need to finally make a metaprogram for offset math...

code/gte_hello/hello_gte.c

@@ -14,11 +14,7 @@
 #include "duffle/gte.h"
 #include "duffle/lottes_tape.h"
 #include "hello_gte.h"
-
-enum {
-	PrimitiveBuff_Len = 4096,
-	OrderingTbl_Len   = 2048
-};
+#include "hello_gte_tape.c"
 
 typedef U4 OrderingTable_Buffer[OrderingTbl_Len];
 typedef Array_(OrderingTable_Buffer, 2);
@@ -104,6 +100,8 @@ typedef Struct_(SMemory) {
 
 	Ent_Cube  cube;
 	Ent_Floor floor;
+
+
 };
 global SMemory static_mem;
 extern SMemory static_mem;
@@ -245,10 +243,10 @@ void update(PrimitiveArena* pa, U4* ordering_buf)
 		LP_ U4 mem_temp_tape[512]; FArena tape_arena; farena_init(& tape_arena, slice_ut_arr(mem_temp_tape));
 		TapeBuilder tb = tb_make(&tape_arena); tb_scope(& tb) {
 			tb_emit(& tb, code_rbind_cube_tri);
-			tb_data(& tb, prim_cursor);
-			tb_data(& tb, u4_(static_mem.cube.faces));
-			tb_data(& tb, u4_(static_mem.cube.verts));
-			tb_data(& tb, u4_(ordering_buf));
+				tb_data(& tb, prim_cursor);
+				tb_data(& tb, u4_(static_mem.cube.faces));
+				tb_data(& tb, u4_(static_mem.cube.verts));
+				tb_data(& tb, u4_(ordering_buf));
 
 			for (U4 i = 0; i < Cube_num_faces; i++) {
 				// Two triangles per quad face: (x,y,z) and (x,z,w)
@@ -256,8 +254,8 @@ void update(PrimitiveArena* pa, U4* ordering_buf)
 			}
 
 			tb_emit(& tb, code_sync_prim_cursor);
-			tb_data(& tb, u4_(& pa->used));
-			tb_data(& tb, prim_base);
+				tb_data(& tb, u4_(& pa->used));
+				tb_data(& tb, prim_base);
 		}
 		tape_run(tb_slice(tb));
 
@@ -334,21 +332,24 @@ void update(PrimitiveArena* pa, U4* ordering_buf)
 		LP_ U4 mem_temp_tape[512]; FArena tape_arena; farena_init(& tape_arena, slice_ut_arr(mem_temp_tape));
 		TapeBuilder tb = tb_make(&tape_arena); tb_scope(& tb) {
 			// Push "Protocol" to tape
-			tb_emit_(& tb, rbind_floor_tri);
+			tb_emit(& tb, code_rbind_floor_tri);
+				// Note(Ed): This is technically argument shuffle and would be better if we did a single reference at most to a global batch context.
+				// Note(Ed): We can technically allocate a single ptr with the global offset to the working context instead of utilizing the tape for this stack of refs.
 				tb_data(& tb, prim_cursor);
 				tb_data(& tb, u4_(static_mem.floor.faces));
 				tb_data(& tb, u4_(static_mem.floor.verts));
 				tb_data(& tb, u4_(ordering_buf));
+
 			tb_emit(& tb, code_set_gte_world);
-			tb_data(& tb, u4_(& static_mem.tform_world));
-            
+				tb_data(& tb, u4_(& static_mem.tform_world));
+
 			for (U4 i = 0; i < Floor_num_faces; i++) {
 				tb_emit(& tb, code_floor_tri);
 			}
 			
 			tb_emit(& tb, code_sync_prim_cursor);
-			tb_data(& tb, u4_(& pa->used));
-			tb_data(& tb, prim_base);
+				tb_data(& tb, u4_(& pa->used));
+				tb_data(& tb, prim_base);
 		}
 
 		// Fire off the tape.

code/gte_hello/hello_gte.h

@@ -5,6 +5,11 @@
 #	include "duffle/gp.h"
 #endif
 
+enum {
+	PrimitiveBuff_Len = 4096,
+	OrderingTbl_Len   = 2048
+};
+
 typedef Struct_(DrawEnv_Packed) { U4 tag; U4 code[15]; };
 typedef Struct_(DrawEnv) {
 	Rect_S2 clip_area;

code/gte_hello/hello_gte_tape.c

@@ -0,0 +1,61 @@
+#ifdef INTELLISENSE_DIRECTIVES
+#	include "duffle/lottes_tape.h"
+#	include "hello_gte.h"
+#endif
+
+#pragma region MACs
+
+/* Words: 3; High: 0x20/B, Low: G/R */
+#define mac_format_f3_color(color_hi, color_lo) \
+	  load_ui(R_AT, color_hi), or_i(R_AT, R_AT, color_lo) \
+	, store_word(R_AT, R_PrimCur, O_(Poly_F3,color))   \
+
+/* Words: 3 */
+#define mac_gte_store_f3() \
+	  gte_sw(C2_SXY0, R_PrimCur, O_(Poly_F3,p0)) \
+	, gte_sw(C2_SXY1, R_PrimCur, O_(Poly_F3,p1)) \
+	, gte_sw(C2_SXY2, R_PrimCur, O_(Poly_F3,p2))
+
+#pragma endregion MACs
+
+#pragma region Baked Atoms
+
+internal MipsAtom_(floor_tri) {
+	// T0-T2 allocated
+	mac_load_tri_indices(R_T0, R_T1, R_T2),
+	mac_load_tri_verts(  R_T0, R_T1, R_T2),
+
+	/* 3. Execute Math */
+	nop, nop, gte_cmdw_rotate_translate_perspective_triple,
+	nop, nop, gte_cmdw_nclip,
+	nop, nop, 
+
+	/* 4. Culling (Branch forward 29 instructions if Backface) */
+	gte_mf(R_T0, C2_MAC0),
+	nop, branch_le_zero(R_T0, 29), 
+	nop,                      
+	
+	/* 5. Format Primitive */
+	mac_format_f3_color(0x20FF, 0xFFFF), 
+	mac_gte_store_f3(),
+	
+	/* 6. Calculate Depth */
+	nop, nop, gte_avg_sort_z3,            
+	nop, nop, gte_mf(R_T1, C2_OTZ),      
+	
+	/* 7. Bounds Check OTZ < 2048 (Branch forward 13 instructions to skip insertion) */
+	add_ui(      R_AT, R_0,  OrderingTbl_Len),   
+	slt_u(       R_AT, R_T1, R_AT), 
+	branch_equal(R_AT, R_0,  13),   
+	nop, 
+	
+	/* 8. Insert into Ordering Table Linked List */
+	mac_insert_ot_tag(R_T1, 0x0400),
+	add_ui(R_PrimCur, R_PrimCur, S_(Poly_F3)), /* Advance Prim Cursor (5 words) */
+
+	/* 9. Advance Input Cursor & Yield (Both branch targets land here) */
+	add_ui(R_FaceCur, R_FaceCur, S_(S2) * 4),  /* Advance Face Cursor (4 * S2 = 8 bytes) */
+	mac_yield()
+};
+
+#pragma endregion Baked Atoms