[Sesse__]

mmap-based I/O still needs to go through the kernel, including memory allocation (in the page cache) and all. If you've got 10x speedup from mmap, it is usually because your explicit I/O was very inefficient; there are situations where mmap is useful, but it's rarely a high-performance strategy, as it's really hard for it to guess what your intended I/O patterns are just from the page faults it's seeing.

[jclarkcom]

Windows uses memory mapped IO for loading all executable processes because it allows you to start executing a process after loading a few pages even if the exe is megabytes. You can use the same to reduce latency for starting to assemble data before the rest of the file loads, the rest can be loaded using more efficienct asynchronous mechanisms. Using for output also means your process doesnt waits on flushes that is also async. And in memory constrained environments the OS doesn’t have to write your data to swap, it can just reload it from the meeting mapped file.

[Sesse__]

Linux also uses mmap for running executables. But explicit I/O does not mean you have to start off by a gigabyte-long read().

[jclarkcom]

More detailed explanation from ChatGPT. As quick estimate you could achieve a >2x speed up using memory mapped files for a typical assembler workload.

https://chatgpt.com/share/68b5e0db-a6d0-8005-9101-d326d2af0a...

[Sesse__]

Why would anyone be interested in arguing against a confused AI?

[jclarkcom]

I was trying to provide a more detailed explanation without typing a lot. I studied this problem a lot as PE at vmware.

[Sesse__]

https://distantprovince.by/posts/its-rude-to-show-ai-output-...

In any case, if you really believe mmap is great for an assembler, then sure, go ahead. But it's not.