u32 span_pixels = 0;
u32 span_pixel_blocks = 0;
-u32 span_pixel_blocks_unaligned = 0;
u32 spans = 0;
u32 triangles = 0;
u32 sprites = 0;
u32 texel_blocks_16bpp = 0;
u32 texel_blocks_untextured = 0;
u32 blend_blocks = 0;
-u32 untextured_pixels = 0;
-u32 blend_pixels = 0;
-u32 transparent_pixels = 0;
u32 render_buffer_flushes = 0;
u32 state_changes = 0;
u32 left_split_triangles = 0;
u32 clipped_triangles = 0;
u32 zero_block_spans = 0;
u32 texture_cache_loads = 0;
-u32 false_modulated_triangles = 0;
-u32 false_modulated_sprites = 0;
+u32 false_modulated_blocks = 0;
-u32 reciprocal_table[512];
+/* double size for enhancement */
+u32 reciprocal_table[512 * 2];
typedef s32 fixed_type;
{
u32 mask = texture_region_mask(x1, y1, x2, y2) &
psx_gpu->viewport_mask;
+
psx_gpu->dirty_textures_4bpp_mask |= mask;
psx_gpu->dirty_textures_8bpp_mask |= mask;
psx_gpu->dirty_textures_8bpp_alternate_mask |= mask;
return mask;
}
+void update_texture_cache_region(psx_gpu_struct *psx_gpu, u32 x1, u32 y1,
+ u32 x2, u32 y2)
+{
+ u32 mask = texture_region_mask(x1, y1, x2, y2);
+ u32 texture_page;
+ u8 *texture_page_ptr;
+ u16 *vram_ptr;
+ u32 texel_block;
+ u32 sub_x, sub_y;
+
+ psx_gpu->dirty_textures_8bpp_mask |= mask;
+ psx_gpu->dirty_textures_8bpp_alternate_mask |= mask;
+
+ if ((psx_gpu->dirty_textures_4bpp_mask & mask) == 0 &&
+ (x1 & 3) == 0 && (y1 & 15) == 0 && x2 - x1 < 4 && y2 - y1 < 16)
+ {
+ texture_page = ((x1 / 64) & 15) + (y1 / 256) * 16;
+ texture_page_ptr = psx_gpu->texture_4bpp_cache[texture_page];
+ texture_page_ptr += (x1 / 4 & 15) * 16*16 + (y1 / 16 & 15) * 16*16*16;
+ vram_ptr = psx_gpu->vram_ptr + x1 + y1 * 1024;
+ sub_x = 4;
+ sub_y = 16;
+
+ while(sub_y)
+ {
+ while(sub_x)
+ {
+ texel_block = *vram_ptr;
+
+ texture_page_ptr[0] = texel_block & 0xF;
+ texture_page_ptr[1] = (texel_block >> 4) & 0xF;
+ texture_page_ptr[2] = (texel_block >> 8) & 0xF;
+ texture_page_ptr[3] = texel_block >> 12;
+
+ vram_ptr++;
+ texture_page_ptr += 4;
+
+ sub_x--;
+ }
+
+ vram_ptr -= 4;
+ sub_x = 4;
+
+ sub_y--;
+ vram_ptr += 1024;
+ }
+ }
+ else
+ {
+ psx_gpu->dirty_textures_4bpp_mask |= mask;
+ }
+}
void update_texture_8bpp_cache_slice(psx_gpu_struct *psx_gpu,
u32 texture_page);
void update_texture_4bpp_cache(psx_gpu_struct *psx_gpu)
{
u32 current_texture_page = psx_gpu->current_texture_page;
- u8 *texture_page_ptr = psx_gpu->texture_page_ptr;
+ u8 *texture_page_ptr = psx_gpu->texture_page_base;
u16 *vram_ptr = psx_gpu->vram_ptr;
u32 texel_block;
while(sub_x)
{
texel_block = *vram_ptr;
+
texture_page_ptr[0] = texel_block & 0xF;
texture_page_ptr[1] = (texel_block >> 4) & 0xF;
texture_page_ptr[2] = (texel_block >> 8) & 0xF;
void update_texture_8bpp_cache_slice(psx_gpu_struct *psx_gpu,
u32 texture_page)
{
- u16 *texture_page_ptr = psx_gpu->texture_page_ptr;
+ u16 *texture_page_ptr = psx_gpu->texture_page_base;
u16 *vram_ptr = psx_gpu->vram_ptr;
u32 tile_x, tile_y;
void flush_render_block_buffer(psx_gpu_struct *psx_gpu)
{
+ if((psx_gpu->render_mode & RENDER_INTERLACE_ENABLED) &&
+ (psx_gpu->primitive_type == PRIMITIVE_TYPE_SPRITE))
+ {
+ u32 num_blocks_dest = 0;
+ block_struct *block_src = psx_gpu->blocks;
+ block_struct *block_dest = psx_gpu->blocks;
+
+ u16 *vram_ptr = psx_gpu->vram_ptr;
+ u32 i;
+
+ if(psx_gpu->render_mode & RENDER_INTERLACE_ODD)
+ {
+ for(i = 0; i < psx_gpu->num_blocks; i++)
+ {
+ u32 fb_offset = (u32)((u8 *)block_src->fb_ptr - (u8 *)vram_ptr);
+ if(fb_offset & (1 << 11))
+ {
+ *block_dest = *block_src;
+ num_blocks_dest++;
+ block_dest++;
+ }
+ block_src++;
+ }
+ }
+ else
+ {
+ for(i = 0; i < psx_gpu->num_blocks; i++)
+ {
+ u32 fb_offset = (u32)((u8 *)block_src->fb_ptr - (u8 *)vram_ptr);
+ if((fb_offset & (1 << 11)) == 0)
+ {
+ *block_dest = *block_src;
+ num_blocks_dest++;
+ block_dest++;
+ }
+ block_src++;
+ }
+ }
+
+ psx_gpu->num_blocks = num_blocks_dest;
+ }
+
if(psx_gpu->num_blocks)
{
render_block_handler_struct *render_block_handler =
render_block_handler->shade_blocks(psx_gpu);
render_block_handler->blend_blocks(psx_gpu);
+#ifdef PROFILE
span_pixel_blocks += psx_gpu->num_blocks;
render_buffer_flushes++;
+#endif
psx_gpu->num_blocks = 0;
}
vec_4x32u uvrg_base;
vec_4x32u b_base;
- vec_4x32u const_0x8000;
+ vec_4x32u uvrgb_phase;
vec_4x16s d0_a_d3_c, d0_b, d0_c;
vec_4x16s d1_a, d1_b, d1_c_d2_a;
setup_gradient_calculation_input(1, b);
setup_gradient_calculation_input(2, c);
- dup_4x32b(const_0x8000, 0x8000);
+ dup_4x32b(uvrgb_phase, psx_gpu->uvrgb_phase);
shl_long_4x16b(uvrg_base, x0_a_y0_c, 16);
shl_long_4x16b(b_base, x0_b, 16);
- add_4x32b(uvrg_base, uvrg_base, const_0x8000);
- add_4x32b(b_base, b_base, const_0x8000);
+ add_4x32b(uvrg_base, uvrg_base, uvrgb_phase);
+ add_4x32b(b_base, b_base, uvrgb_phase);
// Can probably pair these, but it'll require careful register allocation
sub_4x16b(d0_a_d3_c, x1_a_y1_c, x0_a_y0_c);
\
dup_2x32b(edge_shifts, edge_shift); \
