m

@borememore
14 Followers
18 Following
122 Posts

Update on my image chroma compression algorithm: There's now a C/C++ library that implements the algorithm, and it's preforming pretty decently.

Meet Bitfrost CC: https://codeberg.org/mbitsnbites/bitfrostcc

bitfrostcc

The Bitfrost CC library for image chroma channel compression.

Codeberg.org

Latest hack: A spatial domain variable block size luma dependent chroma compression algorithm

https://www.bitsnbites.eu/a-spatial-domain-variable-block-size-luma-dependent-chroma-compression-algorithm/

#image #compression

This weekend's hack: How long does it take to brute force a cryptographic key?

https://bruteforce.bitsnbites.eu/

Brute Force Attack Time Estimator

Another hurdle that I need to get over with Fedora is SELinux. Undecided if I should learn it (i.e. The Right Thing) or disable it and get on with my hacks.
@mattiasb Yes, Snap is a pain. I tend to go with Mint for desktop uses as it's essentially Ubuntu without Snap and Canonical bloat. However the Mint desktop environment (e.g. Cinnamon) isn't as slick as Fedora's gnome shell (if you're on decent hardware).
LFT-Nine https://www.youtube.com/watch?v=Ws4twUyt-MY if you know how the C64 works, this starts out bog standard and then gets increasingly WTF
NINE (Seemingly Impossible C64 Demo)

YouTube

As a @unity user I’d like to:
Read blog posts about the product or plans, like maybe some details about that “Unity next” or some such,

So that I can:
Stop seeing blog posts called “How AI enables the creation of high-quality ads” on the official Unity blog.

Thanks.

@mattiasb I'm still a noob and not really used to dnf etc, I struggled a bit with docker (e.g. it wasn't clear what compose solution to use) and NVIDIA, and I generally find the search-info-on-the-webs experience unsatisfactory (lots of useless guides from older versions, e.g. dnf in 41 doesn't work as in 40).

Other than that, I really like the UI/desktop (much better than Ubuntu), and default configs feel sane (e.g. zram), etc.

Guess who affected my decision to try it out 😉

Didn't have much luck getting docker/podman to work with CUDA in Fedora (too much hassle, mostly blaming NVIDIA), so I went with a user systemd unit to run my llama.cpp instance instead. Works perfectly.