Fast memory allocation

lda commented

2024-06-06 10:14:29 +00:00

Contributor

After messing around with it with Telodendria(and some projects I'm working on), it's fairly stable.
It still doesn't implement checksums(I also need to change the magic number, maybe to some randomly generated value made at runtime), and effectively disable MemoryIterate during mallocs, reallocs, and frees.

This still however removes `MemoryIterate' from allocations, because it's a real performance pickle.
From my tests parsing large sync
replies, this commit is near instant(pv reports 23MiB/s on a 2MB sync).

Still need a good compromise along MemoryIterate(like maybe find out a clever way to only make it run on a small subset, or maybe just randomly run it from time to time, or maybe just roll without it except on some soft of debug mode????)

After messing around with it with Telodendria(and some projects I'm working on), it's fairly stable. It still doesn't implement checksums(I also need to change the magic number, maybe to some randomly generated value made at runtime), and effectively disable MemoryIterate during mallocs, reallocs, and frees. >This still however removes `MemoryIterate' from allocations, because it's a real performance pickle. > From my tests parsing large sync replies, this commit is near instant(pv reports 23MiB/s on a 2MB sync). > > Still need a good compromise along MemoryIterate(like maybe find out a clever way to only make it run on a small subset, or maybe just randomly run it from time to time, or maybe just roll without it except on some soft of debug mode????)

lda added 1 commit 2024-06-06 10:14:29 +00:00

[MOD/WIP] Blazing-fast memory allocator 🚀🚀🚀🚀🚀🚀🚀🚀🚀🚀 007b8f6d43

This still however removes `MemoryIterate' from allocations, because
it's a real performance pickle. From my tests parsing large sync
replies, this commit is near instant(pv reports 23MiB/s on a 2MB sync).

Still need a good compromise along MemoryIterate(like maybe find out a
clever way to only make it run on a small subset, or maybe just randomly
run it from time to time, or maybe just roll without it except on some
soft of debug mode????)

lda added 1 commit 2024-06-08 10:08:50 +00:00

Merge branch 'master' of https://git.telodendria.io/Telodendria/Cytoplasm into mem-moment 5003ddc281

lda added 2 commits 2024-06-15 11:44:29 +00:00

Merge branch 'master' of https://git.telodendria.io/Telodendria/Cytoplasm into mem-moment 9f706102c4

Merge branch 'mem-moment' of https://git.telodendria.io/lda/Cytoplasm into mem-moment 1ad6f0d976

lda added 1 commit 2024-06-15 11:49:16 +00:00

[FIX] Actually do proper stringification. 8b2bdbe220

It just kept bothering me.

lda added 1 commit 2024-06-19 15:51:30 +00:00

[MOD] Use a different constant bec672c92c

lda added 1 commit 2024-06-27 14:43:56 +00:00

Merge branch 'master' of https://git.telodendria.io/Telodendria/Cytoplasm into mem-moment 402d73c866

lda added 1 commit 2024-07-18 11:20:32 +00:00

Merge branch 'master' of https://git.telodendria.io/Telodendria/Cytoplasm into mem-moment 0d122976d4

lda changed title from ~~WIP: Fast memory allocation~~ to Fast memory allocation

2024-08-06 12:03:58 +00:00

jordan commented

2024-08-18 23:25:17 +00:00

Owner

The code looks good, I'll have to find the time to go through it more to see if it's something we want to merge. I'll have to decide if the functionality we're losing can be implemented in any other way. I'll think about it and try to make a decision either way as soon as I can.

jordan commented

2024-08-23 20:26:09 +00:00

Owner

I think, for now, if you're not running into any memory issue that the iterator is detecting, we can probably roll without it for now. We can re-evaluate if we find some gnarly memory bugs that we just can't fix without it.

jordan merged commit 494be7b4dc into master

2024-08-23 20:26:14 +00:00

jordan referenced this pull request from a commit

2024-08-23 20:26:14 +00:00

Merge pull request 'Fast memory allocation' (#36) from lda/Cytoplasm:mem-moment into master

Fast memory allocation #36