sgilissen/llamaplayer
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

GPU/Software: Reduce number of copies by one, enable 16-bit scanout

Browse Source
This commit is contained in:
Connor McLaughlin
2020-10-22 01:25:33 +10:00
parent beffbaee39
commit d3d881aa6b
20 changed files with 875 additions and 187 deletions
Download Patch File Download Diff File Expand all files Collapse all files

449
src/core/gpu_sw.cpp
View File

@ -1,11 +1,24 @@
#include "gpu_sw.h"
#include "common/align.h"
#include "common/assert.h"
#include "common/cpu_detect.h"
#include "common/log.h"
#include "common/make_array.h"
#include "host_display.h"
#include "system.h"
#include <algorithm>
Log_SetChannel(GPU_SW);
#if defined(CPU_X64)
#include <emmintrin.h>
#elif defined(CPU_AARCH64)
#ifdef _MSC_VER
#include <arm64_neon.h>
#else
#include <arm_neon.h>
#endif
#endif
GPU_SW::GPU_SW()
{
m_vram.fill(0);
@ -28,9 +41,27 @@ bool GPU_SW::Initialize(HostDisplay* host_display)
if (!GPU::Initialize(host_display))
return false;
m_display_texture = host_display->CreateTexture(VRAM_WIDTH, VRAM_HEIGHT, nullptr, 0, true);
if (!m_display_texture)
return false;
static constexpr auto formats_for_16bit = make_array(HostDisplayPixelFormat::RGB565, HostDisplayPixelFormat::RGBA5551,
HostDisplayPixelFormat::RGBA8, HostDisplayPixelFormat::BGRA8);
static constexpr auto formats_for_24bit =
make_array(HostDisplayPixelFormat::RGBA8, HostDisplayPixelFormat::BGRA8, HostDisplayPixelFormat::RGB565,
HostDisplayPixelFormat::RGBA5551);
for (const HostDisplayPixelFormat format : formats_for_16bit)
{
if (m_host_display->SupportsDisplayPixelFormat(format))
{
m_16bit_display_format = format;
break;
}
}
for (const HostDisplayPixelFormat format : formats_for_24bit)
{
if (m_host_display->SupportsDisplayPixelFormat(format))
{
m_24bit_display_format = format;
break;
}
}
return true;
}
@ -42,74 +73,323 @@ void GPU_SW::Reset()
m_vram.fill(0);
}
void GPU_SW::CopyOut15Bit(u32 src_x, u32 src_y, u32* dst_ptr, u32 dst_stride, u32 width, u32 height, bool interlaced,
bool interleaved)
template<HostDisplayPixelFormat out_format, typename out_type>
static void CopyOutRow16(const u16* src_ptr, out_type* dst_ptr, u32 width);
template<HostDisplayPixelFormat out_format, typename out_type>
static out_type VRAM16ToOutput(u16 value);
template<>
ALWAYS_INLINE u16 VRAM16ToOutput<HostDisplayPixelFormat::RGBA5551, u16>(u16 value)
{
return (value & 0x3E0) | ((value >> 10) & 0x1F) | ((value & 0x1F) << 10);
}
template<>
ALWAYS_INLINE u16 VRAM16ToOutput<HostDisplayPixelFormat::RGB565, u16>(u16 value)
{
return ((value & 0x3E0) << 1) | ((value & 0x20) << 1) | ((value >> 10) & 0x1F) | ((value & 0x1F) << 11);
