aboutsummaryrefslogtreecommitdiff
path: root/README.md
diff options
context:
space:
mode:
authorMikkel Thestrup <mithe24@student.sdu.dk>2025-12-05 11:33:16 +0100
committerMikkel Thestrup <mithe24@student.sdu.dk>2025-12-05 11:41:09 +0100
commit7a00f88749b00c7eb0404872b21e36adb17d8456 (patch)
tree51c387c8a46b80984526aea74b4e4d703bf15bea /README.md
parentf624eaf7cb786fea5e574637ccd31fbed52eec5a (diff)
downloadcycle-detector-7a00f88749b00c7eb0404872b21e36adb17d8456.tar.gz
cycle-detector-7a00f88749b00c7eb0404872b21e36adb17d8456.zip
Small changes
Diffstat (limited to 'README.md')
-rw-r--r--README.md11
1 files changed, 11 insertions, 0 deletions
diff --git a/README.md b/README.md
index ea29f50..4581aa9 100644
--- a/README.md
+++ b/README.md
@@ -74,3 +74,14 @@ To list all available build targets and their descriptions:
```bash
make help
```
+
+## Adjacency List Storage: Linked List vs Array
+
+Two approaches for storing graph neighbors: linked lists and dynamic arrays.
+Benchmarks show no measurable performance difference
+despite vectors offering better cache locality.
+
+Why? Kahn's algorithm is $O(V+E) $, so memory access patterns
+likely isn't the bottleneck.
+
+Cache locality could become a bottleneck on huge graphs