mirror/openlara
1
0
Fork 0
mirror of https://github.com/XProger/OpenLara.git synced 2025-08-06 05:07:48 +02:00
Code Issues Releases Wiki Activity

#368 micro optimizations

Browse Source
This commit is contained in:
XProger
2022-02-13 18:21:19 +03:00
parent 6656837473
commit f434d45359
6 changed files with 46 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
src/fixed/common.h
View File

@@ -157,16 +157,19 @@
#define X_INLINE inline #define X_INLINE inline
#define X_NOINLINE __declspec(noinline) #define X_NOINLINE __declspec(noinline)
#define ALIGN4 __declspec(align(4)) #define ALIGN4 __declspec(align(4))
#define ALIGN8 __declspec(align(8))
#define ALIGN16 __declspec(align(16)) #define ALIGN16 __declspec(align(16))
#elif defined(__WATCOMC__) || defined(__3DO__) #elif defined(__WATCOMC__) || defined(__3DO__)
#define X_INLINE inline #define X_INLINE inline
#define X_NOINLINE #define X_NOINLINE
#define ALIGN4 #define ALIGN4
#define ALIGN8
#define ALIGN16 #define ALIGN16
#else #else
#define X_INLINE __attribute__((always_inline)) inline #define X_INLINE __attribute__((always_inline)) inline
#define X_NOINLINE __attribute__((noinline)) #define X_NOINLINE __attribute__((noinline))
#define ALIGN4 __attribute__((aligned(4))) #define ALIGN4 __attribute__((aligned(4)))
#define ALIGN8 __attribute__((aligned(8)))
#define ALIGN16 __attribute__((aligned(16))) #define ALIGN16 __attribute__((aligned(16)))
#endif #endif
@@ -689,6 +692,7 @@ struct Face
int32 ccb_HDDX; int32 ccb_HDDX;
int32 ccb_HDDY; int32 ccb_HDDY;
uint32 ccb_PIXC; uint32 ccb_PIXC;
// TODO use 1x1 textures instead of colored faces to remove preamble words (8 bytes per face - 15k)
uint32 ccb_PRE0; uint32 ccb_PRE0;
uint32 ccb_PRE1; uint32 ccb_PRE1;
//int32 ccb_Width; //int32 ccb_Width;

23
src/platform/gba/asm/faceAddMeshQuads.s
View File

@@ -32,6 +32,8 @@ tmp .req flags
vertices .req vg2 vertices .req vg2
next .req vp0 next .req vp0
SP_SIZE = 4
.global faceAddMeshQuads_asm .global faceAddMeshQuads_asm
faceAddMeshQuads_asm: faceAddMeshQuads_asm:
stmfd sp!, {r4-r11, lr} stmfd sp!, {r4-r11, lr}
@@ -39,6 +41,10 @@ faceAddMeshQuads_asm:
ldr vp, =gVerticesBase ldr vp, =gVerticesBase
ldr vp, [vp] ldr vp, [vp]
ldr vertices, =gVertices
lsr vertices, #3
stmfd sp!, {vertices}
ldr face, =gFacesBase ldr face, =gFacesBase
ldr face, [face] ldr face, [face]
@@ -90,16 +96,14 @@ faceAddMeshQuads_asm:
lsr depth, #(2 + OT_SHIFT) lsr depth, #(2 + OT_SHIFT)
// faceAdd // faceAdd
ldr vertices, =gVertices ldr vertices, [sp]
sub vp0, vertices rsb vp0, vertices, vp0, lsr #3
sub vp1, vertices rsb vp1, vertices, vp1, lsr #3
sub vp2, vertices rsb vp2, vertices, vp2, lsr #3
sub vp3, vertices rsb vp3, vertices, vp3, lsr #3
lsr vp0, #3 orr vp1, vp0, vp1, lsl #16
orr vp1, vp0, vp1, lsl #(16 - 3) orr vp3, vp2, vp3, lsl #16
lsr vp2, #3
orr vp3, vp2, vp3, lsl #(16 - 3)
ldr next, [ot, depth, lsl #2] ldr next, [ot, depth, lsl #2]
str face, [ot, depth, lsl #2] str face, [ot, depth, lsl #2]
@@ -111,4 +115,5 @@ faceAddMeshQuads_asm:
ldr tmp, =gFacesBase ldr tmp, =gFacesBase
str face, [tmp] str face, [tmp]
add sp, #SP_SIZE
ldmfd sp!, {r4-r11, pc} ldmfd sp!, {r4-r11, pc}