}
template<>
ALWAYS_INLINE u32 VRAM16ToOutput<HostDisplayPixelFormat::RGBA8, u32>(u16 value)
{
u8 r = Truncate8(value & 31);
u8 g = Truncate8((value >> 5) & 31);
u8 b = Truncate8((value >> 10) & 31);
// 00012345 -> 1234545
b = (b << 3) | (b & 0b111);
g = (g << 3) | (g & 0b111);
r = (r << 3) | (r & 0b111);
return ZeroExtend32(r) | (ZeroExtend32(g) << 8) | (ZeroExtend32(b) << 16) | (0xFF000000u);
}
template<>
ALWAYS_INLINE u32 VRAM16ToOutput<HostDisplayPixelFormat::BGRA8, u32>(u16 value)
{
u8 r = Truncate8(value & 31);
u8 g = Truncate8((value >> 5) & 31);
u8 b = Truncate8((value >> 10) & 31);
// 00012345 -> 1234545
b = (b << 3) | (b & 0b111);
g = (g << 3) | (g & 0b111);
r = (r << 3) | (r & 0b111);
return ZeroExtend32(b) | (ZeroExtend32(g) << 8) | (ZeroExtend32(r) << 16) | (0xFF000000u);
}
template<>
ALWAYS_INLINE void CopyOutRow16<HostDisplayPixelFormat::RGBA5551, u16>(const u16* src_ptr, u16* dst_ptr, u32 width)
{
u32 col = 0;
#if defined(CPU_X64)
const u32 aligned_width = Common::AlignDownPow2(width, 8);
for (; col < aligned_width; col += 8)
{
const __m128i single_mask = _mm_set1_epi16(0x1F);
__m128i value = _mm_loadu_si128(reinterpret_cast<const __m128i*>(src_ptr));
src_ptr += 8;
__m128i a = _mm_and_si128(value, _mm_set1_epi16(static_cast<s16>(static_cast<u16>(0x3E0))));
__m128i b = _mm_and_si128(_mm_srli_epi16(value, 10), single_mask);
__m128i c = _mm_slli_epi16(_mm_and_si128(value, single_mask), 10);
value = _mm_or_si128(_mm_or_si128(a, b), c);
_mm_storeu_si128(reinterpret_cast<__m128i*>(dst_ptr), value);
dst_ptr += 8;
}
#elif defined(CPU_AARCH64)
const u32 aligned_width = Common::AlignDownPow2(width, 8);
for (; col < aligned_width; col += 8)
{
const uint16x8_t single_mask = vdupq_n_u16(0x1F);
uint16x8_t value = vld1q_u16(src_ptr);
src_ptr += 8;
uint16x8_t a = vandq_u16(value, vdupq_n_u16(0x3E0));
uint16x8_t b = vandq_u16(vshrq_n_u16(value, 10), single_mask);
uint16x8_t c = vshlq_n_u16(vandq_u16(value, single_mask), 10);
value = vorrq_u16(vorrq_u16(a, b), c);
vst1q_u16(dst_ptr, value);
dst_ptr += 8;
}
#endif
for (; col < width; col++)
*(dst_ptr++) = VRAM16ToOutput<HostDisplayPixelFormat::RGBA5551, u16>(*(src_ptr++));
}
template<>
ALWAYS_INLINE void CopyOutRow16<HostDisplayPixelFormat::RGB565, u16>(const u16* src_ptr, u16* dst_ptr, u32 width)
{
u32 col = 0;
#if defined(CPU_X64)
const u32 aligned_width = Common::AlignDownPow2(width, 8);
for (; col < aligned_width; col += 8)
{
const __m128i single_mask = _mm_set1_epi16(0x1F);
__m128i value = _mm_loadu_si128(reinterpret_cast<const __m128i*>(src_ptr));
src_ptr += 8;
