Posts

At the start of the week, a friend of mine sent me a link. "What an interesting yet pointless question!", both of us agreed. Later that day, another link was also sent to me. Apparently there was someone who did care a little bit. Out of curiosity, I opened the find_md5.py file, and it hit me: "How can I optimize this task?" The task did shout SIMD to me, and I knew rust has core::arch which provides many SIMD intrinsics. Hence my journey to try out Rust's SIMD and inline assembly began.

Recently I made a presentation for a talk on FWCG2021. As always I went to Beamer. What was different this time was that I needed many figures, I mean, a lot. On top of that, I wanted to do overlays on these figures. Overleaf timed out when trying to compile all the TikZ figures so I turned to external. There are a lot of files to generate when I built the presentation locally and it was unbearably slow. Maybe there's a way to reduce the time, at least a little bit?