From b2e8c878d18d4f1c19bd2f9ba6adccddc008eb9b Mon Sep 17 00:00:00 2001
From: _ <_@_>
Date: Wed, 4 Oct 2023 03:17:24 -0500
Subject: [PATCH] :pencil: doc: failed optimization

---
 notes.md | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/notes.md b/notes.md
index 9d662fd..69cb788 100644
--- a/notes.md
+++ b/notes.md
@@ -64,6 +64,12 @@ Result: Regressed from 3200 to 3600. Not sure why.
 
 Plan: OP_GETFIELD hits the constants a lot. I thought caching it and not dereferencing the chunk and block constantly might help.
 
+## Splitting up the opcode match
+
+Result: No change, 3200 to 3200. Maybe Rust was already optimizing this into an optimal jump table?
+
+Plan: Maybe if the hot inner opcodes, OP_GETFIELD, OP_MUL, and OP_SETFIELD get their own match statement at the top of the function, the step function can exit sooner.
+
 ## Iterating over instruction list
 
 (upcoming)