__m128i a = _mm_slli_epi16(_mm_and_si128(value, _mm_set1_epi16(static_cast<s16>(static_cast<u16>(0x3E0)))), 1);
__m128i b = _mm_slli_epi16(_mm_and_si128(value, _mm_set1_epi16(static_cast<s16>(static_cast<u16>(0x20)))), 1);
__m128i c = _mm_and_si128(_mm_srli_epi16(value, 10), single_mask);
__m128i d = _mm_slli_epi16(_mm_and_si128(value, single_mask), 11);
value = _mm_or_si128(_mm_or_si128(_mm_or_si128(a, b), c), d);
_mm_storeu_si128(reinterpret_cast<__m128i*>(dst_ptr), value);
dst_ptr += 8;
}
#elif defined(CPU_AARCH64)
const u32 aligned_width = Common::AlignDownPow2(width, 8);
const uint16x8_t single_mask = vdupq_n_u16(0x1F);
for (; col < aligned_width; col += 8)
{
uint16x8_t value = vld1q_u16(src_ptr);
src_ptr += 8;
uint16x8_t a = vshlq_n_u16(vandq_u16(value, vdupq_n_u16(0x3E0)), 1); // (value & 0x3E0) << 1
uint16x8_t b = vshlq_n_u16(vandq_u16(value, vdupq_n_u16(0x20)), 1); // (value & 0x20) << 1
uint16x8_t c = vandq_u16(vshrq_n_u16(value, 10), single_mask); // ((value >> 10) & 0x1F)
uint16x8_t d = vshlq_n_u16(vandq_u16(value, single_mask), 11); // ((value & 0x1F) << 11)
value = vorrq_u16(vorrq_u16(vorrq_u16(a, b), c), d);
vst1q_u16(dst_ptr, value);
dst_ptr += 8;
}
#endif
for (; col < width; col++)
*(dst_ptr++) = VRAM16ToOutput<HostDisplayPixelFormat::RGB565, u16>(*(src_ptr++));
}
template<>
ALWAYS_INLINE void CopyOutRow16<HostDisplayPixelFormat::RGBA8, u32>(const u16* src_ptr, u32* dst_ptr, u32 width)
{
for (u32 col = 0; col < width; col++)
*(dst_ptr++) = VRAM16ToOutput<HostDisplayPixelFormat::RGBA8, u32>(*(src_ptr++));
}
template<>
ALWAYS_INLINE void CopyOutRow16<HostDisplayPixelFormat::BGRA8, u32>(const u16* src_ptr, u32* dst_ptr, u32 width)
{
for (u32 col = 0; col < width; col++)
*(dst_ptr++) = VRAM16ToOutput<HostDisplayPixelFormat::BGRA8, u32>(*(src_ptr++));
}
template<HostDisplayPixelFormat display_format>
void GPU_SW::CopyOut15Bit(u32 src_x, u32 src_y, u32 width, u32 height, u32 field, bool interlaced, bool interleaved)
{
u8* dst_ptr;
u32 dst_stride;
using OutputPixelType = std::conditional_t<
display_format == HostDisplayPixelFormat::RGBA8 || display_format == HostDisplayPixelFormat::BGRA8, u32, u16>;
if (!interlaced)
{
if (!m_host_display->BeginSetDisplayPixels(display_format, width, height, reinterpret_cast<void**>(&dst_ptr),
&dst_stride))
{
return;
}
}
else
{
dst_stride = Common::AlignUpPow2<u32>(width * sizeof(OutputPixelType), 4);
dst_ptr = m_display_texture_buffer.data() + (field != 0 ? dst_stride : 0);
}
const u32 output_stride = dst_stride;
const u8 interlaced_shift = BoolToUInt8(interlaced);
const u8 interleaved_shift = BoolToUInt8(interleaved);
// Fast path when not wrapping around.
