From 68b8ca5856619e835ef406c00642a6ad0d708d79 Mon Sep 17 00:00:00 2001
From: Lucas Scharenbroich <lscharen@prowestgis.com>
Date: Tue, 26 Jul 2022 00:11:47 -0500
Subject: [PATCH] Remove a CLC from the inner dispatch. Saves 2 cycles per
 block per sprite at the cost of 2 cycles per sprite

---
 src/Sprite2.s |  1 +
 src/Tiles.s   | 18 +++++++++---------
 2 files changed, 10 insertions(+), 9 deletions(-)

diff --git a/src/Sprite2.s b/src/Sprite2.s
index d622bfb..da52b18 100644
--- a/src/Sprite2.s
+++ b/src/Sprite2.s
@@ -186,6 +186,7 @@ _CalcDirtySprite
         eor   #$FFFF                             ; A = -X - 1
         sec                                      ; C = 1
         adc   _Sprites+SPRITE_DISP,y             ; A = SPRITE_DISP + (-X - 1) + 1 = SPRITE_DISP - X
+        dec                                      ; [!! INTERLOCK !!] pre-decrement to save clc in core blitter. See dobit macros in src/Tiles.s
         sta   _Sprites+TS_VBUFF_BASE,y
 
 ; Create an offset value for loading the calculated VBUFF addresses within the core renderer by
diff --git a/src/Tiles.s b/src/Tiles.s
index c60ba3d..f5ec35d 100644
--- a/src/Tiles.s
+++ b/src/Tiles.s
@@ -571,14 +571,14 @@ dobit       mac
             bcc   next_bit
             beq   last_bit
             tax
-            lda   (SPRITE_VBUFF_PTR+{]1*2}),y
-            clc
-            adc   _Sprites+TS_VBUFF_BASE+{]1*2}
+            lda   (SPRITE_VBUFF_PTR+{]1*2}),y       ; The carry is always set
+;            clc
+            adc   _Sprites+TS_VBUFF_BASE+{]1*2}     ; [!! INTERLOCK !!] The table is pre-decremented in Sprite2.s
             sta   sprite_ptr0+{]2*4}
             txa
             jmp   ]3
 last_bit    lda   (SPRITE_VBUFF_PTR+{]1*2}),y
-            clc    ; pre-adjust these later
+;            clc    ; pre-adjust these later
             adc   _Sprites+TS_VBUFF_BASE+{]1*2}
             sta   sprite_ptr0+{]2*4}
             jmp   ]4
@@ -593,13 +593,13 @@ dobit1      mac
             beq   last_bit
             tax
             lda   (SPRITE_VBUFF_PTR+{]1*2}),y
-            clc
+;            clc
             adc   _Sprites+TS_VBUFF_BASE+{]1*2}
             sta   sprite_ptr0+{]2*4}
             txa
             jmp   ]3
 last_bit    lda   (SPRITE_VBUFF_PTR+{]1*2}),y
-            clc    ; pre-adjust these later
+;            clc    ; pre-adjust these later
             adc   _Sprites+TS_VBUFF_BASE+{]1*2}
             sta   sprite_ptr0+{]2*4}
             tyx
@@ -612,7 +612,7 @@ stpbit      mac
             lsr
             bcc   next_bit
             lda   (SPRITE_VBUFF_PTR+{]1*2}),y
-            clc    ; pre-adjust these later
+;            clc    ; pre-adjust these later
             adc   _Sprites+TS_VBUFF_BASE+{]1*2}
             sta   sprite_ptr0+{]2*4}
             jmp   ]3
@@ -622,7 +622,7 @@ next_bit
 ; Last bit test which *must* be set
 endbit      mac
             lda   (SPRITE_VBUFF_PTR+{]1*2}),y
-            clc    ; pre-adjust these later
+;            clc    ; pre-adjust these later
             adc   _Sprites+TS_VBUFF_BASE+{]1*2}
             sta   sprite_ptr0+{]2*4}
             jmp   ]3
@@ -630,7 +630,7 @@ endbit      mac
 
 endbit1     mac
             lda   (SPRITE_VBUFF_PTR+{]1*2}),y
-            clc    ; pre-adjust these later
+;            clc    ; pre-adjust these later
             adc   _Sprites+TS_VBUFF_BASE+{]1*2}
             sta   sprite_ptr0+{]2*4}
             tyx