Fcalva/cyclone
1
0
Fork 0
mirror of https://github.com/justinethier/cyclone.git synced 2025-05-18 21:29:18 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions

Cleanup

Browse source
This commit is contained in:
Justin Ethier 2022-11-23 22:15:19 -05:00
parent 9cab675f7d
commit 45f97bb85a
1 changed files with 2 additions and 2 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
docs/Garbage-Collector-Revised-2022.md
View file

@ -312,7 +312,7 @@ Performance is improved in several ways:
- Better Locality - Heap slots tend to be used soon after they are swept and sweep only needs to visit a small part of the heap. This allows programs to make better use of the processor cache.
- Thread-Local Data - There is no need to lock the heap for allocation or sweeping since both operations are performed by the same thread.
- Reduced Complexity - According to [[1]](#further-reading) the algorithmic complexity of mark-sweep is reduced to be proportional to the size of the live data in the heap instead of the whole heap, similar to a copying collector. Lazy sweeping will perform best when most of the heap is empty.
- Reduced Complexity - The algorithmic complexity of mark-sweep is reduced to be proportional to the size of the live data in the heap instead of the whole heap, similar to a copying collector. Lazy sweeping will perform best when most of the heap is empty.
This section describes several related changes to our collector that were required to support lazy sweeping.
@ -390,7 +390,7 @@ The existing GC tracked free space and would start a major GC once the amount of
### Performance Measurements
A benchmark suite [[3]](#further-reading) was used to compare performance between the previous version of Cyclone (0.8.1) and the new version with lazy sweeping.
A [benchmark suite](#further-reading) was used to compare performance between the previous version of Cyclone (0.8.1) and the new version with lazy sweeping.
The following table lists the differences in elapsed time (seconds) between versions: