Unroll the core loop of strlen one time.

This makes the core loop about 10% faster at the cost of 5 extra code bytes, which seems like a reasonable tradeoff.
2024-02-26 22:14:59 -06:00 · 2024-02-26 22:14:59 -06:00 · 49ffb1065b
parent 9181b0bd73
commit 49ffb1065b
1 changed files with 3 additions and 0 deletions
--- a/string.asm
+++ b/string.asm
@ -961,6 +961,9 @@ str      equ   4                        pointer to the string
         tyx                             null
         short M
 lb1      lda   [str],Y
+         beq   lb2
+         iny
+         lda   [str],Y
         beq   lb2
         iny
         bne   lb1