#48 Support for larger LZMA2 dictionary sizes
Recent versions of 7-Zip now support even larger dictionaries in the LZMA2 context for data deduplication and higher efficiency when compressing very large amounts of data on a system with large amounts of memory (tens to hundreds of gigabytes, or even terabytes).
While p7zip (just like xz) is still limited at 1536 MiB dictionaries, 7-Zip now supports up to 3840 MiB, resulting in roughly 40 GiB of memory use for single-threaded operation and slightly over 500 GiB for 24 compression threads. This seems useful, as machines evolve to have more and more RAM that can be used for deduplication hash tables.
It's specifically effective for compressing e.g. lots of virtual machine disk images for long-term offline storage, or generally large data sets featuring high block-level redundancy.
I would like to request similar (if not even larger) dictionaries to be implemented in p7zip to make use of the resources of large servers and workstations for more efficient compression.
Thank you!