if ((src_x + width) <= VRAM_WIDTH && (src_y + height) <= VRAM_HEIGHT)
{
const u32 rows = height >> interlaced_shift;
dst_stride <<= interlaced_shift;
height >>= interlaced_shift;
const u16* src_ptr = &m_vram[src_y * VRAM_WIDTH + src_x];
const u32 src_stride = VRAM_WIDTH << interleaved_shift;
for (u32 row = 0; row < height; row++)
const u32 src_step = VRAM_WIDTH << interleaved_shift;
for (u32 row = 0; row < rows; row++)
{
const u16* src_row_ptr = src_ptr;
u32* dst_row_ptr = dst_ptr;
for (u32 col = 0; col < width; col++)
*(dst_row_ptr++) = RGBA5551ToRGBA8888(*(src_row_ptr++));
src_ptr += src_stride;
CopyOutRow16<display_format>(src_ptr, reinterpret_cast<OutputPixelType*>(dst_ptr), width);
src_ptr += src_step;
dst_ptr += dst_stride;
}
}
else
{
const u32 rows = height >> interlaced_shift;
dst_stride <<= interlaced_shift;
height >>= interlaced_shift;
const u32 end_x = src_x + width;
for (u32 row = 0; row < height; row++)
for (u32 row = 0; row < rows; row++)
{
const u16* src_row_ptr = &m_vram[(src_y % VRAM_HEIGHT) * VRAM_WIDTH];
u32* dst_row_ptr = dst_ptr;
OutputPixelType* dst_row_ptr = reinterpret_cast<OutputPixelType*>(dst_ptr);
for (u32 col = src_x; col < end_x; col++)
*(dst_row_ptr++) = RGBA5551ToRGBA8888(src_row_ptr[col % VRAM_WIDTH]);
src_y += (1 << interleaved_shift);
dst_ptr += dst_stride;
{
*(dst_row_ptr++) = VRAM16ToOutput<display_format, OutputPixelType>(src_row_ptr[col % VRAM_WIDTH]);
src_y += (1 << interleaved_shift);
dst_ptr += dst_stride;
}
}
}
if (!interlaced)
{
m_host_display->EndSetDisplayPixels();
}
else
{
m_host_display->SetDisplayPixels(display_format, width, height, m_display_texture_buffer.data(), output_stride);
}
}
void GPU_SW::CopyOut24Bit(u32 src_x, u32 src_y, u32* dst_ptr, u32 dst_stride, u32 width, u32 height, bool interlaced,
void GPU_SW::CopyOut15Bit(HostDisplayPixelFormat display_format, u32 src_x, u32 src_y, u32 width, u32 height, u32 field,
bool interlaced, bool interleaved)
{
switch (display_format)
{
case HostDisplayPixelFormat::RGBA5551:
CopyOut15Bit<HostDisplayPixelFormat::RGBA5551>(src_x, src_y, width, height, field, interlaced, interleaved);
break;
case HostDisplayPixelFormat::RGB565:
CopyOut15Bit<HostDisplayPixelFormat::RGB565>(src_x, src_y, width, height, field, interlaced, interleaved);
break;
case HostDisplayPixelFormat::RGBA8:
CopyOut15Bit<HostDisplayPixelFormat::RGBA8>(src_x, src_y, width, height, field, interlaced, interleaved);
break;
case HostDisplayPixelFormat::BGRA8:
CopyOut15Bit<HostDisplayPixelFormat::BGRA8>(src_x, src_y, width, height, field, interlaced, interleaved);
break;
default:
break;
}
}
template<HostDisplayPixelFormat display_format>
void GPU_SW::CopyOut24Bit(u32 src_x, u32 src_y, u32 skip_x, u32 width, u32 height, u32 field, bool interlaced,
bool interleaved)
{
u8* dst_ptr;
u32 dst_stride;
using OutputPixelType = std::conditional_t<
display_format == HostDisplayPixelFormat::RGBA8 || display_format == HostDisplayPixelFormat::BGRA8, u32, u16>;
if (!interlaced)
{
if (!m_host_display->BeginSetDisplayPixels(display_format, width, height, reinterpret_cast<void**>(&dst_ptr),
&dst_stride))
{
return;
}
}
else
{
