wait, people actually still do that outside of two file toy projects?

I remember when I was traveling to the US for the first time, and some form had "Write your name like it is on your credit card".

They didn't like it when I did 🙃

Eh, I checked with my calipers and it's 1 (and a bit?) mm difference between the left and right end of the calling convention table.

I expected it to be more given how obviously it felt askew when I unpacked it :-D

I twitched when reading this, congrats

When you forget to fact check your LLM's article...

The answers a gamedev will give for these is going to be wildly different from mine, but that doesn't mean that either of us cares about performance less.

And that's still just the simple stuff before one gets into actual optimization work. 🙃

But performance is not just spectrum, it has many dimensions.

Do you care about your p99 latency or p90 (or even lower)? What does the scaling between these look like? Are you expected to scale vertically or horizontally? Are you running on shared HW, or exclusive HW?

This mirrors some of my thoughts in that I have never met anyone who would say that they genuinely don't care about performance.

It's just that for some people that means avoiding pointless copies, and for other it means precisely placed explicit prefetches in their hot loops.

Honestly that's pretty relatable, I've had to stop and think about the actual year since 2020.

I remembered I didn't post them.

Actually, you should bring them to conferences next year.

No reason. 🙃

It was never sold, Phil got some test mugs made and I have one.

On one hand, I want one. On the other, I barely use my Catch2 mug.

decisions, decisions...

Where is this from? The suggestions have wildly different impact levels, but they are all presented next to the 400l number like they are supposed to be equal.