
2
8
The next issue of the Four Steps newsletter goes out next week. It'll be published to the website in time, but if you'd like to see it before then, subscribe here: foursteps.software/
1
Impressive numbers from haproxy on ARM: haproxy.com/blog/haproxy-for…

HAProxy Exceeds 2 Million RPS on a Single Arm Instance

First ever software load balancer exceeds 2 million RPS on a single Arm instance! We're near an era where you get the world’s fastest load balancer for free.

haproxy.com
1
2
Nice! Babl is designed to support many thousands of concurrent connections. There are options to scale out your business logic, so get in touch if you're finding that babl is a limiting factor in your design.
Koth.io@io_koth3 Apr 2021
Just wanted to show off player capacity a bit. Here's the mini-map for koth.io - Every single one of those grey dots is a connected player (at 520~). Many thanks to @epickrram and his amazing project babl-ws #screeenshotsaturday #gamedev
1
1
Load more