Emulator-Archive/fpPS4
1
0
Fork 0
mirror of https://github.com/red-prig/fpPS4.git synced 2024-10-07 03:33:20 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
17f02ea0ff
fpPS4/chip/ps4_tiling.pas
red-prig a39db791b9 +
2022-07-09 19:13:52 +03:00

3601 lines
127 KiB
ObjectPascal
Raw Blame History

unit ps4_Tiling;
{$mode objfpc}{$H+}
interface
uses
Classes,
SysUtils,
Math,
bittype,
si_ci_vi_merged_offset,
si_ci_vi_merged_enum,
si_ci_vi_merged_registers
;
const
// Depth modes (for depth buffers)
kTileModeDepth_2dThin_64 = $00000000; ///< Recommended for depth targets with one fragment per pixel.
kTileModeDepth_2dThin_128 = $00000001; ///< Recommended for depth targets with two or four fragments per pixel, or texture-readable.
kTileModeDepth_2dThin_256 = $00000002; ///< Recommended for depth targets with eight fragments per pixel.
kTileModeDepth_2dThin_512 = $00000003; ///< Recommended for depth targets with 512-byte tiles.
kTileModeDepth_2dThin_1K = $00000004; ///< Recommended for depth targets with 1024-byte tiled.
kTileModeDepth_1dThin = $00000005; ///< Not used; included only for completeness.
kTileModeDepth_2dThinPrt_256 = $00000006; ///< Recommended for partially-resident depth surfaces. Does not support aliasing multiple virtual texture pages to the same physical page.
kTileModeDepth_2dThinPrt_1K = $00000007; ///< Not used; included only for completeness.
// Display modes
kTileModeDisplay_LinearAligned = $00000008; ///< Recommended for any surface to be easily accessed on the CPU.
kTileModeDisplay_1dThin = $00000009; ///< Not used; included only for completeness.
kTileModeDisplay_2dThin = $0000000A; ///< Recommended mode for displayable render targets.
kTileModeDisplay_ThinPrt = $0000000B; ///< Supports aliasing multiple virtual texture pages to the same physical page.
kTileModeDisplay_2dThinPrt = $0000000C; ///< Does not support aliasing multiple virtual texture pages to the same physical page.
// Thin modes (for non-displayable 1D/2D/3D surfaces)
kTileModeThin_1dThin = $0000000D; ///< Recommended for read-only non-volume textures.
kTileModeThin_2dThin = $0000000E; ///< Recommended for non-displayable intermediate render targets and read/write non-volume textures.
kTileModeThin_3dThin = $0000000F; ///< Not used; included only for completeness.
kTileModeThin_ThinPrt = $00000010; ///< Recommended for partially-resident textures (PRTs). Supports aliasing multiple virtual texture pages to the same physical page.
kTileModeThin_2dThinPrt = $00000011; ///< Does not support aliasing multiple virtual texture pages to the same physical page.
kTileModeThin_3dThinPrt = $00000012; ///< Does not support aliasing multiple virtual texture pages to the same physical page.
// Thick modes (for 3D textures)
kTileModeThick_1dThick = $00000013; ///< Recommended for read-only volume textures.
kTileModeThick_2dThick = $00000014; ///< Recommended for volume textures to which pixel shaders will write.
kTileModeThick_3dThick = $00000015; ///< Not used; included only for completeness.
kTileModeThick_ThickPrt = $00000016; ///< Supports aliasing multiple virtual texture pages to the same physical page.
kTileModeThick_2dThickPrt = $00000017; ///< Does not support aliasing multiple virtual texture pages to the same physical page.
kTileModeThick_3dThickPrt = $00000018; ///< Does not support aliasing multiple virtual texture pages to the same physical page.
kTileModeThick_2dXThick = $00000019; ///< Recommended for volume textures to which pixel shaders will write.
kTileModeThick_3dXThick = $0000001A; ///< Not used; included only for completeness.
// Hugely inefficient linear display mode -- do not use!
kTileModeDisplay_LinearGeneral = $0000001F; ///< Unsupported; do not use!
kGpuModeBase = 0; ///< GPU mode that the original PlayStation 4 uses.
kGpuModeNeo = 1; ///< GPU mode that NEO uses.
kNumSamples1 = $0; ///< 1 sample per pixel.
kNumSamples2 = $1; ///< 2 samples per pixel.
kNumSamples4 = $2; ///< 4 samples per pixel.
kNumSamples8 = $3; ///< 8 samples per pixel.
kNumSamples16 = $4; ///< 16 samples per pixel.
kNumFragments1 = $0; ///< 1 fragment per pixel.
kNumFragments2 = $1; ///< 2 fragments per pixel.
kNumFragments4 = $2; ///< 4 fragments per pixel.
kNumFragments8 = $3; ///< 8 fragments per pixel.
kMicroTileModeDisplay = $00000000; ///< Only for 64 bpp and below.
kMicroTileModeThin = $00000001; ///< Non-displayable. Can be used for thin, thick, or X thick.
kMicroTileModeDepth = $00000002; ///< Only mode supported by DB.
kMicroTileModeRotated = $00000003; ///< Rotated. Not supported by Gnm.
kMicroTileModeThick = $00000004; ///< Thick and X thick, non-AA only.
kArrayModeLinearGeneral = $00000000; ///< Linear pixel storage; no alignment or padding restrictions. DEPRECATED -- Do not use!
kArrayModeLinearAligned = $00000001; ///< Linear pixel storage with some minor alignment requirements and internal padding.
kArrayMode1dTiledThin = $00000002; ///< Micro-tile-only tiling for non-volume surfaces. Not valid for AA modes.
kArrayMode1dTiledThick = $00000003; ///< Micro-tile-only tiling for volume surfaces (8x8x4 pixel micro-tiles). Not valid for AA modes.
kArrayMode2dTiledThin = $00000004; ///< Macro-tile tiling for non-volume surfaces.
kArrayModeTiledThinPrt = $00000005; ///< Macro-tile tiling for non-volume partially-resident texture (PRT) surfaces. Supports aliasing multiple virtual texture pages to the same physical page.
