Honestly not really, but what I work on is bottlenecked by api calls and not actually DDOSing our apis.
The other day I wanted to create a 10GB file with a perl one liner and it was taking over 10 minutes and doing some napkin math I figured out that should only take 4s at 1 character per clock cycle, so something is wrong (it was flushing every single character). I tried versions that flushed less and could get it to 7s by printing chunks of 1MB. That's ultimately the goal of the class, get a rough idea of what should be possible.
What I really like about that class though is it really makes you realize even assembly is a high level language now. I had heard about it, but it didn't really make sense to me, but the course basically has you writing doing assembly that looks like it should be doing the same thing (e.g. same thing but with loop unrolling), and explains what's happening in the background making one version twice as fast as the other.
maximus-decimus|1 year ago
The other day I wanted to create a 10GB file with a perl one liner and it was taking over 10 minutes and doing some napkin math I figured out that should only take 4s at 1 character per clock cycle, so something is wrong (it was flushing every single character). I tried versions that flushed less and could get it to 7s by printing chunks of 1MB. That's ultimately the goal of the class, get a rough idea of what should be possible.
What I really like about that class though is it really makes you realize even assembly is a high level language now. I had heard about it, but it didn't really make sense to me, but the course basically has you writing doing assembly that looks like it should be doing the same thing (e.g. same thing but with loop unrolling), and explains what's happening in the background making one version twice as fast as the other.