11
src/platform/gba/asm/faceAddMeshTriangles.s
View File

@@ -42,6 +42,7 @@ faceAddMeshTriangles_asm:
ldr ot, =gOT ldr ot, =gOT
ldr vertices, =gVertices ldr vertices, =gVertices
lsr vertices, #3
add polys, #2 // skip flags add polys, #2 // skip flags
@@ -82,13 +83,11 @@ faceAddMeshTriangles_asm:
lsr depth, #(2 + OT_SHIFT) lsr depth, #(2 + OT_SHIFT)
// faceAdd // faceAdd
sub vp0, vertices rsb vp0, vertices, vp0, lsr #3
sub vp1, vertices rsb vp1, vertices, vp1, lsr #3
sub vp2, vertices rsb vp2, vertices, vp2, lsr #3
lsr vp0, #3 orr vp1, vp0, vp1, lsl #16
orr vp1, vp0, vp1, lsl #(16 - 3)
lsr vp2, #3
orr flags, #FACE_TRIANGLE orr flags, #FACE_TRIANGLE

23
src/platform/gba/asm/faceAddRoomQuads.s
View File

@@ -32,6 +32,8 @@ tmp .req flags
vertices .req vg2 vertices .req vg2
next .req vp0 next .req vp0
SP_SIZE = 4
.global faceAddRoomQuads_asm .global faceAddRoomQuads_asm
faceAddRoomQuads_asm: faceAddRoomQuads_asm:
stmfd sp!, {r4-r11, lr} stmfd sp!, {r4-r11, lr}
@@ -39,6 +41,10 @@ faceAddRoomQuads_asm:
ldr vp, =gVerticesBase ldr vp, =gVerticesBase
ldr vp, [vp] ldr vp, [vp]
ldr vertices, =gVertices
lsr vertices, #3
stmfd sp!, {vertices}
ldr face, =gFacesBase ldr face, =gFacesBase
ldr face, [face] ldr face, [face]
@@ -101,16 +107,14 @@ faceAddRoomQuads_asm:
mov depth, vz0, lsr #OT_SHIFT mov depth, vz0, lsr #OT_SHIFT
// faceAdd // faceAdd
ldr vertices, =gVertices ldr vertices, [sp]
sub vp0, vertices rsb vp0, vertices, vp0, lsr #3
sub vp1, vertices rsb vp1, vertices, vp1, lsr #3
sub vp2, vertices rsb vp2, vertices, vp2, lsr #3
sub vp3, vertices rsb vp3, vertices, vp3, lsr #3
lsr vp0, #3 orr vp1, vp0, vp1, lsl #16
orr vp1, vp0, vp1, lsl #(16 - 3) orr vp3, vp2, vp3, lsl #16
lsr vp2, #3
orr vp3, vp2, vp3, lsl #(16 - 3)
ldr next, [ot, depth, lsl #2] ldr next, [ot, depth, lsl #2]
str face, [ot, depth, lsl #2] str face, [ot, depth, lsl #2]
@@ -122,4 +126,5 @@ faceAddRoomQuads_asm:
ldr tmp, =gFacesBase ldr tmp, =gFacesBase
str face, [tmp] str face, [tmp]
add sp, #SP_SIZE
ldmfd sp!, {r4-r11, pc} ldmfd sp!, {r4-r11, pc}

11
src/platform/gba/asm/faceAddRoomTriangles.s
View File

@@ -42,6 +42,7 @@ faceAddRoomTriangles_asm:
ldr ot, =gOT ldr ot, =gOT
ldr vertices, =gVertices ldr vertices, =gVertices
lsr vertices, #3
add polys, #2 // skip flags add polys, #2 // skip flags
@@ -91,13 +92,11 @@ faceAddRoomTriangles_asm:
mov depth, vz0, lsr #OT_SHIFT mov depth, vz0, lsr #OT_SHIFT
// faceAdd // faceAdd
sub vp0, vertices rsb vp0, vertices, vp0, lsr #3
sub vp1, vertices rsb vp1, vertices, vp1, lsr #3
sub vp2, vertices rsb vp2, vertices, vp2, lsr #3
lsr vp0, #3 orr vp1, vp0, vp1, lsl #16
orr vp1, vp0, vp1, lsl #(16 - 3)
lsr vp2, #3
orr flags, #FACE_TRIANGLE orr flags, #FACE_TRIANGLE

4
src/platform/gba/render.iwram.cpp
View File

@@ -66,8 +66,8 @@ Vertex* gVerticesBase;
Face* gFacesBase; Face* gFacesBase;
EWRAM_DATA uint8 gBackgroundCopy[FRAME_WIDTH * FRAME_HEIGHT]; // EWRAM 37.5k EWRAM_DATA uint8 gBackgroundCopy[FRAME_WIDTH * FRAME_HEIGHT]; // EWRAM 37.5k
EWRAM_DATA Vertex gVertices[MAX_VERTICES]; // EWRAM 16k EWRAM_DATA ALIGN8 Vertex gVertices[MAX_VERTICES]; // EWRAM 16k
EWRAM_DATA Face gFaces[MAX_FACES]; // EWRAM 5k EWRAM_DATA Face gFaces[MAX_FACES]; // EWRAM 30k
Face* gOT[OT_SIZE]; // IWRAM 2.5k Face* gOT[OT_SIZE]; // IWRAM 2.5k
enum ClipFlags { enum ClipFlags {