kArrayMode2dTiledThinPrt = $00000006; ///< Macro-tile tiling for non-volume partially-resident texture (PRT) surfaces. Does not support aliasing multiple virtual texture pages to the same physical page.
kArrayMode2dTiledThick = $00000007; ///< Macro-tile tiling for volume surfaces (8x8x4 pixel micro-tiles).
kArrayMode2dTiledXThick = $00000008; ///< Macro-tile tiling for volume surfaces (8x8x8 pixel micro-tiles).
kArrayModeTiledThickPrt = $00000009; ///< Micro-tile-only tiling for partially-resident texture (PRT) volume surfaces (8x8x4 pixel micro-tiles). Supports aliasing multiple virtual texture pages to the same physical page.
kArrayMode2dTiledThickPrt = $0000000a; ///< Macro-tile tiling for partially-resident texture (PRT) volume surfaces (8x8x4 pixel micro-tiles). Does not support aliasing multiple virtual texture pages to the same physical page.
kArrayMode3dTiledThinPrt = $0000000b; ///< Macro-tile tiling for partially-resident texture (PRT) non-volume surfaces. Z slices are rotated by pipe. Does not support aliasing multiple virtual texture pages to the same physical page.
kArrayMode3dTiledThin = $0000000c; ///< Macro-tile tiling for non-volume surfaces. Z slices are rotated by pipe.
kArrayMode3dTiledThick = $0000000d; ///< Macro-tile tiling for volume surfaces (8x8x4 pixel micro-tiles). Z slices are rotated by pipe.
kArrayMode3dTiledXThick = $0000000e; ///< Macro-tile tiling for volume surfaces (8x8x8 pixel micro-tiles). Z slices are rotated by pipe.
kArrayMode3dTiledThickPrt = $0000000f; ///< Macro-tile tiling for partially-resident texture (PRT) volume surfaces (8x8x4 pixel micro-tiles). Z slices are rotated by pipe. Does not support aliasing multiple virtual texture pages to the same physical page.
kPipeConfigP8_32x32_8x16 = $0000000a;
kPipeConfigP8_32x32_16x16 = $0000000c;
kPipeConfigP16 = $00000012;
kDramRowSize = $400;
kNumLogicalBanks = 16;
kPipeInterleaveBytes = 256;
kBankInterleave = 1;
kMicroTileWidth = 8;
kMicroTileHeight = 8;
kNumMicroTilePixels = kMicroTileWidth*kMicroTileHeight;
kCmaskCacheBits = $400;
kHtileCacheBits = $4000;
kSurfaceFormatInvalid = $00000000; ///< Invalid surface format.
kSurfaceFormat8 = $00000001; ///< One 8-bit channel. X=0xFF
kSurfaceFormat16 = $00000002; ///< One 16-bit channel. X=0xFFFF
kSurfaceFormat8_8 = $00000003; ///< Two 8-bit channels. X=0x00FF, Y=0xFF00
kSurfaceFormat32 = $00000004; ///< One 32-bit channel. X=0xFFFFFFFF
kSurfaceFormat16_16 = $00000005; ///< Two 16-bit channels. X=0x0000FFFF, Y=0xFFFF0000
kSurfaceFormat10_11_11 = $00000006; ///< One 10-bit channel (Z) and two 11-bit channels (Y,X). X=0x000007FF, Y=0x003FF800, Z=0xFFC00000 Interpreted only as floating-point by texture unit, but also as integer by rasterizer.
kSurfaceFormat11_11_10 = $00000007; ///< Two 11-bit channels (Z,Y) and one 10-bit channel (X). X=0x000003FF, Y=0x001FFC00, Z=0xFFE00000 Interpreted only as floating-point by texture unit, but also as integer by rasterizer.
kSurfaceFormat10_10_10_2 = $00000008; ///< Three 10-bit channels (W,Z,Y) and one 2-bit channel (X). X=0x00000003, Y=0x00000FFC, Z=0x003FF000, W=0xFFC00000 X is never negative, even when YZW are.
kSurfaceFormat2_10_10_10 = $00000009; ///< One 2-bit channel (W) and three 10-bit channels (Z,Y,X). X=0x000003FF, Y=0x000FFC00, Z=0x3FF00000, W=0xC0000000 W is never negative, even when XYZ are.
kSurfaceFormat8_8_8_8 = $0000000a; ///< Four 8-bit channels. X=0x000000FF, Y=0x0000FF00, Z=0x00FF0000, W=0xFF000000
kSurfaceFormat32_32 = $0000000b; ///< Two 32-bit channels.
kSurfaceFormat16_16_16_16 = $0000000c; ///< Four 16-bit channels.
kSurfaceFormat32_32_32 = $0000000d; ///< Three 32-bit channels.
kSurfaceFormat32_32_32_32 = $0000000e; ///< Four 32-bit channels.
kSurfaceFormat5_6_5 = $00000010; ///< One 5-bit channel (Z), one 6-bit channel (Y), and a second 5-bit channel (X). X=0x001F, Y=0x07E0, Z=0xF800
kSurfaceFormat1_5_5_5 = $00000011; ///< One 1-bit channel (W) and three 5-bit channels (Z,Y,X). X=0x001F, Y=0x03E0, Z=0x7C00, W=0x8000
kSurfaceFormat5_5_5_1 = $00000012; ///< Three 5-bit channels (W,Z,Y) and one 1-bit channel (X). X=0x0001, Y=0x003E, Z=0x07C0, W=0xF800
kSurfaceFormat4_4_4_4 = $00000013; ///< Four 4-bit channels. X=0x000F, Y=0x00F0, Z=0x0F00, W=0xF000
kSurfaceFormat8_24 = $00000014; ///< One 8-bit channel and one 24-bit channel.
kSurfaceFormat24_8 = $00000015; ///< One 24-bit channel and one 8-bit channel.
kSurfaceFormatX24_8_32 = $00000016; ///< One 24-bit channel, one 8-bit channel, and one 32-bit channel.
kSurfaceFormatGB_GR = $00000020; ///< To be documented.
kSurfaceFormatBG_RG = $00000021; ///< To be documented.
