Snes9x - Use for loops for DRAW_TILE. 3-4% speed benefit. Backport Select brightness function in tile renderer. (#828)

- Use for loops for DRAW_TILE. 3-4% speed benefit.
- Select brightness function in tile renderer.
This commit is contained in:
bladeoner 2019-03-27 22:17:24 +01:00 committed by dborth
parent 6dcba441ea
commit 56fb0046ba
2 changed files with 67 additions and 51 deletions

View File

@ -131,8 +131,9 @@ extern struct SGFX GFX;
((((((C1) & RGB_REMOVE_LOW_BITS_MASK) + \ ((((((C1) & RGB_REMOVE_LOW_BITS_MASK) + \
((C2) & RGB_REMOVE_LOW_BITS_MASK)) >> 1) + \ ((C2) & RGB_REMOVE_LOW_BITS_MASK)) >> 1) + \
((C1) & (C2) & RGB_LOW_BITS_MASK)) | ALPHA_BITS_MASK) ((C1) & (C2) & RGB_LOW_BITS_MASK)) | ALPHA_BITS_MASK)
#define COLOR_ADD_BRIGHTNESS1_2 COLOR_ADD1_2
inline uint16 COLOR_ADD(uint16 C1, uint16 C2) inline uint16 COLOR_ADD_BRIGHTNESS(uint16 C1, uint16 C2)
{ {
return ((brightness_cap[ (C1 >> RED_SHIFT_BITS) + (C2 >> RED_SHIFT_BITS) ] << RED_SHIFT_BITS) | return ((brightness_cap[ (C1 >> RED_SHIFT_BITS) + (C2 >> RED_SHIFT_BITS) ] << RED_SHIFT_BITS) |
(brightness_cap[((C1 >> GREEN_SHIFT_BITS) & 0x1f) + ((C2 >> GREEN_SHIFT_BITS) & 0x1f)] << GREEN_SHIFT_BITS) | (brightness_cap[((C1 >> GREEN_SHIFT_BITS) & 0x1f) + ((C2 >> GREEN_SHIFT_BITS) & 0x1f)] << GREEN_SHIFT_BITS) |
@ -143,27 +144,39 @@ inline uint16 COLOR_ADD(uint16 C1, uint16 C2)
(brightness_cap[ (C1 & 0x1f) + (C2 & 0x1f)] )); (brightness_cap[ (C1 & 0x1f) + (C2 & 0x1f)] ));
} }
inline uint16 COLOR_ADD(uint16 C1, uint16 C2)
{
const int RED_MASK = 0x1F << RED_SHIFT_BITS;
const int GREEN_MASK = 0x1F << GREEN_SHIFT_BITS;
const int BLUE_MASK = 0x1F;
int rb = C1 & (RED_MASK | BLUE_MASK);
rb += C2 & (RED_MASK | BLUE_MASK);
int rbcarry = rb & ((0x20 << RED_SHIFT_BITS) | (0x20 << 0));
int g = (C1 & (GREEN_MASK)) + (C2 & (GREEN_MASK));
int rgbsaturate = (((g & (0x20 << GREEN_SHIFT_BITS)) | rbcarry) >> 5) * 0x1f;
return (uint16)((rb & (RED_MASK | BLUE_MASK)) | (g & GREEN_MASK) | rgbsaturate);
}
#define COLOR_SUB1_2(C1, C2) \ #define COLOR_SUB1_2(C1, C2) \
GFX.ZERO[(((C1) | RGB_HI_BITS_MASKx2) - \ GFX.ZERO[(((C1) | RGB_HI_BITS_MASKx2) - \
((C2) & RGB_REMOVE_LOW_BITS_MASK)) >> 1] ((C2) & RGB_REMOVE_LOW_BITS_MASK)) >> 1]
inline uint16 COLOR_SUB (uint16 C1, uint16 C2) inline uint16 COLOR_SUB (uint16 C1, uint16 C2)
{ {
uint16 mC1, mC2, v = ALPHA_BITS_MASK; int a = (C1 & (THIRD_COLOR_MASK | FIRST_COLOR_MASK)) | ((0x20 << 0) | (0x20 << RED_SHIFT_BITS));
int b = C2 & (THIRD_COLOR_MASK | FIRST_COLOR_MASK);
mC1 = C1 & FIRST_COLOR_MASK; int c = a - b;
mC2 = C2 & FIRST_COLOR_MASK; int d = c & ((0x20 << RED_SHIFT_BITS) | (0x20 << 0));
if (mC1 > mC2) v += (mC1 - mC2); int e = (C1 & (SECOND_COLOR_MASK)) | (0x20 << GREEN_SHIFT_BITS);
int f = C2 & (SECOND_COLOR_MASK);
mC1 = C1 & SECOND_COLOR_MASK; int g = e - f;
mC2 = C2 & SECOND_COLOR_MASK; int h = (g & (0x20 << GREEN_SHIFT_BITS)) | d;
if (mC1 > mC2) v += (mC1 - mC2); int i = h >> 5;
int j = i * 0x1F;
mC1 = C1 & THIRD_COLOR_MASK; int k = ((c & (THIRD_COLOR_MASK | FIRST_COLOR_MASK)) | (g & SECOND_COLOR_MASK)) & j;
mC2 = C2 & THIRD_COLOR_MASK; return (uint16)k;
if (mC1 > mC2) v += (mC1 - mC2);
return (v);
} }
void S9xStartScreenRefresh (void); void S9xStartScreenRefresh (void);