dst_stride = Common::AlignUpPow2<u32>(width * sizeof(OutputPixelType), 4);
dst_ptr = m_display_texture_buffer.data() + (field != 0 ? dst_stride : 0);
}
const u32 output_stride = dst_stride;
const u8 interlaced_shift = BoolToUInt8(interlaced);
const u8 interleaved_shift = BoolToUInt8(interleaved);
const u32 rows = height >> interlaced_shift;
dst_stride <<= interlaced_shift;
if ((src_x + width) <= VRAM_WIDTH && (src_y + height) <= VRAM_HEIGHT)
if ((src_x + width) <= VRAM_WIDTH && (src_y + (rows << interleaved_shift)) <= VRAM_HEIGHT)
{
dst_stride <<= interlaced_shift;
height >>= interlaced_shift;
const u8* src_ptr = reinterpret_cast<const u8*>(&m_vram[src_y * VRAM_WIDTH + src_x]);
const u8* src_ptr = reinterpret_cast<const u8*>(&m_vram[src_y * VRAM_WIDTH + src_x]) + (skip_x * 3);
const u32 src_stride = (VRAM_WIDTH << interleaved_shift) * sizeof(u16);
for (u32 row = 0; row < height; row++)
for (u32 row = 0; row < rows; row++)
{
const u8* src_row_ptr = src_ptr;
u8* dst_row_ptr = reinterpret_cast<u8*>(dst_ptr);
for (u32 col = 0; col < width; col++)
if constexpr (display_format == HostDisplayPixelFormat::RGBA8)
{
*(dst_row_ptr++) = *(src_row_ptr++);
*(dst_row_ptr++) = *(src_row_ptr++);
*(dst_row_ptr++) = *(src_row_ptr++);
*(dst_row_ptr++) = 0xFF;
const u8* src_row_ptr = src_ptr;
u8* dst_row_ptr = reinterpret_cast<u8*>(dst_ptr);
for (u32 col = 0; col < width; col++)
{
*(dst_row_ptr++) = *(src_row_ptr++);
*(dst_row_ptr++) = *(src_row_ptr++);
*(dst_row_ptr++) = *(src_row_ptr++);
*(dst_row_ptr++) = 0xFF;
}
}
else if constexpr (display_format == HostDisplayPixelFormat::BGRA8)
{
const u8* src_row_ptr = src_ptr;
u8* dst_row_ptr = reinterpret_cast<u8*>(dst_ptr);
for (u32 col = 0; col < width; col++)
{
*(dst_row_ptr++) = src_row_ptr[2];
*(dst_row_ptr++) = src_row_ptr[1];
*(dst_row_ptr++) = src_row_ptr[0];
*(dst_row_ptr++) = 0xFF;
src_row_ptr += 3;
}
}
else if constexpr (display_format == HostDisplayPixelFormat::RGB565)
{
const u8* src_row_ptr = src_ptr;
u16* dst_row_ptr = reinterpret_cast<u16*>(dst_ptr);
for (u32 col = 0; col < width; col++)
{
*(dst_row_ptr++) = ((static_cast<u16>(src_row_ptr[0]) >> 3) << 11) |
((static_cast<u16>(src_row_ptr[1]) >> 2) << 5) | (static_cast<u16>(src_row_ptr[2]) >> 3);
src_row_ptr += 3;
}
}
else if constexpr (display_format == HostDisplayPixelFormat::RGBA5551)
{
const u8* src_row_ptr = src_ptr;
u16* dst_row_ptr = reinterpret_cast<u16*>(dst_ptr);
for (u32 col = 0; col < width; col++)
{
*(dst_row_ptr++) = ((static_cast<u16>(src_row_ptr[0]) >> 3) << 10) |
((static_cast<u16>(src_row_ptr[1]) >> 3) << 5) | (static_cast<u16>(src_row_ptr[2]) >> 3);
src_row_ptr += 3;
}
}
src_ptr += src_stride;
@ -118,39 +398,83 @@ void GPU_SW::CopyOut24Bit(u32 src_x, u32 src_y, u32* dst_ptr, u32 dst_stride, u3
}
else
{
dst_stride <<= interlaced_shift;
height >>= interlaced_shift;
for (u32 row = 0; row < height; row++)
for (u32 row = 0; row < rows; row++)
{
const u16* src_row_ptr = &m_vram[(src_y % VRAM_HEIGHT) * VRAM_WIDTH];
u32* dst_row_ptr = dst_ptr;
OutputPixelType* dst_row_ptr = reinterpret_cast<OutputPixelType*>(dst_ptr);
for (u32 col = 0; col < width; col++)
{