sub_2x32b(heights_b, heights, c_0x01); \
- shr_2x32b(height_reciprocals, edge_shifts, 12); \
+ shr_2x32b(height_reciprocals, edge_shifts, 10); \
\
mla_2x32b(heights_b, x_starts, heights); \
bic_immediate_4x16b(vector_cast(vec_4x16u, edge_shifts), 0xE0); \
sub_2x32b(widths, x_ends, x_starts); \
width_alt = x_c - start_c; \
\
- shr_2x32b(height_reciprocals, edge_shifts, 12); \
- height_reciprocal_alt = edge_shift_alt >> 12; \
+ shr_2x32b(height_reciprocals, edge_shifts, 10); \
+ height_reciprocal_alt = edge_shift_alt >> 10; \
\
bic_immediate_4x16b(vector_cast(vec_4x16u, edge_shifts), 0xE0); \
edge_shift_alt &= 0x1F; \
} \
#define setup_blocks_add_blocks_direct() \
+ texel_blocks_untextured += span_num_blocks; \
+ span_pixel_blocks += span_num_blocks \
#define setup_blocks_builder(shading, texturing, dithering, sw, target) \
if(span_num_blocks) \
{ \
y = span_edge_data->y; \
- fb_ptr = psx_gpu->vram_ptr + span_edge_data->left_x + (y * 1024); \
+ fb_ptr = psx_gpu->vram_out_ptr + span_edge_data->left_x + (y * 1024); \
\
setup_blocks_span_initialize_##shading##_##texturing(); \
setup_blocks_span_initialize_##dithering(texturing); \
s32 pixel_span = span_num_blocks * 8; \
pixel_span -= __builtin_popcount(span_edge_data->right_mask & 0xFF); \
span_pixels += pixel_span; \
- span_pixel_blocks_unaligned += (pixel_span + 7) / 8; \
\
span_num_blocks--; \
while(span_num_blocks) \
} \
-#define shade_blocks_textured_modulated_shaded_primitive_load() \
+#define shade_blocks_textured_false_modulated_check_dithered(target) \
+ if(psx_gpu->triangle_color == 0x808080) \
+ { \
+ false_modulated_blocks += num_blocks; \
+ } \
+
+#define shade_blocks_textured_false_modulated_check_undithered(target) \
+ if(psx_gpu->triangle_color == 0x808080) \
+ { \
+ \
+ shade_blocks_textured_unmodulated_##target(psx_gpu); \
+ false_modulated_blocks += num_blocks; \
+ return; \
+ } \
+
-#define shade_blocks_textured_modulated_unshaded_primitive_load() \
+#define shade_blocks_textured_modulated_shaded_primitive_load(dithering, \
+ target) \
+
+#define shade_blocks_textured_modulated_unshaded_primitive_load(dithering, \
+ target) \
{ \
u32 color = psx_gpu->triangle_color; \
dup_8x8b(colors_r, color); \
dup_8x8b(colors_g, color >> 8); \
dup_8x8b(colors_b, color >> 16); \
- if(psx_gpu->triangle_color == 0x808080) \
- false_modulated_triangles++; \
+ shade_blocks_textured_false_modulated_check_##dithering(target); \
} \
#define shade_blocks_textured_modulated_shaded_block_load() \
\
dup_8x16b(d128_0x8000, 0x8000); \
\
- shade_blocks_textured_modulated_##shading##_primitive_load(); \
+ shade_blocks_textured_modulated_##shading##_primitive_load(dithering, \
+ target); \
\
while(num_blocks) \
{ \
void shade_blocks_unshaded_textured_modulated_undithered_indirect(psx_gpu_struct
*psx_gpu);
+void shade_blocks_textured_unmodulated_indirect(psx_gpu_struct *psx_gpu);
+void shade_blocks_textured_unmodulated_direct(psx_gpu_struct *psx_gpu);
+
#ifndef NEON_BUILD
shade_blocks_textured_modulated_builder(shaded, dithered, direct);
} \
} \
-void shade_blocks_textured_unmodulated_indirect(psx_gpu_struct *psx_gpu);
-void shade_blocks_textured_unmodulated_direct(psx_gpu_struct *psx_gpu);
+#define shade_blocks_textured_unmodulated_dithered_builder(target) \
+void shade_blocks_textured_unmodulated_dithered_##target(psx_gpu_struct \
+ *psx_gpu) \
+{ \
+ block_struct *block = psx_gpu->blocks; \
+ u32 num_blocks = psx_gpu->num_blocks; \
+ vec_8x16u draw_mask; \
+ vec_8x16u test_mask = psx_gpu->test_mask; \
+ u32 draw_mask_bits; \
+ \
+ vec_8x16u pixels; \
+ shade_blocks_load_msb_mask_##target(); \
+ \
+ while(num_blocks) \
+ { \
+ vec_8x16u zero_mask; \
+ \
+ draw_mask_bits = block->draw_mask_bits; \
+ dup_8x16b(draw_mask, draw_mask_bits); \
+ tst_8x16b(draw_mask, draw_mask, test_mask); \
+ \
+ pixels = block->texels; \
+ \
+ cmpeqz_8x16b(zero_mask, pixels); \
+ or_8x16b(zero_mask, draw_mask, zero_mask); \
+ \
+ shade_blocks_store_##target(zero_mask, pixels); \
+ \
+ num_blocks--; \
+ block++; \
+ } \
+} \
#ifndef NEON_BUILD
void shade_blocks_unshaded_untextured_direct(psx_gpu_struct *psx_gpu);
#ifndef NEON_BUILD
-
+
void shade_blocks_unshaded_untextured_indirect(psx_gpu_struct *psx_gpu)
{
}
(triangle_y_direction_##direction_c << 4) | \
(triangle_winding_##winding << 6)) \
-psx_gpu_struct __attribute__((aligned(64))) psx_gpu_alt;
-
-void render_triangle(psx_gpu_struct *psx_gpu, vertex_struct *vertexes,
- u32 flags)
+static int prepare_triangle(psx_gpu_struct *psx_gpu, vertex_struct *vertexes,
+ vertex_struct *vertexes_out[3])
{
s32 y_top, y_bottom;
s32 triangle_area;
triangle_area = triangle_signed_area_x2(a->x, a->y, b->x, b->y, c->x, c->y);
+#ifdef PROFILE
triangles++;
+#endif
if(triangle_area == 0)
{
+#ifdef PROFILE
trivial_rejects++;
- return;
+#endif
+ return 0;
}
if(b->y < a->y)
if((y_bottom - y_top) >= 512)
{
+#ifdef PROFILE
