My understanding is mamba, like conda, just call pip. So it likely wouldn't make a difference.
The pip section in a env file is just a list of arguments passed through to the pip install command. Prior to pip 20.3 we had to add `--use-feature=2020-resolver` to get an install that resolved for our teams that used mamba.
I got a bit confused by the statements in the thread. If the issue is downloading wheels to check dependencies, mamba should be a better alternative, as not only it is downloading packages in parallel but it uses a different dependency solver.
Conda installs conda packages and conda uses pip to install pip packages. However a pip package can be converted to a conda package, and then in that case the dependency will be installed by conda and not pip.
You’re correct, Conda does not call pip when it installs packages. But pip does have a subtle involvement here: Many Python packages available to Conda are packaged based on installations made by pip, and (IMO lazily and incorrectly) inherits a lot of pip characteristics. This makes the packages “appear” to be installed by pip, and sometimes induce interpolation inconsistencies.
Given the amount of unique objects and texture maps in the game, this is not a surprise, specially thinking that graphics are the biggest selling points of today's consoles.
If Call of Duty is anything like any other AAA game, it's not because of the number of unique objects and textures, but the size is on purpose. Lots of video games are optimized by _duplicating_ the assets to where they are needed to be loaded, so since they are all in sequence in storage, it'll be faster to load. But this increases the storage requirements, as lots of assets ends up being duplicated all over the place.
I suspect they're not using techniques like procedural generation, which can pack a lot of highly detailed graphics into a tiny space. It's the difference between a vector and raster image.
You do not use realtime procedural textures in AAA games. The resolution for those needs to be quite high (2048px or 4096px) and compiling the shaders into that resolution takes ages.
They however do use procedural textures when texturing objects (Check Substance Designer), but having to compile them during playtime would make load times unacceptably high. Maybe even hours.
I’ve heard it’s literally this. Every asset is duplicated for every map or level it appears in to speed up load times and to make it easier for development teams to work independently on different maps without stepping on each others toes. Also the assets aren’t compressed for faster load times, especially on console. Don’t know how true it is.
It's pretty much a straight trade of load times vs storage— didn't matter back when games were streamed from disk, but it sucks now that they're installed. It's just that load times are measurable and often criticized, whereas no one really talks about installation size, except maybe to remark on it when it's unexpectedly small— for example, FromSoftware's Sekiro, which clocks in at just under 13GB.
for large "open world" games, you sometimes dupe things to lay out the data in memory to be near each other when the items are near each other in the world, so you can stream data in or chunk it in efficiently. If a player drives toward a city you have to be able to load all the buildings and mailboxes in without any pause. it isn't easy
I was constantly reading/writing config files and the structure and comments were what convinced me TOML was worth to try. I did end up going back to JSON for the schema validator bit, but I just found out that TOML is still working on it: https://github.com/toml-lang/toml/pull/116
[1] https://github.com/mamba-org/mamba