View File

@ -21,7 +21,6 @@ static uint32 pixbit[8][16];
static uint8 hrbit_odd[256]; static uint8 hrbit_odd[256];
static uint8 hrbit_even[256]; static uint8 hrbit_even[256];
void S9xInitTileRenderer (void) void S9xInitTileRenderer (void)
{ {
int i; int i;
@ -400,6 +399,14 @@ void S9xSelectTileRenderers (int BGMode, bool8 sub, bool8 obj)
if (Memory.FillRAM[0x2130] & 2) if (Memory.FillRAM[0x2130] & 2)
i++; i++;
} }
if (IPPU.MaxBrightness != 0xf)
{
if (i == 1)
i = 7;
else if (i == 3)
i = 8;
}
} }
GFX.DrawTileMath = DT[i]; GFX.DrawTileMath = DT[i];
@ -578,14 +585,9 @@ void S9xSelectTileConverter (int depth, bool8 hires, bool8 sub, bool8 mosaic)
OFFSET_IN_LINE; \ OFFSET_IN_LINE; \
for (l = LineCount; l > 0; l--, bp += 8 * PITCH, Offset += GFX.PPL) \ for (l = LineCount; l > 0; l--, bp += 8 * PITCH, Offset += GFX.PPL) \
{ \ { \
DRAW_PIXEL(0, Pix = bp[0]); \ for (int x = 0; x < 8; x++) { \
DRAW_PIXEL(1, Pix = bp[1]); \ DRAW_PIXEL(x, Pix = bp[x]); \
DRAW_PIXEL(2, Pix = bp[2]); \ } \
DRAW_PIXEL(3, Pix = bp[3]); \
DRAW_PIXEL(4, Pix = bp[4]); \
DRAW_PIXEL(5, Pix = bp[5]); \
DRAW_PIXEL(6, Pix = bp[6]); \
DRAW_PIXEL(7, Pix = bp[7]); \
} \ } \
} \ } \
else \ else \
@ -595,14 +597,9 @@ void S9xSelectTileConverter (int depth, bool8 hires, bool8 sub, bool8 mosaic)
OFFSET_IN_LINE; \ OFFSET_IN_LINE; \
for (l = LineCount; l > 0; l--, bp += 8 * PITCH, Offset += GFX.PPL) \ for (l = LineCount; l > 0; l--, bp += 8 * PITCH, Offset += GFX.PPL) \
{ \ { \
DRAW_PIXEL(0, Pix = bp[7]); \ for (int x = 0; x < 8; x++) { \
DRAW_PIXEL(1, Pix = bp[6]); \ DRAW_PIXEL(x, Pix = bp[7 - x]); \
DRAW_PIXEL(2, Pix = bp[5]); \ } \
DRAW_PIXEL(3, Pix = bp[4]); \
DRAW_PIXEL(4, Pix = bp[3]); \
DRAW_PIXEL(5, Pix = bp[2]); \
DRAW_PIXEL(6, Pix = bp[1]); \
DRAW_PIXEL(7, Pix = bp[0]); \
} \ } \
} \ } \
else \ else \
@ -612,14 +609,9 @@ void S9xSelectTileConverter (int depth, bool8 hires, bool8 sub, bool8 mosaic)
OFFSET_IN_LINE; \ OFFSET_IN_LINE; \
for (l = LineCount; l > 0; l--, bp -= 8 * PITCH, Offset += GFX.PPL) \ for (l = LineCount; l > 0; l--, bp -= 8 * PITCH, Offset += GFX.PPL) \
{ \ { \
DRAW_PIXEL(0, Pix = bp[0]); \ for (int x = 0; x < 8; x++) { \
DRAW_PIXEL(1, Pix = bp[1]); \ DRAW_PIXEL(x, Pix = bp[x]); \
DRAW_PIXEL(2, Pix = bp[2]); \ } \
DRAW_PIXEL(3, Pix = bp[3]); \
DRAW_PIXEL(4, Pix = bp[4]); \
DRAW_PIXEL(5, Pix = bp[5]); \
DRAW_PIXEL(6, Pix = bp[6]); \