trivial_rejects++;
- return;
+#endif
+ return 0;
}
if(triangle_area < 0)
vertex_swap(a, b);
}
- if((c->x - a->x) >= 1024)
+ if((c->x - psx_gpu->offset_x) >= 1024 || (c->x - a->x) >= 1024)
{
+#ifdef PROFILE
trivial_rejects++;
- return;
+#endif
+ return 0;
}
if(invalidate_texture_cache_region_viewport(psx_gpu, a->x, y_top, c->x,
y_bottom) == 0)
{
+#ifdef PROFILE
trivial_rejects++;
- return;
+#endif
+ return 0;
}
- psx_gpu->num_spans = 0;
psx_gpu->triangle_area = triangle_area;
psx_gpu->triangle_winding = triangle_winding;
+ vertexes_out[0] = a;
+ vertexes_out[1] = b;
+ vertexes_out[2] = c;
+
+ return 1;
+}
+
+static void render_triangle_p(psx_gpu_struct *psx_gpu,
+ vertex_struct *vertex_ptrs[3], u32 flags)
+{
+ psx_gpu->num_spans = 0;
+
+ vertex_struct *a = vertex_ptrs[0];
+ vertex_struct *b = vertex_ptrs[1];
+ vertex_struct *c = vertex_ptrs[2];
+
s32 y_delta_a = b->y - a->y;
s32 y_delta_b = c->y - b->y;
s32 y_delta_c = c->y - a->y;
compute_all_gradients(psx_gpu, a, b, c);
switch(y_direction_a | (y_direction_b << 2) | (y_direction_c << 4) |
- (triangle_winding << 6))
+ (psx_gpu->triangle_winding << 6))
{
triangle_case(up, up, up, negative):
triangle_case(up, up, flat, negative):
break;
}
+#ifdef PROFILE
spans += psx_gpu->num_spans;
+#endif
+
+ if(unlikely(psx_gpu->render_mode & RENDER_INTERLACE_ENABLED))
+ {
+ u32 i;
+
+ if(psx_gpu->render_mode & RENDER_INTERLACE_ODD)
+ {
+ for(i = 0; i < psx_gpu->num_spans; i++)
+ {
+ if((psx_gpu->span_edge_data[i].y & 1) == 0)
+ psx_gpu->span_edge_data[i].num_blocks = 0;
+ }
+ }
+ else
+ {
+ for(i = 0; i < psx_gpu->num_spans; i++)
+ {
+ if(psx_gpu->span_edge_data[i].y & 1)
+ psx_gpu->span_edge_data[i].num_blocks = 0;
+ }
+ }
+ }
u32 render_state = flags &
(RENDER_FLAGS_MODULATE_TEXELS | RENDER_FLAGS_BLEND |
{
psx_gpu->render_state = render_state;
flush_render_block_buffer(psx_gpu);
+#ifdef PROFILE
state_changes++;
+#endif
}
psx_gpu->primitive_type = PRIMITIVE_TYPE_TRIANGLE;
(psx_gpu);
}
+void render_triangle(psx_gpu_struct *psx_gpu, vertex_struct *vertexes,
+ u32 flags)
+{
+ vertex_struct *vertex_ptrs[3];
+ if (prepare_triangle(psx_gpu, vertexes, vertex_ptrs))
+ render_triangle_p(psx_gpu, vertex_ptrs, flags);
+}
+
void texture_sprite_blocks_8bpp(psx_gpu_struct *psx_gpu);
#endif
-#define setup_sprite_tiled_initialize_4bpp() \
+#define setup_sprite_tiled_initialize_4bpp_clut() \
u16 *clut_ptr = psx_gpu->clut_ptr; \
vec_8x16u clut_a, clut_b; \
vec_16x8u clut_low, clut_high; \
\
load_8x16b(clut_a, clut_ptr); \
load_8x16b(clut_b, clut_ptr + 8); \
- unzip_16x8b(clut_low, clut_high, clut_a, clut_b); \
+ unzip_16x8b(clut_low, clut_high, clut_a, clut_b) \
+
+#define setup_sprite_tiled_initialize_4bpp() \
+ setup_sprite_tiled_initialize_4bpp_clut(); \
\
if(psx_gpu->current_texture_mask & psx_gpu->dirty_textures_4bpp_mask) \
update_texture_4bpp_cache(psx_gpu) \
load_64b(texels, texture_block_ptr) \
-#define setup_sprite_tile_setup_block_yes(side, offset, texture_mode) \
-
-#define setup_sprite_tile_setup_block_no(side, offset, texture_mode) \
-
#define setup_sprite_tile_add_blocks(tile_num_blocks) \
num_blocks += tile_num_blocks; \
sprite_blocks += tile_num_blocks; \
#define setup_sprite_tile_column_edge_post_adjust_full(edge) \
-#define setup_sprite_tile_column_height_single(edge_mode, edge, texture_mode) \
+#define setup_sprite_tile_column_height_single(edge_mode, edge, texture_mode, \
+ x4mode) \
do \
{ \
sub_tile_height = column_data; \
- setup_sprite_tile_column_edge_pre_adjust_##edge_mode(edge); \
- setup_sprite_tile_##edge_mode##_##texture_mode(edge); \
- setup_sprite_tile_column_edge_post_adjust_##edge_mode(edge); \
+ setup_sprite_tile_column_edge_pre_adjust_##edge_mode##x4mode(edge); \
+ setup_sprite_tile_##edge_mode##_##texture_mode##x4mode(edge); \
+ setup_sprite_tile_column_edge_post_adjust_##edge_mode##x4mode(edge); \
} while(0) \
-#define setup_sprite_tile_column_height_multi(edge_mode, edge, texture_mode) \
+#define setup_sprite_tile_column_height_multi(edge_mode, edge, texture_mode, \
+ x4mode) \
do \
{ \
u32 tiles_remaining = column_data >> 16; \
sub_tile_height = column_data & 0xFF; \
- setup_sprite_tile_column_edge_pre_adjust_##edge_mode(edge); \
- setup_sprite_tile_##edge_mode##_##texture_mode(edge); \
+ setup_sprite_tile_column_edge_pre_adjust_##edge_mode##x4mode(edge); \
+ setup_sprite_tile_##edge_mode##_##texture_mode##x4mode(edge); \
tiles_remaining -= 1; \
\
while(tiles_remaining) \
{ \
sub_tile_height = 16; \
- setup_sprite_tile_##edge_mode##_##texture_mode(edge); \
+ setup_sprite_tile_##edge_mode##_##texture_mode##x4mode(edge); \
tiles_remaining--; \
} \
\
sub_tile_height = (column_data >> 8) & 0xFF; \
- setup_sprite_tile_##edge_mode##_##texture_mode(edge); \
- setup_sprite_tile_column_edge_post_adjust_##edge_mode(edge); \
+ setup_sprite_tile_##edge_mode##_##texture_mode##x4mode(edge); \
