I take my shitposts very seriously.

  • 0 Posts
  • 181 Comments
Joined 2 years ago
cake
Cake day: June 24th, 2023

help-circle
  • LGA sockets (including AM5) have flat contacts on the CPU and spring-loaded contacts on the motherboard. No pins, no holes. You could take a small tweezer or precision flathead screwdriver and lift the CPU from a corner. As long as you don’t reach in too deep, it won’t damage either side. You should be able to lift it with minimal force.


  • Find the motherboard’s manual. It will have instructions for installing and removing both the CPU and the cooler. If the CPU uses a PGA socket, removing it might require a little force.

    If you’re sure the clip and the retaining frame are released, but can’t/don’t want to lift it by its edges, you could use suction.

    • If there’s still some moisture in the thermal compound, you can use a flat piece of glass or plastic. Press it against the CPU’s surface and lift gently. If the thermal compound is dry, apply a little from a new tube.
    • A small suction cup might work.
    • There are vacuum pens made specifically for this kind of work.

    (edit) The MSI B650 uses an AM5 socket, which is an LGA package. The CPU itself doesn’t have any pins that could be damaged, so you can be a bit more forceful. You could even take a small tweezer and pry it out from one of the corners (as long as the retaining frame is off, of course).









  • I simply use Nextcloud to sync the vault directory. It has clients for both desktop and mobile and works perfectly fine. I use it to sync basically everything between my work, home, laptop, and mobile.

    The only drawback is that I don’t know if Obsidian automatically reloads a file if it is changed - if not, and you leave the file open in the editor, you might accidentally overwrite the new file with old data.








  • You’ll encounter math eventually. It could be as simple as implementing linear interpolation for a custom type, or understanding why a type is not suited for a particular application (e.g. never use floating points to represent money). If you delve into low-level networking, you’ll need a good understanding of binary/decimal/hexadecimal conversions and operations. If you go into game development or graphics, you won’t survive without a deep understanding of vectors, matrices, and quaternions. Any kind of data science is just math translated to a machine-readable language.

    In my opinion, knowledge of the basic concepts is more important than being good at actually performing mathematics with pen and paper. For example, if you need to apply a transformation to a vector, nobody expects you to whip up a program that does the thing. Instead, you should immediately know:

    • what a transformation is (translation, rotation, scaling, projection, etc),
    • that each transformation has a corresponding transformation matrix,
    • that you’ll have to deal with inhomogeneous and homogeneous coordinates, and
    • that you’ll have to combine the transformation matrices and the original vector.

    That abstract knowledge will give you a starting point. Then you can look up the particulars – the corresponding transformation matrices, the method to convert between inhomogeneous and homogeneous coordinates, and the process of matrix multiplication. I know because I failed calculus.