Date: Tue, 16 Mar 2021 22:08:17 -0500 From: Dennis Williamson <dennistwilliam...@gmail.com> Message-ID: <canaoh6kr2pnymc79+gd6xepnywxnuctmx3zu0nyfekjkwhs...@mail.gmail.com>
| In any case this points to larger chunks being more efficient. This is not news, doing reads (or writes) using bigger buffers (more bytes at a time) means less system calls, and compared to almost anything else that happens, system calls are expensive - doing less will make things run faster. Unfortunately, it can also alter the semantics of what is happening - whether it does or not entirely depends upon the application. Lastly, if you really care about any of this - that is, if it really matters to something more than your ego - then you shouldn't be writing in shell at all, but using a compiled language. Lots of things in shell involve running other programs, and running another program usually involves large numbers of system calls (including fork() and exec() - two of the most costly that exist). Write shell scripts to do things where all the time is either spent waiting for something to happen (like the user to type the next command) or in some other program that the shell runs (like a compiler, or document processor) and no-one cares just how efficient the shell might be in making all of that happen, as in the overall scheme of things it makes no difference (kind of like worrying how quickly you can drive down your driveway to get in or out of your house when the real issue is how long it takes you to get to/from work). kre