+ setup_sprite_tile_column_edge_post_adjust_##edge_mode##x4mode(edge); \
} while(0) \
column_data |= (tile_height - 1) << 16 \
+#define RIGHT_MASK_BIT_SHIFT 8
+#define RIGHT_MASK_BIT_SHIFT_4x 16
+
#define setup_sprite_tile_column_width_single(texture_mode, multi_height, \
- edge_mode, edge) \
+ edge_mode, edge, x4mode) \
{ \
setup_sprite_column_data_##multi_height(); \
left_mask_bits = left_block_mask | right_block_mask; \
- right_mask_bits = left_mask_bits >> 8; \
+ right_mask_bits = left_mask_bits >> RIGHT_MASK_BIT_SHIFT##x4mode; \
\
setup_sprite_tile_column_height_##multi_height(edge_mode, edge, \
- texture_mode); \
+ texture_mode, x4mode); \
} \
#define setup_sprite_tiled_advance_column() \
if((texture_offset_base & 0xF00) == 0) \
texture_offset_base -= (0x100 + 0xF00) \
+#define FB_PTR_MULTIPLIER 1
+#define FB_PTR_MULTIPLIER_4x 2
+
#define setup_sprite_tile_column_width_multi(texture_mode, multi_height, \
- left_mode, right_mode) \
+ left_mode, right_mode, x4mode) \
{ \
setup_sprite_column_data_##multi_height(); \
- s32 fb_ptr_advance_column = 16 - (1024 * height); \
+ s32 fb_ptr_advance_column = (16 - (1024 * height)) \
+ * FB_PTR_MULTIPLIER##x4mode; \
\
tile_width -= 2; \
left_mask_bits = left_block_mask; \
- right_mask_bits = left_mask_bits >> 8; \
+ right_mask_bits = left_mask_bits >> RIGHT_MASK_BIT_SHIFT##x4mode; \
\
setup_sprite_tile_column_height_##multi_height(left_mode, right, \
- texture_mode); \
+ texture_mode, x4mode); \
fb_ptr += fb_ptr_advance_column; \
\
left_mask_bits = 0x00; \
while(tile_width) \
{ \
setup_sprite_tiled_advance_column(); \
- setup_sprite_tile_column_height_##multi_height(full, none, texture_mode); \
+ setup_sprite_tile_column_height_##multi_height(full, none, \
+ texture_mode, x4mode); \
fb_ptr += fb_ptr_advance_column; \
tile_width--; \
} \
\
left_mask_bits = right_block_mask; \
- right_mask_bits = left_mask_bits >> 8; \
+ right_mask_bits = left_mask_bits >> RIGHT_MASK_BIT_SHIFT##x4mode; \
\
setup_sprite_tiled_advance_column(); \
setup_sprite_tile_column_height_##multi_height(right_mode, left, \
- texture_mode); \
+ texture_mode, x4mode); \
} \
-#define setup_sprite_tiled_builder(texture_mode) \
-void setup_sprite_##texture_mode(psx_gpu_struct *psx_gpu, s32 x, s32 y, \
+/* 4x stuff */
+#define setup_sprite_tiled_initialize_4bpp_4x() \
+ setup_sprite_tiled_initialize_4bpp_clut() \
+
+#define setup_sprite_tiled_initialize_8bpp_4x() \
+
+
+#define setup_sprite_tile_full_4bpp_4x(edge) \
+{ \
+ vec_8x8u texels_low, texels_high; \
+ vec_8x16u pixels, pixels_wide; \
+ setup_sprite_tile_add_blocks(sub_tile_height * 2 * 4); \
+ u32 left_mask_bits_a = left_mask_bits & 0xFF; \
+ u32 left_mask_bits_b = left_mask_bits >> 8; \
+ u32 right_mask_bits_a = right_mask_bits & 0xFF; \
+ u32 right_mask_bits_b = right_mask_bits >> 8; \
+ \
+ while(sub_tile_height) \
+ { \
+ setup_sprite_tile_fetch_texel_block_8bpp(0); \
+ tbl_16(texels_low, texels, clut_low); \
+ tbl_16(texels_high, texels, clut_high); \
+ zip_8x16b(pixels, texels_low, texels_high); \
+ \
+ zip_4x32b(vector_cast(vec_4x32u, pixels_wide), pixels.low, pixels.low); \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = left_mask_bits_a; \
+ block->fb_ptr = fb_ptr; \
+ block++; \
+ \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = left_mask_bits_a; \
+ block->fb_ptr = fb_ptr + 1024; \
+ block++; \
+ \
+ zip_4x32b(vector_cast(vec_4x32u, pixels_wide), pixels.high, pixels.high); \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = left_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 8; \
+ block++; \
+ \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = left_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 1024 + 8; \
+ block++; \
+ \
+ setup_sprite_tile_fetch_texel_block_8bpp(8); \
+ tbl_16(texels_low, texels, clut_low); \
+ tbl_16(texels_high, texels, clut_high); \
+ zip_8x16b(pixels, texels_low, texels_high); \
+ \
+ zip_4x32b(vector_cast(vec_4x32u, pixels_wide), pixels.low, pixels.low); \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = right_mask_bits_a; \
+ block->fb_ptr = fb_ptr + 16; \
+ block++; \
+ \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = right_mask_bits_a; \
+ block->fb_ptr = fb_ptr + 1024 + 16; \
+ block++; \
+ \
+ zip_4x32b(vector_cast(vec_4x32u, pixels_wide), pixels.high, pixels.high); \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = right_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 24; \
+ block++; \
+ \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = right_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 1024 + 24; \
+ block++; \
+ \
+ fb_ptr += 2048; \
+ texture_offset += 0x10; \
+ sub_tile_height--; \
+ } \
+ texture_offset += 0xF00; \
+ psx_gpu->num_blocks = num_blocks; \
+} \
+
+#define setup_sprite_tile_half_4bpp_4x(edge) \
+{ \
+ vec_8x8u texels_low, texels_high; \
+ vec_8x16u pixels, pixels_wide; \
+ setup_sprite_tile_add_blocks(sub_tile_height * 4); \