kSurfaceFormat5_9_9_9 = $00000022; ///< One 5-bit channel (W) and three 9-bit channels (Z,Y,X). X=0x000001FF, Y=0x0003FE00, Z=0x07FC0000, W=0xF8000000. Interpreted only as three 9-bit denormalized mantissas, and one shared 5-bit exponent.
kSurfaceFormatBc1 = $00000023; ///< BC1 block-compressed surface.
kSurfaceFormatBc2 = $00000024; ///< BC2 block-compressed surface.
kSurfaceFormatBc3 = $00000025; ///< BC3 block-compressed surface.
kSurfaceFormatBc4 = $00000026; ///< BC4 block-compressed surface.
kSurfaceFormatBc5 = $00000027; ///< BC5 block-compressed surface.
kSurfaceFormatBc6 = $00000028; ///< BC6 block-compressed surface.
kSurfaceFormatBc7 = $00000029; ///< BC7 block-compressed surface.
kSurfaceFormatFmask8_S2_F1 = $0000002C; ///< 8 bits-per-element FMASK surface (2 samples, 1 fragment).
kSurfaceFormatFmask8_S4_F1 = $0000002D; ///< 8 bits-per-element FMASK surface (4 samples, 1 fragment).
kSurfaceFormatFmask8_S8_F1 = $0000002E; ///< 8 bits-per-element FMASK surface (8 samples, 1 fragment).
kSurfaceFormatFmask8_S2_F2 = $0000002F; ///< 8 bits-per-element FMASK surface (2 samples, 2 fragments).
kSurfaceFormatFmask8_S4_F2 = $00000030; ///< 8 bits-per-element FMASK surface (8 samples, 2 fragments).
kSurfaceFormatFmask8_S4_F4 = $00000031; ///< 8 bits-per-element FMASK surface (4 samples, 4 fragments).
kSurfaceFormatFmask16_S16_F1 = $00000032; ///< 16 bits-per-element FMASK surface (16 samples, 1 fragment).
kSurfaceFormatFmask16_S8_F2 = $00000033; ///< 16 bits-per-element FMASK surface (8 samples, 2 fragments).
kSurfaceFormatFmask32_S16_F2 = $00000034; ///< 32 bits-per-element FMASK surface (16 samples, 2 fragments).
kSurfaceFormatFmask32_S8_F4 = $00000035; ///< 32 bits-per-element FMASK surface (8 samples, 4 fragments).
kSurfaceFormatFmask32_S8_F8 = $00000036; ///< 32 bits-per-element FMASK surface (8 samples, 8 fragments).
kSurfaceFormatFmask64_S16_F4 = $00000037; ///< 64 bits-per-element FMASK surface (16 samples, 4 fragments).
kSurfaceFormatFmask64_S16_F8 = $00000038; ///< 64 bits-per-element FMASK surface (16 samples, 8 fragments).
kSurfaceFormat4_4 = $00000039; ///< Two 4-bit channels (Y,X). X=0x0F, Y=0xF0
kSurfaceFormat6_5_5 = $0000003A; ///< One 6-bit channel (Z) and two 5-bit channels (Y,X). X=0x001F, Y=0x03E0, Z=0xFC00
kSurfaceFormat1 = $0000003B; ///< One 1-bit channel. 8 pixels per byte, with pixel index increasing from LSB to MSB.
kSurfaceFormat1Reversed = $0000003C; ///< One 1-bit channel. 8 pixels per byte, with pixel index increasing from MSB to LSB.
kSurfaceTypeColorTargetDisplayable = 1 ; ///< A render target that will be sent to scan-out. This may have different restrictions than a general-purpose color buffer.
kSurfaceTypeColorTarget = 2 ; ///< An intermediate render target.
kSurfaceTypeDepthTarget = 3 ; ///< A depth target that supports a depth buffer and/or a stencil buffer.
kSurfaceTypeDepthOnlyTarget = 4 ; ///< A depth target that supports a depth buffer, but may not also support a stencil buffer.
kSurfaceTypeStencilOnlyTarget = 5 ; ///< A depth target that supports a stencil buffer, but may not also support a depth buffer.
kSurfaceTypeFmaskBuffer = 6 ; ///< An FMASK surface.
kSurfaceTypeTextureFlat = 7 ; ///< A read-only 1D or 2D texture (or texture array).
kSurfaceTypeTextureVolume = 8 ; ///< A read-only 3D texture.
kSurfaceTypeTextureCubemap = 9 ; ///< A read-only cubic environment map (or cubemap array).
kSurfaceTypeRwTextureFlat = 10; ///< A read/write 1D or 2D texture (or texture array).
kSurfaceTypeRwTextureVolume = 11; ///< A read/write 3D texture.
kSurfaceTypeRwTextureCubemap = 12; ///< A read/write cubic environment map (or cubemap array).
kSurfaceMipmapDisable = 0; ///< Surface has no mipmaps beyond the base level.
kSurfaceMipmapEnable = 1; ///< Surface has one or more mipmaps beyond the base level.
kDccBlockSize64 = 0; ///< 64-byte blocks.
kDccBlockSize128 = 1; ///< 128-byte blocks.
kDccBlockSize256 = 2; ///< 256-byte blocks.
type
TDATA_FORMAT=bitpacked record
m_surfaceFormat :bit8; //0 < Gnm::SurfaceFormat.
m_channelType :bit4; //8 < Gnm::TextureChannelType. Can be easily converted to BufferChannelType or RenderTargetChannelType.
m_channelX :bit3; //12 < Gnm::TextureChannel.
m_channelY :bit3; //15 < Gnm::TextureChannel.
m_channelZ :bit3; //18 < Gnm::TextureChannel.
m_channelW :bit3; //21 < Gnm::TextureChannel.
m_unused :bit8; //29 < Unused.
end;
RenderTargetInitFlags=bitpacked record
enableCmaskFastClear :0..1; ///< Set to 1 to enable CMASK fast clears for this target.
enableFmaskCompression :0..1; ///< Set to 1 to enable FMASK compression for this target. Has no effect for non-MSAA surfaces. Ignored if <c>enableCmaskFastClear=0</c>.