const u32 offset = (src_x + ((col * 3) / 2));
const u32 offset = (src_x + (((skip_x + col) * 3) / 2));
const u16 s0 = src_row_ptr[offset % VRAM_WIDTH];
const u16 s1 = src_row_ptr[(offset + 1) % VRAM_WIDTH];
const u8 shift = static_cast<u8>(col & 1u) * 8;
*(dst_row_ptr++) = (((ZeroExtend32(s1) << 16) | ZeroExtend32(s0)) >> shift) | 0xFF000000u;
const u32 rgb = (((ZeroExtend32(s1) << 16) | ZeroExtend32(s0)) >> shift);
if constexpr (display_format == HostDisplayPixelFormat::RGBA8)
{
*(dst_row_ptr++) = rgb | 0xFF000000u;
}
else if constexpr (display_format == HostDisplayPixelFormat::BGRA8)
{
*(dst_row_ptr++) = (rgb & 0x00FF00) | ((rgb & 0xFF) << 16) | ((rgb >> 16) & 0xFF) | 0xFF000000u;
}
else if constexpr (display_format == HostDisplayPixelFormat::RGB565)
{
*(dst_row_ptr++) = ((rgb >> 3) & 0x1F) | (((rgb >> 10) << 5) & 0x7E0) | (((rgb >> 19) << 11) & 0x3E0000);
}
else if constexpr (display_format == HostDisplayPixelFormat::RGBA5551)
{
*(dst_row_ptr++) = ((rgb >> 3) & 0x1F) | (((rgb >> 11) << 5) & 0x3E0) | (((rgb >> 19) << 10) & 0x1F0000);
}
}
src_y += (1 << interleaved_shift);
dst_ptr += dst_stride;
}
}
if (!interlaced)
{
m_host_display->EndSetDisplayPixels();
}
else
{
m_host_display->SetDisplayPixels(display_format, width, height, m_display_texture_buffer.data(), output_stride);
}
}
void GPU_SW::CopyOut24Bit(HostDisplayPixelFormat display_format, u32 src_x, u32 src_y, u32 skip_x, u32 width,
u32 height, u32 field, bool interlaced, bool interleaved)
{
switch (display_format)
{
case HostDisplayPixelFormat::RGBA5551:
CopyOut24Bit<HostDisplayPixelFormat::RGBA5551>(src_x, src_y, skip_x, width, height, field, interlaced,
interleaved);
break;
case HostDisplayPixelFormat::RGB565:
CopyOut24Bit<HostDisplayPixelFormat::RGB565>(src_x, src_y, skip_x, width, height, field, interlaced, interleaved);
break;
case HostDisplayPixelFormat::RGBA8:
CopyOut24Bit<HostDisplayPixelFormat::RGBA8>(src_x, src_y, skip_x, width, height, field, interlaced, interleaved);
break;
case HostDisplayPixelFormat::BGRA8:
CopyOut24Bit<HostDisplayPixelFormat::BGRA8>(src_x, src_y, skip_x, width, height, field, interlaced, interleaved);
break;
default:
break;
}
}
void GPU_SW::ClearDisplay()
{
std::memset(m_display_texture_buffer.data(), 0, sizeof(u32) * m_display_texture_buffer.size());
std::memset(m_display_texture_buffer.data(), 0, m_display_texture_buffer.size());
}
void GPU_SW::UpdateDisplay()
{
// fill display texture
m_display_texture_buffer.resize(VRAM_WIDTH * VRAM_HEIGHT);
if (!g_settings.debugging.show_vram)
{
if (IsDisplayDisabled())
@ -162,39 +486,37 @@ void GPU_SW::UpdateDisplay()
const u32 vram_offset_y = m_crtc_state.display_vram_top;
const u32 display_width = m_crtc_state.display_vram_width;
const u32 display_height = m_crtc_state.display_vram_height;
const u32 texture_offset_x = m_crtc_state.display_vram_left - m_crtc_state.regs.X;
if (IsInterlacedDisplayEnabled())
{
const u32 field = GetInterlacedDisplayField();
if (m_GPUSTAT.display_area_color_depth_24)
{
CopyOut24Bit(m_crtc_state.regs.X, vram_offset_y + field, m_display_texture_buffer.data() + field * VRAM_WIDTH,