+ u32 edge##_mask_bits_a = edge##_mask_bits & 0xFF; \
+ u32 edge##_mask_bits_b = edge##_mask_bits >> 8; \
+ \
+ while(sub_tile_height) \
+ { \
+ setup_sprite_tile_fetch_texel_block_8bpp(0); \
+ tbl_16(texels_low, texels, clut_low); \
+ tbl_16(texels_high, texels, clut_high); \
+ zip_8x16b(pixels, texels_low, texels_high); \
+ \
+ zip_4x32b(vector_cast(vec_4x32u, pixels_wide), pixels.low, pixels.low); \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = edge##_mask_bits_a; \
+ block->fb_ptr = fb_ptr; \
+ block++; \
+ \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = edge##_mask_bits_a; \
+ block->fb_ptr = fb_ptr + 1024; \
+ block++; \
+ \
+ zip_4x32b(vector_cast(vec_4x32u, pixels_wide), pixels.high, pixels.high); \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = edge##_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 8; \
+ block++; \
+ \
+ block->texels = pixels_wide; \
+ block->draw_mask_bits = edge##_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 1024 + 8; \
+ block++; \
+ \
+ fb_ptr += 2048; \
+ texture_offset += 0x10; \
+ sub_tile_height--; \
+ } \
+ texture_offset += 0xF00; \
+ psx_gpu->num_blocks = num_blocks; \
+} \
+
+
+#define setup_sprite_tile_full_8bpp_4x(edge) \
+{ \
+ setup_sprite_tile_add_blocks(sub_tile_height * 2 * 4); \
+ vec_16x8u texels_wide; \
+ u32 left_mask_bits_a = left_mask_bits & 0xFF; \
+ u32 left_mask_bits_b = left_mask_bits >> 8; \
+ u32 right_mask_bits_a = right_mask_bits & 0xFF; \
+ u32 right_mask_bits_b = right_mask_bits >> 8; \
+ \
+ while(sub_tile_height) \
+ { \
+ setup_sprite_tile_fetch_texel_block_8bpp(0); \
+ zip_8x16b(vector_cast(vec_8x16u, texels_wide), texels, texels); \
+ block->r = texels_wide.low; \
+ block->draw_mask_bits = left_mask_bits_a; \
+ block->fb_ptr = fb_ptr; \
+ block++; \
+ \
+ block->r = texels_wide.low; \
+ block->draw_mask_bits = left_mask_bits_a; \
+ block->fb_ptr = fb_ptr + 1024; \
+ block++; \
+ \
+ block->r = texels_wide.high; \
+ block->draw_mask_bits = left_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 8; \
+ block++; \
+ \
+ block->r = texels_wide.high; \
+ block->draw_mask_bits = left_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 1024 + 8; \
+ block++; \
+ \
+ setup_sprite_tile_fetch_texel_block_8bpp(8); \
+ zip_8x16b(vector_cast(vec_8x16u, texels_wide), texels, texels); \
+ block->r = texels_wide.low; \
+ block->draw_mask_bits = right_mask_bits_a; \
+ block->fb_ptr = fb_ptr + 16; \
+ block++; \
+ \
+ block->r = texels_wide.low; \
+ block->draw_mask_bits = right_mask_bits_a; \
+ block->fb_ptr = fb_ptr + 1024 + 16; \
+ block++; \
+ \
+ block->r = texels_wide.high; \
+ block->draw_mask_bits = right_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 24; \
+ block++; \
+ \
+ block->r = texels_wide.high; \
+ block->draw_mask_bits = right_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 24 + 1024; \
+ block++; \
+ \
+ fb_ptr += 2048; \
+ texture_offset += 0x10; \
+ sub_tile_height--; \
+ } \
+ texture_offset += 0xF00; \
+ psx_gpu->num_blocks = num_blocks; \
+} \
+
+#define setup_sprite_tile_half_8bpp_4x(edge) \
+{ \
+ setup_sprite_tile_add_blocks(sub_tile_height * 4); \
+ vec_16x8u texels_wide; \
+ u32 edge##_mask_bits_a = edge##_mask_bits & 0xFF; \
+ u32 edge##_mask_bits_b = edge##_mask_bits >> 8; \
+ \
+ while(sub_tile_height) \
+ { \
+ setup_sprite_tile_fetch_texel_block_8bpp(0); \
+ zip_8x16b(vector_cast(vec_8x16u, texels_wide), texels, texels); \
+ block->r = texels_wide.low; \
+ block->draw_mask_bits = edge##_mask_bits_a; \
+ block->fb_ptr = fb_ptr; \
+ block++; \
+ \
+ block->r = texels_wide.low; \
+ block->draw_mask_bits = edge##_mask_bits_a; \
+ block->fb_ptr = fb_ptr + 1024; \
+ block++; \
+ \
+ block->r = texels_wide.high; \
+ block->draw_mask_bits = edge##_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 8; \
+ block++; \
+ \
+ block->r = texels_wide.high; \
+ block->draw_mask_bits = edge##_mask_bits_b; \
+ block->fb_ptr = fb_ptr + 8 + 1024; \
+ block++; \
+ \
+ fb_ptr += 2048; \
+ texture_offset += 0x10; \
+ sub_tile_height--; \
+ } \
+ texture_offset += 0xF00; \
+ psx_gpu->num_blocks = num_blocks; \
+} \
+
+
+#define setup_sprite_tile_column_edge_pre_adjust_half_right_4x() \
+ texture_offset = texture_offset_base + 8; \
+ fb_ptr += 16 \
+
+#define setup_sprite_tile_column_edge_pre_adjust_half_left_4x() \
+ texture_offset = texture_offset_base \
+
+#define setup_sprite_tile_column_edge_pre_adjust_half_4x(edge) \
+ setup_sprite_tile_column_edge_pre_adjust_half_##edge##_4x() \
+
+#define setup_sprite_tile_column_edge_pre_adjust_full_4x(edge) \
+ texture_offset = texture_offset_base \
+
+#define setup_sprite_tile_column_edge_post_adjust_half_right_4x() \
+ fb_ptr -= 16 \
+
+#define setup_sprite_tile_column_edge_post_adjust_half_left_4x() \
+
+#define setup_sprite_tile_column_edge_post_adjust_half_4x(edge) \
