Subscribe
Sign in
Home
Podcast
Notes
Chat
Alternate Payment Processor
Archive
About
Latest
Top
Discussions
I Spent 6 Months Fighting My Compiler. Here's What I Learned About Loop Unrolling (Part 1)
The Day the Pragmas Stopped Working
13 hrs ago
•
Henrique Bucher
1
1
“Stop Teaching C”: When Academic Abstraction Meets Reality
Kate Gregory’s 2015 CppCon talk argues for skipping C fundamentals when teaching C++.
Dec 24
•
Henrique Bucher
10
6
From iPhone to Data Center: How One CppCon Talk Changed How We Think About Performance
Let me walk you through why this 2014 CppCon talk remains one of the most watched C++ presentations ever, and more importantly, why the lessons Carruth…
Dec 22
•
Henrique Bucher
2
Dissecting Optiver’s HFT Optimization Playbook
Carl Cook from Optiver delivered what became one of CppCon’s most-watched talks on low-latency trading systems. Seven years later, some of his…
Dec 19
•
Henrique Bucher
17
QuickFix vs hffix vs SIMD FIX shootout.
After building what I thought was the ultimate SIMD FIX parser in just a couple of days, I decided to benchmark it against hffix - a header-only FIX…
Dec 16
•
Henrique Bucher
9
6
Your Shipped Code Runs Like Garbage Despite Their NASA Computer
Game developers ship at 60fps.
Dec 15
•
Henrique Bucher
5
Why C Still Beats C++ Where It Matters: A High-Frequency Trading Perspective
Part 4: The Systems That Actually Run Financial Markets (Spoiler: They’re C)
Dec 10
•
Henrique Bucher
6
2
Why C Still Beats C++ Where It Matters: A High-Frequency Trading Perspective
Part 3: Bare Metal Superiority and Why Your FPGA Team Uses C
Dec 8
•
Henrique Bucher
6
Why C Still Beats C++ Where It Matters: A High-Frequency Trading Perspective
Part 2: Simplicity, Debuggability, and Why Your Template Errors Are Killing Productivity
Dec 5
•
Henrique Bucher
4
Why C Still Beats C++ Where It Matters: A High-Frequency Trading Perspective
Part 1: Performance, Predictability, and the Zero-Abstraction Reality
Dec 3
•
Henrique Bucher
6
2
1
November 2025
Your Triple-Loop Matrix Multiply Is 50x Slower Than It Needs To Be (And SIMD Isn’t Why)
Ask any developer why BLAS is fast and they’ll say “SIMD” or “vectorization.” They’re wrong.
Nov 29
•
Henrique Bucher
3
1
48 Hours, 48 Videos, 1 Month: What November 2025's C++ Talks Taught Me About Writing Faster Trading Systems
November 2025 was packed with C++ content.
Nov 28
•
Henrique Bucher
9
1
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
