4 ms·

Afaik memory-mapped files don't count towards the commit limit. Couldn't they roll a custom file-backed allocator (essentially manual swap) to get around the wi

Afaik memory-mapped files don't count towards the commit limit. Couldn't they roll a custom file-backed allocator (essentially manual swap) to get around the windows limit?

323 4y ago
Not sure, but probably just the file-backed memory mapped files (as opposed to the anonymous ones) don't count.
dr-detroit 4y ago
dundarious 4y ago
It’s good to somewhat limit the amount of memory that gets paged to disk, as it is incredibly slow memory. There should be backpressure on memory allocation. Getting around the OS on this matter would be silly, IMO. We’re talking about an interactive UI environment where massive page files would be doubly bad, swallowing up surprising amounts of disk space in order to provide incredibly slow memory. However, the type of solution you’re talking about would apply well to a caching server, and Varnish (on FreeBSD I believe) uses it quite successfully.
zyx321 4y ago
Please don't. I've set a maximum swap size per partition, and you are suggesting to bypass it by falsely declaring your swap as cache? Who died and made you NT_AUTHORITY\SYSTEM?
justsomehnguy 4y ago
Reminds me how some folk (uTorrent team, AFAIR) thought they are smarter than those guys in Redmond and rolled out their own caching algorithm. Except they weren't in fact smarter and fucked up the memory usage by allocating all the availble memory for a disk cache... pushing everything else including OS to the swap.
- pas 4y ago
  if the OS was pushable it's not an OS it's just an app. memory management is hard, and all major OSes still have serious problems with it.
guenthert 4y ago
Don't give them ideas about how to use even more memory. For someone who's first computer had 64KiB RAM, it's incomprehensible why rendering a single web page should consume 100MiB.
- astrange 4y ago
  Web pages are Turing complete and can do whatever they want. Web browsers are VMs in a world where nobody wants to help them do their job efficiently.

Afaik memory-mapped files don't count towards the commit limit. Couldn't they roll a custom file-backed allocator (essentially manual swap) to get around the windows limit?

323 4y ago

Not sure, but probably just the file-backed memory mapped files (as opposed to the anonymous ones) don't count.

dr-detroit 4y ago

dundarious 4y ago

It’s good to somewhat limit the amount of memory that gets paged to disk, as it is incredibly slow memory. There should be backpressure on memory allocation. Getting around the OS on this matter would be silly, IMO. We’re talking about an interactive UI environment where massive page files would be doubly bad, swallowing up surprising amounts of disk space in order to provide incredibly slow memory. However, the type of solution you’re talking about would apply well to a caching server, and Varnish (on FreeBSD I believe) uses it quite successfully.

zyx321 4y ago

Please don't. I've set a maximum swap size per partition, and you are suggesting to bypass it by falsely declaring your swap as cache? Who died and made you NT_AUTHORITY\SYSTEM?

justsomehnguy 4y ago

Reminds me how some folk (uTorrent team, AFAIR) thought they are smarter than those guys in Redmond and rolled out their own caching algorithm. Except they weren't in fact smarter and fucked up the memory usage by allocating all the availble memory for a disk cache... pushing everything else including OS to the swap.

pas 4y ago
if the OS was pushable it's not an OS it's just an app. memory management is hard, and all major OSes still have serious problems with it.

guenthert 4y ago

Don't give them ideas about how to use even more memory. For someone who's first computer had 64KiB RAM, it's incomprehensible why rendering a single web page should consume 100MiB.

astrange 4y ago
Web pages are Turing complete and can do whatever they want. Web browsers are VMs in a world where nobody wants to help them do their job efficiently.