enableColorTextureWithoutDecompress :0..1; ///< Set to 1 to allow the compressed color buffer to be sampled without an explicit DCC decompression pass. Ignored if <c>enableDccCompression=0</c>.
enableFmaskTextureWithoutDecompress :0..1; ///< Set to 1 to allow the FMASK surface to be sampled without an explicit FMASK decompression pass. Ignored if <c>enableFmaskCompression=0</c> or <c>enableDccCompression=0</c>.
enableDccCompression :0..1; ///< Set to 1 to enable DCC color compression for this target. NEO mode only.
reserved :0..134217727; ///< This field must be set to zero.
end;
TRENDER_TARGET=packed object
BASE :TCB_COLOR0_BASE ; //0 mmCB_COLOR0_BASE_DEFAULT
PITCH :TCB_COLOR0_PITCH ; //1 mmCB_COLOR0_PITCH_DEFAULT
SLICE :TCB_COLOR0_SLICE ; //2 mmCB_COLOR0_SLICE_DEFAULT
VIEW :TCB_COLOR0_VIEW ; //3 mmCB_COLOR0_VIEW_DEFAULT
INFO :TCB_COLOR0_INFO ; //4 mmCB_COLOR0_INFO_DEFAULT
ATTRIB :TCB_COLOR0_ATTRIB ; //5 mmCB_COLOR0_ATTRIB_DEFAULT
DCC_CONTROL:TCB_COLOR0_DCC_CONTROL; //6 mmCB_COLOR0_DCC_CONTROL_DEFAULT
CMASK :TCB_COLOR0_CMASK ; //7 mmCB_COLOR0_CMASK_DEFAULT
CMASK_SLICE:TCB_COLOR0_CMASK_SLICE; //8 mmCB_COLOR0_CMASK_SLICE_DEFAULT
FMASK :TCB_COLOR0_FMASK ; //9 mmCB_COLOR0_FMASK_DEFAULT
FMASK_SLICE:TCB_COLOR0_FMASK_SLICE; //10 mmCB_COLOR0_FMASK_SLICE_DEFAULT
CLEAR_WORD :QWORD; //11 mmCB_COLOR0_CLEAR_WORD0_DEFAULT
//12 mmCB_COLOR0_CLEAR_WORD1_DEFAULT
DCC_BASE :TCB_COLOR0_DCC_BASE ; //13 mmCB_COLOR0_DCC_BASE_DEFAULT
Width,Height:WORD; //14not a reg
function getTileMode:Byte; inline;
function getWidth:WORD; inline;
function getHeight:WORD; inline;
function getMinimumGpuMode:Byte; inline;
function getNumFragments:Byte; inline;
function getPitchDiv8Minus1:Word; inline;
function getPitch:DWORD; inline;
function getDccCompressionEnable:Boolean; inline;
function getDataFormat:TDATA_FORMAT;
function getTileSwizzleMask:Byte;
end;
RenderTargetSpec=object
m_width:DWORD; ///< The requested width, in pixels. The actual surface width may be padded to accommodate hardware restrictions. Valid range is <c>[1..16384]</c>.
m_height:DWORD; ///< The requested height, in pixels. The actual surface width may be padded to accommodate hardware restrictions. Valid range is <c>[1..16384]</c>.
m_pitch:DWORD; ///< The requested pitch in pixels. If this value is zero, the library will compute the minimum valid pitch for the surface given the restrictions
///< imposed by other surface parameters; otherwise the provided pitch will be used, provided it also conforms to hardware restrictions. A non-zero
///< pitch that does not conform to hardware restrictions will cause initialization to fail. The valid range is [0..16384] subject to hardware restrictions.
m_numSlices:DWORD; ///< The requested number of array slices. The actual number of slices may be padded to accommodate hardware restrictions. Valid range is [1..2048].
m_colorFormat:TDATA_FORMAT; ///< The requested format for each color fragment. This format must be one that is supported for render targets (see DataFormat::supportsRenderTarget()).
m_colorTileModeHint:DWORD; ///< The requested tiling mode. The actual tiling mode by be different to accommodate hardware restrictions; use RenderTarget::getTileMode() to determine the object's final tiling mode.
m_minGpuMode:DWORD; ///< The minimum GPU mode on which this surface should be supported. This setting may affect surface sizes, memory layout, available features, and so on.
m_numSamples:DWORD; ///< The number of samples per pixel. This must not be less than <c><i>numFragments</i></c>.
m_numFragments:DWORD; ///< The number of fragments per pixel. This must not be greater than <c><i>numSamples</i></c>.
m_flags:RenderTargetInitFlags; ///< Used to enable additional RenderTarget features.
m_regs:TRENDER_TARGET;
end;
PSurfaceFlags=^SurfaceFlags;
SurfaceFlags=bitpacked record
m_colorTarget :bit1; //< DEPRECATED -- Unused.
m_depthTarget :bit1; //< Flag indicates whether the surface is a depth-only buffer.
m_stencilTarget :bit1; //< Flag indicates whether the surface is a stencil-only buffer.
m_texture :bit1; //< DEPRECATED -- Unused.
m_cube :bit1; //< Flag indicates whether the surface is a cubemap.
m_volume :bit1; //< Flag indicates whether the surface is a volume texture.
m_fmask :bit1; //< Flag indicates whether the surface is an FMASK surface.
m_cubeAsArray :bit1; //< DEPRECATED -- Unused.
m_overlay :bit1; //< DEPRECATED -- Unused.
m_noStencil :bit1; //< DEPRECATED -- Unused.
m_display :bit1; //< DEPRECATED -- Unused.
m_prt :bit1; //< DEPRECATED -- Unused.
m_pow2Pad :bit1; //< If set, all dimensions will be padded to powers of 2. <i>Must</i> be set for any texture with mipmaps, including the base level.
m_texCompatible :bit1; //< Flag indicates whether the surface must be configured such that it can be used/aliased as a Texture. NEO ONLY.
m_reserved :bit18; //< Reserved bits.