+ setup_sprite_tile_column_edge_post_adjust_half_##edge##_4x() \
+
+#define setup_sprite_tile_column_edge_post_adjust_full_4x(edge) \
+
+
+#define setup_sprite_offset_u_adjust() \
+
+#define setup_sprite_comapre_left_block_mask() \
+ ((left_block_mask & 0xFF) == 0xFF) \
+
+#define setup_sprite_comapre_right_block_mask() \
+ (((right_block_mask >> 8) & 0xFF) == 0xFF) \
+
+
+#define setup_sprite_offset_u_adjust_4x() \
+ offset_u *= 2; \
+ offset_u_right = offset_u_right * 2 + 1 \
+
+#define setup_sprite_comapre_left_block_mask_4x() \
+ ((left_block_mask & 0xFFFF) == 0xFFFF) \
+
+#define setup_sprite_comapre_right_block_mask_4x() \
+ (((right_block_mask >> 16) & 0xFFFF) == 0xFFFF) \
+
+
+#define setup_sprite_tiled_builder(texture_mode, x4mode) \
+void setup_sprite_##texture_mode##x4mode(psx_gpu_struct *psx_gpu, s32 x, s32 y,\
s32 u, s32 v, s32 width, s32 height, u32 color) \
{ \
s32 offset_u = u & 0xF; \
s32 tile_width = width_rounded / 16; \
u32 offset_u_right = width_rounded & 0xF; \
\
- u32 left_block_mask = ~(0xFFFF << offset_u); \
- u32 right_block_mask = 0xFFFE << offset_u_right; \
+ setup_sprite_offset_u_adjust##x4mode(); \
+ \
+ u32 left_block_mask = ~(0xFFFFFFFF << offset_u); \
+ u32 right_block_mask = 0xFFFFFFFE << offset_u_right; \
\
u32 left_mask_bits; \
u32 right_mask_bits; \
u32 texture_offset_base = texture_offset; \
u32 control_mask; \
\
- u16 *fb_ptr = psx_gpu->vram_ptr + (y * 1024) + (x - offset_u); \
+ u16 *fb_ptr = psx_gpu->vram_out_ptr + (y * 1024) + (x - offset_u); \
u32 num_blocks = psx_gpu->num_blocks; \
block_struct *block = psx_gpu->blocks + num_blocks; \
\
u16 *texture_block_ptr; \
vec_8x8u texels; \
\
- setup_sprite_tiled_initialize_##texture_mode(); \
+ setup_sprite_tiled_initialize_##texture_mode##x4mode(); \
\
control_mask = tile_width == 1; \
control_mask |= (tile_height == 1) << 1; \
- control_mask |= ((left_block_mask & 0xFF) == 0xFF) << 2; \
- control_mask |= (((right_block_mask >> 8) & 0xFF) == 0xFF) << 3; \
+ control_mask |= setup_sprite_comapre_left_block_mask##x4mode() << 2; \
+ control_mask |= setup_sprite_comapre_right_block_mask##x4mode() << 3; \
\
sprites_##texture_mode++; \
\
{ \
default: \
case 0x0: \
- setup_sprite_tile_column_width_multi(texture_mode, multi, full, full); \
+ setup_sprite_tile_column_width_multi(texture_mode, multi, full, full, \
+ x4mode); \
break; \
\
case 0x1: \
- setup_sprite_tile_column_width_single(texture_mode, multi, full, none); \
+ setup_sprite_tile_column_width_single(texture_mode, multi, full, none, \
+ x4mode); \
break; \
\
case 0x2: \
- setup_sprite_tile_column_width_multi(texture_mode, single, full, full); \
+ setup_sprite_tile_column_width_multi(texture_mode, single, full, full, \
+ x4mode); \
break; \
\
case 0x3: \
- setup_sprite_tile_column_width_single(texture_mode, single, full, none); \
+ setup_sprite_tile_column_width_single(texture_mode, single, full, none, \
+ x4mode); \
break; \
\
case 0x4: \
- setup_sprite_tile_column_width_multi(texture_mode, multi, half, full); \
+ setup_sprite_tile_column_width_multi(texture_mode, multi, half, full, \
+ x4mode); \
break; \
\
case 0x5: \
- setup_sprite_tile_column_width_single(texture_mode, multi, half, right); \
+ setup_sprite_tile_column_width_single(texture_mode, multi, half, right, \
+ x4mode); \
break; \
\
case 0x6: \
- setup_sprite_tile_column_width_multi(texture_mode, single, half, full); \
+ setup_sprite_tile_column_width_multi(texture_mode, single, half, full, \
+ x4mode); \
break; \
\
case 0x7: \
- setup_sprite_tile_column_width_single(texture_mode, single, half, right);\
+ setup_sprite_tile_column_width_single(texture_mode, single, half, right, \
+ x4mode); \
break; \
\
case 0x8: \
- setup_sprite_tile_column_width_multi(texture_mode, multi, full, half); \
+ setup_sprite_tile_column_width_multi(texture_mode, multi, full, half, \
+ x4mode); \
break; \
\
case 0x9: \
- setup_sprite_tile_column_width_single(texture_mode, multi, half, left); \
+ setup_sprite_tile_column_width_single(texture_mode, multi, half, left, \
+ x4mode); \
break; \
\
case 0xA: \
- setup_sprite_tile_column_width_multi(texture_mode, single, full, half); \
+ setup_sprite_tile_column_width_multi(texture_mode, single, full, half, \
+ x4mode); \
break; \
\
case 0xB: \
- setup_sprite_tile_column_width_single(texture_mode, single, half, left); \
+ setup_sprite_tile_column_width_single(texture_mode, single, half, left, \
+ x4mode); \
break; \
\
case 0xC: \
- setup_sprite_tile_column_width_multi(texture_mode, multi, half, half); \
+ setup_sprite_tile_column_width_multi(texture_mode, multi, half, half, \
+ x4mode); \
break; \
\
case 0xE: \
- setup_sprite_tile_column_width_multi(texture_mode, single, half, half); \
+ setup_sprite_tile_column_width_multi(texture_mode, single, half, half, \
+ x4mode); \
break; \
} \
} \
-
void setup_sprite_4bpp(psx_gpu_struct *psx_gpu, s32 x, s32 y, s32 u, s32 v,
