It's not a bug, but it's misreading the htop output as mmap doesn't show up as a...

		l33tman on April 1, 2023 \| parent \| context \| favorite \| on: Llama.cpp 30B runs with only 6GB of RAM now It's not a bug, but it's misreading the htop output as mmap doesn't show up as a resident set size there. The pages are RO and not dirty so it's "on the OS" to count it and the OP had lots of RAM on the computer so the model just resides in his page cache instead.

Ahh, this would do it, thanks :).