end;
TilingParameters=object
m_tileMode:DWORD;
m_minGpuMode:DWORD;
m_linearWidth :DWORD;
m_linearHeight :DWORD;
m_linearDepth :DWORD;
m_numFragmentsPerPixel:DWORD;
m_baseTiledPitch :DWORD;
m_mipLevel :DWORD;
m_arraySlice :DWORD;
m_surfaceFlags :SurfaceFlags;
m_bitsPerFragment :DWORD;
m_isBlockCompressed :Boolean;
m_tileSwizzleMask :Byte;
function initFromRenderTarget(var target:TRENDER_TARGET;arraySlice:DWORD):Integer;
function initFromRenderTargetSpec(var target:RenderTargetSpec;arraySlice:DWORD):Integer;
end;
Tiler2d=object
m_minGpuMode:DWORD;
m_tileMode:DWORD;
m_arrayMode:DWORD;
m_linearWidth:DWORD;
m_linearHeight:DWORD;
m_linearDepth:DWORD;
m_paddedWidth:DWORD;
m_paddedHeight:DWORD;
m_paddedDepth:DWORD;
m_bitsPerElement:DWORD;
m_linearSizeBytes:DWORD;
m_tiledSizeBytes:DWORD;
m_microTileMode:DWORD;
m_pipeConfig:DWORD;
m_arraySlice:DWORD;
m_numFragmentsPerPixel:DWORD;
m_bankWidth:DWORD;
m_bankHeight:DWORD;
m_numBanks:DWORD;
m_macroTileAspect:DWORD;
m_tileSplitBytes:DWORD;
m_numPipes:DWORD;
m_tileThickness:DWORD;
m_macroTileWidth:DWORD;
m_macroTileHeight:DWORD;
m_pipeInterleaveBytes:DWORD;
m_pipeInterleaveBits:DWORD;
m_pipeInterleaveMask:DWORD;
m_pipeBits:DWORD;
m_bankBits:DWORD;
m_pipeMask:DWORD;
m_bankSwizzleMask:DWORD;
m_pipeSwizzleMask:DWORD;
function init(var tp:TilingParameters):integer;
function getTiledElementBitOffset(var outTiledBitOffset:QWORD;x,y,z,fragmentIndex:DWORD):integer;
procedure getTiledElementByteOffset_2d_32(var outTiledByteOffset:QWORD;x,y,z:DWORD);
end;
Tiler1d=object
m_minGpuMode:DWORD;
m_tileMode:DWORD;
m_arrayMode:DWORD;
m_linearWidth:DWORD;
m_linearHeight:DWORD;
m_linearDepth:DWORD;
m_paddedWidth:DWORD;
m_paddedHeight:DWORD;
m_paddedDepth:DWORD;
m_bitsPerElement:DWORD;
m_linearSizeBytes:DWORD;
m_tiledSizeBytes:DWORD;
m_microTileMode:DWORD;
m_tileThickness:DWORD;
m_tileBytes:DWORD;
m_tilesPerRow:DWORD;
m_tilesPerSlice:DWORD;
function getTiledElementBitOffset(var outTiledBitOffset:QWORD;x,y,z:DWORD):integer;
end;
{
m_minGpuMode:1
m_tileMode:10
m_arrayMode:4
m_linearWidth:1920
m_linearHeight:1080
m_linearDepth:1
m_paddedWidth:1920
m_paddedHeight:1152
m_paddedDepth:1
m_bitsPerElement:32
m_linearSizeBytes:8294400
m_tiledSizeBytes:8847360
m_microTileMode:0
m_pipeConfig:18
m_arraySlice:0
m_numFragmentsPerPixel:1
m_bankWidth:1
m_bankHeight:2
m_numBanks:8
m_macroTileAspect:1
m_tileSplitBytes:512
m_numPipes:16
m_tileThickness:1
m_macroTileWidth:128
m_macroTileHeight:128
m_pipeInterleaveBytes:256
m_pipeInterleaveBits:8
m_pipeInterleaveMask:255
m_pipeBits:4
m_bankBits:3
m_pipeMask:3840
m_bankSwizzleMask:0
m_pipeSwizzleMask:0
}
const
Texture2d_32:Tiler1d=(
m_minGpuMode:0 ;
m_tileMode:13 ;
m_arrayMode:2 ;
m_linearWidth:8 ;
m_linearHeight:8 ;
m_linearDepth:1 ;
m_paddedWidth:8 ;
m_paddedHeight:8 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:256;
m_tiledSizeBytes:256 ;
m_microTileMode:1 ;
m_tileThickness:1 ;
m_tileBytes:256 ;
m_tilesPerRow:1 ;
m_tilesPerSlice:1 ;
);
Texture2d_8:Tiler1d=(
m_minGpuMode:0 ;
m_tileMode:13 ;
m_arrayMode:2 ;
m_linearWidth:8 ;
m_linearHeight:8 ;
m_linearDepth:1 ;
m_paddedWidth:32 ;
m_paddedHeight:8 ;
m_paddedDepth:1 ;
m_bitsPerElement:8 ;
m_linearSizeBytes:64;
m_tiledSizeBytes:256;
m_microTileMode:1 ;
m_tileThickness:1 ;
m_tileBytes:64 ;
m_tilesPerRow:4 ;
m_tilesPerSlice:4 ;
);
const
Tiler2d_1280_720_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1280 ;
m_linearHeight:720 ;
m_linearDepth:1 ;
m_paddedWidth:1280 ;
m_paddedHeight:768 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:3686400 ;
m_tiledSizeBytes:3932160 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1312_738_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1312 ;
m_linearHeight:738 ;
m_linearDepth:1 ;
m_paddedWidth:1408 ;
m_paddedHeight:768 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:3873024 ;
m_tiledSizeBytes:4325376 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1344_756_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1344 ;
m_linearHeight:756 ;
m_linearDepth:1 ;
m_paddedWidth:1408 ;
m_paddedHeight:768 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:4064256 ;
m_tiledSizeBytes:4325376 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1376_774_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1376 ;
m_linearHeight:774 ;
m_linearDepth:1 ;
m_paddedWidth:1408 ;
m_paddedHeight:832 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:4260096 ;
m_tiledSizeBytes:4685824 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1408_792_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1408 ;
m_linearHeight:792 ;
m_linearDepth:1 ;
m_paddedWidth:1408 ;
m_paddedHeight:832 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:4460544 ;
m_tiledSizeBytes:4685824 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1440_810_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1440 ;
m_linearHeight:810 ;
m_linearDepth:1 ;
m_paddedWidth:1536 ;
m_paddedHeight:832 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:4665600 ;