s32 width, s32 height, u32 color);
void setup_sprite_8bpp(psx_gpu_struct *psx_gpu, s32 x, s32 y, s32 u, s32 v,
void setup_sprite_16bpp(psx_gpu_struct *psx_gpu, s32 x, s32 y, s32 u, s32 v,
s32 width, s32 height, u32 color);
+void setup_sprite_4bpp_4x(psx_gpu_struct *psx_gpu, s32 x, s32 y, s32 u, s32 v,
+ s32 width, s32 height, u32 color);
+void setup_sprite_8bpp_4x(psx_gpu_struct *psx_gpu, s32 x, s32 y, s32 u, s32 v,
+ s32 width, s32 height, u32 color);
+void setup_sprite_16bpp_4x(psx_gpu_struct *psx_gpu, s32 x, s32 y, s32 u, s32 v,
+ s32 width, s32 height, u32 color);
+
#ifndef NEON_BUILD
-setup_sprite_tiled_builder(4bpp);
-setup_sprite_tiled_builder(8bpp);
+setup_sprite_tiled_builder(4bpp,);
+setup_sprite_tiled_builder(8bpp,);
+
+setup_sprite_tiled_builder(4bpp,_4x);
+setup_sprite_tiled_builder(8bpp,_4x);
void setup_sprite_16bpp(psx_gpu_struct *psx_gpu, s32 x, s32 y, s32 u,
s32 v, s32 width, s32 height, u32 color)
u32 left_offset = u & 0x7;
u32 width_rounded = width + left_offset + 7;
- u16 *fb_ptr = psx_gpu->vram_ptr + (y * 1024) + (x - left_offset);
+ u16 *fb_ptr = psx_gpu->vram_out_ptr + (y * 1024) + (s32)(x - left_offset);
u32 right_width = width_rounded & 0x7;
u32 block_width = width_rounded / 8;
u32 fb_ptr_pitch = (1024 + 8) - (block_width * 8);
{
u32 right_width = ((width - 1) & 0x7) + 1;
u32 right_mask_bits = (0xFF << right_width);
- u16 *fb_ptr = psx_gpu->vram_ptr + (y * 1024) + x;
+ u16 *fb_ptr = psx_gpu->vram_out_ptr + (y * 1024) + x;
u32 block_width = (width + 7) / 8;
u32 fb_ptr_pitch = 1024 - ((block_width - 1) * 8);
u32 blocks_remaining;
{
blocks_remaining = block_width - 1;
num_blocks += block_width;
+
+#ifdef PROFILE
sprite_blocks += block_width;
+#endif
if(num_blocks > MAX_BLOCKS)
{
s32 x_right = x + width - 1;
s32 y_bottom = y + height - 1;
+#ifdef PROFILE
+ sprites++;
+#endif
+
if(invalidate_texture_cache_region_viewport(psx_gpu, x, y, x_right,
y_bottom) == 0)
{
if((width <= 0) || (height <= 0))
return;
- sprites++;
-
+#ifdef PROFILE
span_pixels += width * height;
spans += height;
+#endif
u32 render_state = flags &
(RENDER_FLAGS_MODULATE_TEXELS | RENDER_FLAGS_BLEND |
RENDER_FLAGS_TEXTURE_MAP);
render_state |=
(psx_gpu->render_state_base & ~RENDER_STATE_DITHER);
-
+
if((psx_gpu->render_state != render_state) ||
(psx_gpu->primitive_type != PRIMITIVE_TYPE_SPRITE))
{
psx_gpu->render_state = render_state;
flush_render_block_buffer(psx_gpu);
+#ifdef PROFILE
state_changes++;
+#endif
}
psx_gpu->primitive_type = PRIMITIVE_TYPE_SPRITE;
#define set_line_gradients(minor) \
{ \
s32 gradient_divisor = delta_##minor; \
- gradient_r = int_to_fixed(vertex_b->r - vertex_a->r) / gradient_divisor; \
- gradient_g = int_to_fixed(vertex_b->g - vertex_a->g) / gradient_divisor; \
- gradient_b = int_to_fixed(vertex_b->b - vertex_a->b) / gradient_divisor; \
+ if(gradient_divisor != 0) \
+ { \
+ gradient_r = int_to_fixed(vertex_b->r - vertex_a->r) / gradient_divisor; \
+ gradient_g = int_to_fixed(vertex_b->g - vertex_a->g) / gradient_divisor; \
+ gradient_b = int_to_fixed(vertex_b->b - vertex_a->b) / gradient_divisor; \
+ } \
+ else \
+ { \
+ gradient_r = 0; \
+ gradient_g = 0; \
+ gradient_b = 0; \
+ } \
current_r = fixed_center(vertex_a->r); \
current_g = fixed_center(vertex_a->g); \
current_b = fixed_center(vertex_a->b); \
{ \
delta_y *= -1; \
\
- if(delta_y >= 512) \
- return; \
- \
if(delta_x > delta_y) \
{ \
draw_line_span_horizontal(decrement, shading, blending, dithering, \
} \
else \
{ \
- if(delta_y >= 512) \
- return; \
- \
if(delta_x > delta_y) \
{ \
draw_line_span_horizontal(increment, shading, blending, dithering, \
void render_line(psx_gpu_struct *psx_gpu, vertex_struct *vertexes, u32 flags,
- u32 color)
+ u32 color, int double_resolution)
{
s32 color_r, color_g, color_b;
u32 triangle_winding = 0;
u32 control_mask;
+#ifdef PROFILE
lines++;
+#endif
if(vertex_a->x >= vertex_b->x)
{
delta_x = x_b - x_a;
delta_y = y_b - y_a;
- if(delta_x >= 1024)
+ if(delta_x >= 1024 || delta_y >= 512 || delta_y <= -512)
return;
+ if(double_resolution)
+ {
+ x_a *= 2;
+ x_b *= 2;
+ y_a *= 2;
+ y_b *= 2;
+ delta_x *= 2;
+ delta_y *= 2;
+ }
+
flags &= ~RENDER_FLAGS_TEXTURE_MAP;
- vram_ptr = psx_gpu->vram_ptr + (y_a * 1024) + x_a;
+ vram_ptr = psx_gpu->vram_out_ptr + (y_a * 1024) + x_a;
control_mask = 0x0;
void render_block_fill(psx_gpu_struct *psx_gpu, u32 color, u32 x, u32 y,
u32 width, u32 height)
{
+ if((width == 0) || (height == 0))
+ return;
+
invalidate_texture_cache_region(psx_gpu, x, y, x + width - 1, y + height - 1);
-#ifndef NEON_BUILD
u32 r = color & 0xFF;
u32 g = (color >> 8) & 0xFF;
u32 b = (color >> 16) & 0xFF;
- u32 color_16bpp = (r >> 3) | ((g >> 3) << 5) | ((b >> 3) << 10);
+ u32 color_16bpp = (r >> 3) | ((g >> 3) << 5) | ((b >> 3) << 10) |
+ psx_gpu->mask_msb;
+ u32 color_32bpp = color_16bpp | (color_16bpp << 16);
- u16 *vram_ptr = psx_gpu->vram_ptr + x + (y * 1024);
