I want to have the option to customize application buffer sizes so as to optimize IO operations.
Point of case: I was trying to extract a single-file, 45+GB in size, from a RAR archive. This took almost seven hours, because the PA Toolbox kept reading and writing data to HDD almost simultaneously.
On computers with decent amounts of RAM (> 8GB), I would like to have the option to control the buffer sizes for compression:
Input: Up to 128 MB.
Output: Up to 1 GB.
This way, PowerArchiver Toolbox 2015 does not read 32 KB of data, or whatever small amount it currently does, but reads my proposed 128 MB in one pass, leaves the HDD alone while it’s performing the in-memory extraction, then writes data back to the HDD.
I believe this will add significant performance improvements, because of no longer having to wait for the hard drive to recover from reading/ reading in a location different than the previous one.
On modern hard drives, reading or writing to consecutive blocks is very optimized. Let’s use that. And let’s also better use the amounts of RAM on modern computers.
If we could also imagine a compression mechanism that uses larger-size buffers, that would be great. It will definitely help us improve on our IO usage and performance (due to not congesting IO but by coordinating I and O for smart operations).