m_tiledSizeBytes:5111808 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1472_828_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1472 ;
m_linearHeight:828 ;
m_linearDepth:1 ;
m_paddedWidth:1536 ;
m_paddedHeight:832 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:4875264 ;
m_tiledSizeBytes:5111808 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1504_846_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1504 ;
m_linearHeight:846 ;
m_linearDepth:1 ;
m_paddedWidth:1536 ;
m_paddedHeight:896 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:5089536 ;
m_tiledSizeBytes:5505024 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1536_864_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1536 ;
m_linearHeight:864 ;
m_linearDepth:1 ;
m_paddedWidth:1536 ;
m_paddedHeight:896 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:5308416 ;
m_tiledSizeBytes:5505024 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1568_882_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1568 ;
m_linearHeight:882 ;
m_linearDepth:1 ;
m_paddedWidth:1664 ;
m_paddedHeight:896 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:5531904 ;
m_tiledSizeBytes:5963776 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1600_900_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1600 ;
m_linearHeight:900 ;
m_linearDepth:1 ;
m_paddedWidth:1664 ;
m_paddedHeight:960 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:5760000 ;
m_tiledSizeBytes:6389760 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1632_918_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1632 ;
m_linearHeight:918 ;
m_linearDepth:1 ;
m_paddedWidth:1664 ;
m_paddedHeight:960 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:5992704 ;
m_tiledSizeBytes:6389760 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1664_936_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1664 ;
m_linearHeight:936 ;
m_linearDepth:1 ;
m_paddedWidth:1664 ;
m_paddedHeight:960 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:6230016 ;
m_tiledSizeBytes:6389760 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1696_954_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1696 ;
m_linearHeight:954 ;
m_linearDepth:1 ;
m_paddedWidth:1792 ;
m_paddedHeight:960 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:6471936 ;
m_tiledSizeBytes:6881280 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1728_972_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1728 ;
m_linearHeight:972 ;
m_linearDepth:1 ;
m_paddedWidth:1792 ;
m_paddedHeight:1024 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:6718464 ;
m_tiledSizeBytes:7340032 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1760_990_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1760 ;
m_linearHeight:990 ;
m_linearDepth:1 ;
m_paddedWidth:1792 ;
m_paddedHeight:1024 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:6969600 ;
m_tiledSizeBytes:7340032 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1792_1008_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1792 ;
m_linearHeight:1008 ;
m_linearDepth:1 ;
m_paddedWidth:1792 ;
m_paddedHeight:1024 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:7225344 ;
m_tiledSizeBytes:7340032 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1824_1026_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1824 ;
m_linearHeight:1026 ;
m_linearDepth:1 ;
m_paddedWidth:1920 ;
m_paddedHeight:1088 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:7485696 ;
m_tiledSizeBytes:8355840 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1856_1044_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1856 ;
m_linearHeight:1044 ;
m_linearDepth:1 ;
m_paddedWidth:1920 ;
m_paddedHeight:1088 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:7750656 ;
m_tiledSizeBytes:8355840 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1888_1062_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1888 ;
m_linearHeight:1062 ;
m_linearDepth:1 ;
m_paddedWidth:1920 ;
m_paddedHeight:1088 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:8020224 ;
m_tiledSizeBytes:8355840 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_1920_1080_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:1920 ;
m_linearHeight:1080 ;
m_linearDepth:1 ;
m_paddedWidth:1920 ;
m_paddedHeight:1088 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:8294400 ;
m_tiledSizeBytes:8355840 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_2240_1260_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:2240 ;
m_linearHeight:1260 ;
m_linearDepth:1 ;
m_paddedWidth:2304 ;
m_paddedHeight:1280 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:11289600 ;
m_tiledSizeBytes:11796480 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_2560_1440_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:2560 ;
m_linearHeight:1440 ;
m_linearDepth:1 ;
m_paddedWidth:2560 ;
m_paddedHeight:1472 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:14745600 ;
m_tiledSizeBytes:15073280 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_2880_1620_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:2880 ;