- u32 draw_x, draw_y;
+ u32 *vram_ptr = (u32 *)(psx_gpu->vram_out_ptr + x + (y * 1024));
- for(draw_y = 0; draw_y < height; draw_y++)
+ u32 pitch = 512 - (width / 2);
+ u32 num_width;
+
+ if(psx_gpu->render_mode & RENDER_INTERLACE_ENABLED)
{
- for(draw_x = 0; draw_x < width; draw_x++)
+ pitch += 512;
+ height /= 2;
+
+ if(psx_gpu->render_mode & RENDER_INTERLACE_ODD)
+ vram_ptr += 512;
+ }
+
+ while(height)
+ {
+ num_width = width;
+ while(num_width)
{
- vram_ptr[draw_x] = color_16bpp;
+ vram_ptr[0] = color_32bpp;
+ vram_ptr[1] = color_32bpp;
+ vram_ptr[2] = color_32bpp;
+ vram_ptr[3] = color_32bpp;
+ vram_ptr[4] = color_32bpp;
+ vram_ptr[5] = color_32bpp;
+ vram_ptr[6] = color_32bpp;
+ vram_ptr[7] = color_32bpp;
+
+ vram_ptr += 8;
+ num_width -= 16;
}
- vram_ptr += 1024;
+ vram_ptr += pitch;
+ height--;
}
-#else
- void render_block_fill_body(psx_gpu_struct *psx_gpu, u32 color, u32 x, u32 y,
- u32 width, u32 height);
+}
- render_block_fill_body(psx_gpu, color, x, y, width, height);
-#endif
+void render_block_fill_enh(psx_gpu_struct *psx_gpu, u32 color, u32 x, u32 y,
+ u32 width, u32 height)
+{
+ if((width == 0) || (height == 0))
+ return;
+
+ if(width > 1024)
+ width = 1024;
+
+ u32 r = color & 0xFF;
+ u32 g = (color >> 8) & 0xFF;
+ u32 b = (color >> 16) & 0xFF;
+ u32 color_16bpp = (r >> 3) | ((g >> 3) << 5) | ((b >> 3) << 10) |
+ psx_gpu->mask_msb;
+ u32 color_32bpp = color_16bpp | (color_16bpp << 16);
+
+ u32 *vram_ptr = (u32 *)(psx_gpu->vram_out_ptr + x + (y * 1024));
+
+ u32 pitch = 1024 / 2 - (width / 2);
+ u32 num_width;
+
+ while(height)
+ {
+ num_width = width;
+ while(num_width)
+ {
+ vram_ptr[0] = color_32bpp;
+ vram_ptr[1] = color_32bpp;
+ vram_ptr[2] = color_32bpp;
+ vram_ptr[3] = color_32bpp;
+ vram_ptr[4] = color_32bpp;
+ vram_ptr[5] = color_32bpp;
+ vram_ptr[6] = color_32bpp;
+ vram_ptr[7] = color_32bpp;
+
+ vram_ptr += 8;
+ num_width -= 16;
+ }
+
+ vram_ptr += pitch;
+ height--;
+ }
}
void render_block_copy(psx_gpu_struct *psx_gpu, u16 *source, u32 x, u32 y,
{
u16 *vram_ptr = psx_gpu->vram_ptr + x + (y * 1024);
u32 draw_x, draw_y;
+ u32 mask_msb = psx_gpu->mask_msb;
- invalidate_texture_cache_region(psx_gpu, x, y, x + width - 1, y + height - 1);
+ if((width == 0) || (height == 0))
+ return;
- //printf("copy for %d, %d\n", width, height);
+ flush_render_block_buffer(psx_gpu);
+ invalidate_texture_cache_region(psx_gpu, x, y, x + width - 1, y + height - 1);
for(draw_y = 0; draw_y < height; draw_y++)
{
for(draw_x = 0; draw_x < width; draw_x++)
{
- vram_ptr[draw_x] = source[draw_x];
+ vram_ptr[draw_x] = source[draw_x] | mask_msb;
}
source += pitch;
u32 height_reciprocal;
s32 shift;
- for(height = 1; height < 512; height++)
+ for(height = 1; height < sizeof(reciprocal_table)
+ / sizeof(reciprocal_table[0]); height++)
{
shift = __builtin_clz(height);
height_normalized = height << shift;
- height_reciprocal = ((1ULL << 50) + (height_normalized - 1)) /
+ height_reciprocal = ((1ULL << 52) + (height_normalized - 1)) /
height_normalized;
- shift = 32 - (50 - shift);
+ shift = 32 - (52 - shift);
- reciprocal_table[height] = (height_reciprocal << 12) | shift;
+ reciprocal_table[height] = (height_reciprocal << 10) | shift;
}
}
psx_gpu->test_mask = test_mask;
- psx_gpu->pixel_count_mode = 0;
- psx_gpu->pixel_compare_mode = 0;
-
- psx_gpu->vram_pixel_counts_a = malloc(sizeof(u8) * 1024 * 512);
- psx_gpu->vram_pixel_counts_b = malloc(sizeof(u8) * 1024 * 512);
- memset(psx_gpu->vram_pixel_counts_a, 0, sizeof(u8) * 1024 * 512);
- memset(psx_gpu->vram_pixel_counts_b, 0, sizeof(u8) * 1024 * 512);
- psx_gpu->compare_vram = malloc(sizeof(u16) * 1024 * 512);
-
psx_gpu->dirty_textures_4bpp_mask = 0xFFFFFFFF;
psx_gpu->dirty_textures_8bpp_mask = 0xFFFFFFFF;
psx_gpu->dirty_textures_8bpp_alternate_mask = 0xFFFFFFFF;
psx_gpu->render_state = 0;
psx_gpu->render_state_base = 0;
psx_gpu->num_blocks = 0;
+ psx_gpu->uvrgb_phase = 0x8000;
psx_gpu->vram_ptr = vram;
+ psx_gpu->vram_out_ptr = vram;
+ psx_gpu->texture_page_base = psx_gpu->vram_ptr;
psx_gpu->texture_page_ptr = psx_gpu->vram_ptr;
psx_gpu->clut_ptr = psx_gpu->vram_ptr;
psx_gpu->mask_msb = 0;
+ psx_gpu->texture_window_x = 0;
+ psx_gpu->texture_window_y = 0;
+ psx_gpu->texture_mask_width = 0xFF;
+ psx_gpu->texture_mask_height = 0xFF;
+
+ psx_gpu->render_mode = 0;
+
memset(psx_gpu->vram_ptr, 0, sizeof(u16) * 1024 * 512);
initialize_reciprocal_table();
// d1: (2 3 6 7): y0
// d2: (4 5 6 7): x0 ^ y0
-
psx_gpu->dither_table[0] = dither_table_row(-4, 0, -3, 1);
psx_gpu->dither_table[1] = dither_table_row(2, -2, 3, -1);
psx_gpu->dither_table[2] = dither_table_row(-3, 1, -4, 0);
psx_gpu->dither_table[3] = dither_table_row(3, -1, 2, -2);
psx_gpu->primitive_type = PRIMITIVE_TYPE_UNKNOWN;
+
+ psx_gpu->enhancement_x_threshold = 256;
}
u64 get_us(void)
#endif
+#include "psx_gpu_4x.c"
notaz.gp2x.de / pcsx_rearmed.git / blobdiff