Myfetch
Myfetch
I spent the majority of my day on this project – and it was well worth it!
❯ hyperfine 'target/release/myfetch' 'fastfetch -c ~/.config/fastfetch/config\ \(Copy\).jsonc --logo none' 'neofetch --fast --off'
Benchmark 1: target/release/myfetch
Time (mean ± σ): 609.1 µs ± 103.2 µs [User: 491.5 µs, System: 478.5 µs]
Range (min … max): 353.1 µs … 1058.0 µs 1981 runs
Warning: Command took less than 5 ms to complete. Note that the results might be inaccurate because hyperfine can not calibrate the shell startup time much more precise than this limit. You can try to use the `-N`/`--shell=none` option to disable the shell completely.
Benchmark 2: fastfetch -c ~/.config/fastfetch/config\ \(Copy\).jsonc --logo none
Time (mean ± σ): 9.4 ms ± 0.3 ms [User: 1.2 ms, System: 8.0 ms]
Range (min … max): 8.7 ms … 10.7 ms 270 runs
Benchmark 3: neofetch --fast --off
Time (mean ± σ): 176.8 ms ± 1.5 ms [User: 105.2 ms, System: 79.2 ms]
Range (min … max): 174.0 ms … 180.0 ms 17 runs
Summary
target/release/myfetch ran
15.51 ± 2.68 times faster than fastfetch -c ~/.config/fastfetch/config\ \(Copy\).jsonc --logo none
290.25 ± 49.22 times faster than neofetch --fast --off
(Note: I won’t bring Neofetch back up, and this benchmark is technically rigged because they are both still outputting more info, but that is the Myfetch charm.)
This is with v1.0.0. Compare that to the first prototype:
Summary
fastfetch -c ~/.config/fastfetch/config\ \(Copy\).jsonc ran
13.48 ± 0.48 times faster than target/release/myfetch
Which is bad. The issue was that I was polling my entire system, not just what I needed, which was stupid.
Optimizations
The first time I removed some of the bloat was when I made my own function to fetch memory information, which brought Fastfetch only 9.89 ± 0.29 times faster from 9.92 ± 0.30, which might be a fluke but it was a win, especially long term. It made some of the later optimizations possible!
I beat Fastfetch with this:
Summary
target/release/myfetch ran
8.39 ± 1.04 times faster than fastfetch -c ~/.config/fastfetch/config\ \(Copy\).jsonc
I was running let mut sys = System::new_all(); instead of let mut sys = System::new();, and then I was refreshing on top of that! Oops!
After that, thw program plateaued at ~1.1ms, though I brought it down to 375.4 µs to 1059.1 µs when I removed let mut sys = System::new();, so no more polling!
Right now, the only call to sysinfo is here:
{
// KERNEL
let kernel = System::kernel_version().unwrap_or("Unknown".into());
print_formatted("Kernel", &kernel);
}
Making one call is not that bad, and getting rid of it would just mean more work on my side.
The output:
easyonhard@arch
Kernel: 6.12.31-1-lts
OS: Arch Linux (x86_64)
Shell: /usr/bin/zsh
Uptime: 11 hours, 56 minutes
Memory: 5.21 GiB / 31.19 GiB (16.7%)
It has a bit of formatting, but markdown doesn’t pick it up in code blocks.
Disclaimer: I don’t review these. This is my brain regurgitating onto my keyboard out off my nose.