m_linearHeight:1620 ;
m_linearDepth:1 ;
m_paddedWidth:2944 ;
m_paddedHeight:1664 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:18662400 ;
m_tiledSizeBytes:19595264 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_3200_1800_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:3200 ;
m_linearHeight:1800 ;
m_linearDepth:1 ;
m_paddedWidth:3200 ;
m_paddedHeight:1856 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:23040000 ;
m_tiledSizeBytes:23756800 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_3360_1890_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:3360 ;
m_linearHeight:1890 ;
m_linearDepth:1 ;
m_paddedWidth:3456 ;
m_paddedHeight:1920 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:25401600 ;
m_tiledSizeBytes:26542080 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_3520_1980_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:3520 ;
m_linearHeight:1980 ;
m_linearDepth:1 ;
m_paddedWidth:3584 ;
m_paddedHeight:1984 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:27878400 ;
m_tiledSizeBytes:28442624 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_3680_2070_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:3680 ;
m_linearHeight:2070 ;
m_linearDepth:1 ;
m_paddedWidth:3712 ;
m_paddedHeight:2112 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:30470400 ;
m_tiledSizeBytes:31358976 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
Tiler2d_3840_2160_32:Tiler2d=(
m_minGpuMode:0 ;
m_tileMode:10 ;
m_arrayMode:4 ;
m_linearWidth:3840 ;
m_linearHeight:2160 ;
m_linearDepth:1 ;
m_paddedWidth:3840 ;
m_paddedHeight:2176 ;
m_paddedDepth:1 ;
m_bitsPerElement:32 ;
m_linearSizeBytes:33177600 ;
m_tiledSizeBytes:33423360 ;
m_microTileMode:0 ;
m_pipeConfig:12 ;
m_arraySlice:0 ;
m_numFragmentsPerPixel:1 ;
m_bankWidth:1 ;
m_bankHeight:1 ;
m_numBanks:16 ;
m_macroTileAspect:2 ;
m_tileSplitBytes:512 ;
m_numPipes:8 ;
m_tileThickness:1 ;
m_macroTileWidth:128 ;
m_macroTileHeight:64 ;
m_pipeInterleaveBytes:256 ;
m_pipeInterleaveBits:8 ;
m_pipeInterleaveMask:255 ;
m_pipeBits:3 ;
m_bankBits:4 ;
m_pipeMask:1792 ;
m_bankSwizzleMask:0 ;
m_pipeSwizzleMask:0 ;
);
function GetTiler2d(Width,m_bitsPerElement:DWORD):Tiler2d;
procedure detile32bppDisplaySse2(dst,src:Pointer;destPitch:DWORD); assembler; MS_ABI_CDecl;
procedure detile32bppBuf(var T:Tiler2d;src,dst:Pointer);
function getMicroTileMode(outMicroTileMode:PByte;tmode:Byte):Integer;
Function computeSurfaceMacroTileMode(outMacroTileMode:PByte;tileMode,bitsPerElement,numFragmentsPerPixel:Byte):Integer;
implementation
function GetTiler2d(Width,m_bitsPerElement:DWORD):Tiler2d;
begin
Result:=Default(Tiler2d);
if (m_bitsPerElement<>32) then Exit;
Case Width of
1280:Result:=Tiler2d_1280_720_32;
1312:Result:=Tiler2d_1312_738_32;
1344:Result:=Tiler2d_1344_756_32;
1376:Result:=Tiler2d_1376_774_32;
1408:Result:=Tiler2d_1408_792_32;
1440:Result:=Tiler2d_1440_810_32;
1472:Result:=Tiler2d_1472_828_32;
1504:Result:=Tiler2d_1504_846_32;
1536:Result:=Tiler2d_1536_864_32;
1568:Result:=Tiler2d_1568_882_32;
1600:Result:=Tiler2d_1600_900_32;
1632:Result:=Tiler2d_1632_918_32;
1664:Result:=Tiler2d_1664_936_32;
1696:Result:=Tiler2d_1696_954_32;
1728:Result:=Tiler2d_1728_972_32;
1760:Result:=Tiler2d_1760_990_32;
1792:Result:=Tiler2d_1792_1008_32;
1824:Result:=Tiler2d_1824_1026_32;
1856:Result:=Tiler2d_1856_1044_32;
1888:Result:=Tiler2d_1888_1062_32;
1920:Result:=Tiler2d_1920_1080_32;
2240:Result:=Tiler2d_2240_1260_32;
2560:Result:=Tiler2d_2560_1440_32;
2880:Result:=Tiler2d_2880_1620_32;
3200:Result:=Tiler2d_3200_1800_32;
3360:Result:=Tiler2d_3360_1890_32;
3520:Result:=Tiler2d_3520_1980_32;
3680:Result:=Tiler2d_3680_2070_32;
3840:Result:=Tiler2d_3840_2160_32;
end;
end;
//Resolution (width x height)
//1280x720
//1312x738
//1344x756
//1376x774
//1408x792
//1440x810
//1472x828
//1504x846
//1536x864
//1568x882
//1600x900
//1632x918
//1664x936
//1696x954
//1728x972
//1760x990
//1792x1008
//1824x1026
//1856x1044
//1888x1062
//1920x1080
//2240x1260
//2560x1440
//2880x1620
//3200x1800
//3360x1890
//3520x1980
//3680x2070
//3840x2160
const
kCbColorBase = mmCB_COLOR0_BASE - mmCB_COLOR0_BASE;
kCbColorPitch = mmCB_COLOR0_PITCH - mmCB_COLOR0_BASE;
kCbColorSlice = mmCB_COLOR0_SLICE - mmCB_COLOR0_BASE;
kCbColorView = mmCB_COLOR0_VIEW - mmCB_COLOR0_BASE;
kCbColorInfo = mmCB_COLOR0_INFO - mmCB_COLOR0_BASE;
kCbColorAttrib = mmCB_COLOR0_ATTRIB - mmCB_COLOR0_BASE;
kCbColorDccControl = mmCB_COLOR0_DCC_CONTROL - mmCB_COLOR0_BASE;
kCbColorCmask = mmCB_COLOR0_CMASK - mmCB_COLOR0_BASE;
kCbColorCmaskSlice = mmCB_COLOR0_CMASK_SLICE - mmCB_COLOR0_BASE;
kCbColorFmask = mmCB_COLOR0_FMASK - mmCB_COLOR0_BASE;
kCbColorFmaskSlice = mmCB_COLOR0_FMASK_SLICE - mmCB_COLOR0_BASE;
kCbColorClearWord0 = mmCB_COLOR0_CLEAR_WORD0 - mmCB_COLOR0_BASE;
kCbColorClearWord1 = mmCB_COLOR0_CLEAR_WORD1 - mmCB_COLOR0_BASE;
kCbColorDccBase = mmCB_COLOR0_DCC_BASE - mmCB_COLOR0_BASE;
// 14: unused
kCbWidthHeight = 15; // not a GPU register. width in [15:0], height in [31:16].