DRAW_PIXEL(7, Pix = bp[7]); \
} \ } \
} \ } \
else \ else \
@ -628,14 +620,9 @@ void S9xSelectTileConverter (int depth, bool8 hires, bool8 sub, bool8 mosaic)
OFFSET_IN_LINE; \ OFFSET_IN_LINE; \
for (l = LineCount; l > 0; l--, bp -= 8 * PITCH, Offset += GFX.PPL) \ for (l = LineCount; l > 0; l--, bp -= 8 * PITCH, Offset += GFX.PPL) \
{ \ { \
DRAW_PIXEL(0, Pix = bp[7]); \ for (int x = 0; x < 8; x++) { \
DRAW_PIXEL(1, Pix = bp[6]); \ DRAW_PIXEL(x, Pix = bp[7 - x]); \
DRAW_PIXEL(2, Pix = bp[5]); \ } \
DRAW_PIXEL(3, Pix = bp[4]); \
DRAW_PIXEL(4, Pix = bp[3]); \
DRAW_PIXEL(5, Pix = bp[2]); \
DRAW_PIXEL(6, Pix = bp[1]); \
DRAW_PIXEL(7, Pix = bp[0]); \
} \ } \
} }
@ -1307,6 +1294,13 @@ static void MAKENAME(NAME1, Add_, NAME2) (ARGS)
#undef MATH #undef MATH
} }
static void MAKENAME(NAME1, Add_Brightness_, NAME2) (ARGS)
{
#define MATH(A, B, C) REGMATH(ADD_BRIGHTNESS, A, B, C)
DRAW_TILE();
#undef MATH
}
static void MAKENAME(NAME1, AddF1_2_, NAME2) (ARGS) static void MAKENAME(NAME1, AddF1_2_, NAME2) (ARGS)
{ {
#define MATH(A, B, C) MATHF1_2(ADD, A, B, C) #define MATH(A, B, C) MATHF1_2(ADD, A, B, C)
@ -1321,6 +1315,13 @@ static void MAKENAME(NAME1, AddS1_2_, NAME2) (ARGS)
#undef MATH #undef MATH
} }
static void MAKENAME(NAME1, AddS1_2_Brightness_, NAME2) (ARGS)
{
#define MATH(A, B, C) MATHS1_2(ADD_BRIGHTNESS, A, B, C)
DRAW_TILE();
#undef MATH
}
static void MAKENAME(NAME1, Sub_, NAME2) (ARGS) static void MAKENAME(NAME1, Sub_, NAME2) (ARGS)
{ {
#define MATH(A, B, C) REGMATH(SUB, A, B, C) #define MATH(A, B, C) REGMATH(SUB, A, B, C)
@ -1342,7 +1343,7 @@ static void MAKENAME(NAME1, SubS1_2_, NAME2) (ARGS)
#undef MATH #undef MATH
} }
static void (*MAKENAME(Renderers_, NAME1, NAME2)[7]) (ARGS) = static void (*MAKENAME(Renderers_, NAME1, NAME2)[9]) (ARGS) =
{ {
MAKENAME(NAME1, _, NAME2), MAKENAME(NAME1, _, NAME2),
MAKENAME(NAME1, Add_, NAME2), MAKENAME(NAME1, Add_, NAME2),
@ -1350,7 +1351,9 @@ static void (*MAKENAME(Renderers_, NAME1, NAME2)[7]) (ARGS) =
MAKENAME(NAME1, AddS1_2_, NAME2), MAKENAME(NAME1, AddS1_2_, NAME2),
MAKENAME(NAME1, Sub_, NAME2), MAKENAME(NAME1, Sub_, NAME2),
MAKENAME(NAME1, SubF1_2_, NAME2), MAKENAME(NAME1, SubF1_2_, NAME2),
MAKENAME(NAME1, SubS1_2_, NAME2) MAKENAME(NAME1, SubS1_2_, NAME2),
MAKENAME(NAME1, Add_Brightness_, NAME2),
MAKENAME(NAME1, AddS1_2_Brightness_, NAME2)
}; };
#undef MAKENAME #undef MAKENAME