VRAM_WIDTH, display_width + texture_offset_x, display_height, true, m_GPUSTAT.vertical_resolution);
CopyOut24Bit(m_24bit_display_format, m_crtc_state.regs.X, vram_offset_y + field,
m_crtc_state.display_vram_left - m_crtc_state.regs.X, display_width, display_height, field, true,
m_GPUSTAT.vertical_resolution);
}
else
{
CopyOut15Bit(m_crtc_state.regs.X, vram_offset_y + field, m_display_texture_buffer.data() + field * VRAM_WIDTH,
VRAM_WIDTH, display_width + texture_offset_x, display_height, true, m_GPUSTAT.vertical_resolution);
CopyOut15Bit(m_16bit_display_format, m_crtc_state.display_vram_left, vram_offset_y + field, display_width,
display_height, field, true, m_GPUSTAT.vertical_resolution);
}
}
else
{
if (m_GPUSTAT.display_area_color_depth_24)
{
CopyOut24Bit(m_crtc_state.regs.X, vram_offset_y, m_display_texture_buffer.data(), VRAM_WIDTH,
display_width + texture_offset_x, display_height, false, false);
CopyOut24Bit(m_24bit_display_format, m_crtc_state.regs.X, vram_offset_y,
m_crtc_state.display_vram_left - m_crtc_state.regs.X, display_width, display_height, 0, false,
false);
}
else
{
CopyOut15Bit(m_crtc_state.regs.X, vram_offset_y, m_display_texture_buffer.data(), VRAM_WIDTH,
display_width + texture_offset_x, display_height, false, false);
CopyOut15Bit(m_16bit_display_format, m_crtc_state.display_vram_left, vram_offset_y, display_width,
display_height, 0, false, false);
}
}
m_host_display->UpdateTexture(m_display_texture.get(), 0, 0, display_width, display_height,
m_display_texture_buffer.data(), VRAM_WIDTH * sizeof(u32));
m_host_display->SetDisplayTexture(m_display_texture->GetHandle(), VRAM_WIDTH, VRAM_HEIGHT, texture_offset_x, 0,
display_width, display_height);
m_host_display->SetDisplayParameters(m_crtc_state.display_width, m_crtc_state.display_height,
m_crtc_state.display_origin_left, m_crtc_state.display_origin_top,
m_crtc_state.display_vram_width, m_crtc_state.display_vram_height,
@ -202,11 +524,7 @@ void GPU_SW::UpdateDisplay()
}
else
{
CopyOut15Bit(0, 0, m_display_texture_buffer.data(), VRAM_WIDTH, VRAM_WIDTH, VRAM_HEIGHT, false, false);
m_host_display->UpdateTexture(m_display_texture.get(), 0, 0, VRAM_WIDTH, VRAM_HEIGHT,
m_display_texture_buffer.data(), VRAM_WIDTH * sizeof(u32));
m_host_display->SetDisplayTexture(m_display_texture->GetHandle(), VRAM_WIDTH, VRAM_HEIGHT, 0, 0, VRAM_WIDTH,
VRAM_HEIGHT);
CopyOut15Bit(m_16bit_display_format, 0, 0, VRAM_WIDTH, VRAM_HEIGHT, 0, false, false);
m_host_display->SetDisplayParameters(VRAM_WIDTH, VRAM_HEIGHT, 0, 0, VRAM_WIDTH, VRAM_HEIGHT,
static_cast<float>(VRAM_WIDTH) / static_cast<float>(VRAM_HEIGHT));
}
@ -379,7 +697,8 @@ constexpr GPU_SW::DitherLUT GPU_SW::ComputeDitherLUT()
static constexpr GPU_SW::DitherLUT s_dither_lut = GPU_SW::ComputeDitherLUT();
template<bool texture_enable, bool raw_texture_enable, bool transparency_enable, bool dithering_enable>
void ALWAYS_INLINE_RELEASE GPU_SW::ShadePixel(u32 x, u32 y, u8 color_r, u8 color_g, u8 color_b, u8 texcoord_x, u8 texcoord_y)
void ALWAYS_INLINE_RELEASE GPU_SW::ShadePixel(u32 x, u32 y, u8 color_r, u8 color_g, u8 color_b, u8 texcoord_x,
u8 texcoord_y)
{
VRAMPixel color;
bool transparent;