const
g_bitsPerElement:array[0..60] of Integer=(
$0, $8, $10, $10,
$20, $20, $20, $20,
$20, $20, $20, $40,
$40, $60, $80, -1,
$10, $10, $10, $10,
$20, $20, $40, -1,
-1, -1, -1, -1,
-1, -1, -1, -1,
$10, $10, $20, $4,
$8, $8, $4, $8,
$8, $8, -1, -1,
$8, $8, $8, $8,
$8, $8, $10, $10,
$20, $20, $20, $40,
$40, $8, $10, $1,
$1);
function getTotalBitsPerElement(this:TDATA_FORMAT):Integer;
var
ret:Integer;
m_surfaceFormat:Byte;
begin
m_surfaceFormat:=this.m_surfaceFormat;
ret:=0;
if (m_surfaceFormat < $3d) then
begin
ret := $10;
if (6 < Byte(m_surfaceFormat - $23)) then
begin
if (Byte(m_surfaceFormat - $3b) < 2) then
begin
ret := 8;
end else
begin
ret := 1;
end;
end;
ret := ret * g_bitsPerElement[m_surfaceFormat];
end;
Result:=ret;
end;
function getTexelsPerElement(this:TDATA_FORMAT):DWORD;
var
ret:DWORD;
begin
ret := $10;
if (6 < byte(this.m_surfaceFormat - $23)) then
begin
if (byte(this.m_surfaceFormat - $3b) < 2) then
begin
Exit(8);
end;
ret := 1;
end;
Result:=ret;
end;
type
TTILE_MODE_REG=bitpacked record
RESERVED0 :bit2;
ARRAY_MODE :bit4; ///< Gnm::ArrayMode
PIPE_CONFIG :bit5; ///< Gnm::PipeConfig
TILE_SPLIT :bit3; ///< Gnm::TileSplit
RESERVED1 :bit8;
MICRO_TILE_MODE_NEW:bit3; ///< Gnm::MicroTileMode
SAMPLE_SPLIT :bit2; ///< Gnm::SampleSplit
ALT_PIPE_CONFIG :bit5; ///< NEO ONLY
end;
TMACRO_TILE_MODE_REG=bitpacked record
BANK_WIDTH :bit2; ///< Gnm::BankWidth
BANK_HEIGHT :bit2; ///< Gnm::BankHeight
MACRO_TILE_ASPECT :bit2; ///< Gnm::MacroTileAspect
NUM_BANKS :bit2; ///< Gnm::NumBanks
ALT_BANK_HEIGHT :bit2; ///< NEO ONLY
ALT_MACRO_TILE_ASPECT:bit2; ///< NEO ONLY
ALT_NUM_BANKS :bit2; ///< NEO ONLY
RESERVED0 :bit18;
end;
TTILE_MODE=packed record
Case Byte of
0:(B:TTILE_MODE_REG);
1:(D:DWORD);
end;
TMACRO_TILE_MODE=packed record
Case Byte of
0:(B:TMACRO_TILE_MODE_REG);
1:(D:DWORD);
end;
const
GB_TILE_MODE:array[0..31] of TTILE_MODE=(
(D:$90800310), // GB_TILE_MODE0 0x00 kTileModeDepth_2dThin_64 am=2dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Depth ss=1
(D:$90800B10), // GB_TILE_MODE1 0x01 kTileModeDepth_2dThin_128 am=2dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 128 mtm=Depth ss=1
(D:$90801310), // GB_TILE_MODE2 0x02 kTileModeDepth_2dThin_256 am=2dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 256 mtm=Depth ss=1
(D:$90801B10), // GB_TILE_MODE3 0x03 kTileModeDepth_2dThin_512 am=2dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 512 mtm=Depth ss=1
(D:$90802310), // GB_TILE_MODE4 0x04 kTileModeDepth_2dThin_1K am=2dTiledThin pipe/alt=P8_32x32_16x16/P16 ts=1024 mtm=Depth ss=1
(D:$90800308), // GB_TILE_MODE5 0x05 kTileModeDepth_1dThin am=1dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Depth ss=1
(D:$90801318), // GB_TILE_MODE6 0x06 kTileModeDepth_2dThinPrt_256 am=2dTiledThinPrt pipe/alt=P8_32x32_16x16/P16 ts= 256 mtm=Depth ss=1
(D:$90802318), // GB_TILE_MODE7 0x07 kTileModeDepth_2dThinPrt_1K am=2dTiledThinPrt pipe/alt=P8_32x32_16x16/P16 ts=1024 mtm=Depth ss=1
(D:$90000304), // GB_TILE_MODE8 0x08 kTileModeDisplay_LinearAligned am=LinearAligned pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Display ss=1
(D:$90000308), // GB_TILE_MODE9 0x09 kTileModeDisplay_1dThin am=1dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Display ss=1
(D:$92000310), // GB_TILE_MODE10 0x0A kTileModeDisplay_2dThin am=2dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Display ss=2
(D:$92000294), // GB_TILE_MODE11 0x0B kTileModeDisplay_ThinPrt am=TiledThinPrt pipe/alt=P8_32x32_8x16 /P16 ts= 64 mtm=Display ss=2
(D:$92000318), // GB_TILE_MODE12 0x0C kTileModeDisplay_2dThinPrt am=2dTiledThinPrt pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Display ss=2
(D:$90400308), // GB_TILE_MODE13 0x0D kTileModeThin_1dThin am=1dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Thin ss=1
(D:$92400310), // GB_TILE_MODE14 0x0E kTileModeThin_2dThin am=2dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Thin ss=2
(D:$924002B0), // GB_TILE_MODE15 0x0F kTileModeThin_3dThin am=3dTiledThin pipe/alt=P8_32x32_8x16 /P16 ts= 64 mtm=Thin ss=2
(D:$92400294), // GB_TILE_MODE16 0x10 kTileModeThin_ThinPrt am=TiledThinPrt pipe/alt=P8_32x32_8x16 /P16 ts= 64 mtm=Thin ss=2
(D:$92400318), // GB_TILE_MODE17 0x11 kTileModeThin_2dThinPrt am=2dTiledThinPrt pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Thin ss=2
(D:$9240032C), // GB_TILE_MODE18 0x12 kTileModeThin_3dThinPrt am=3dTiledThinPrt pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Thin ss=2
(D:$9100030C), // GB_TILE_MODE19 0x13 kTileModeThick_1dThick am=1dTiledThick pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Thick ss=1
(D:$9100031C), // GB_TILE_MODE20 0x14 kTileModeThick_2dThick am=2dTiledThick pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Thick ss=1
(D:$910002B4), // GB_TILE_MODE21 0x15 kTileModeThick_3dThick am=3dTiledThick pipe/alt=P8_32x32_8x16 /P16 ts= 64 mtm=Thick ss=1
(D:$910002A4), // GB_TILE_MODE22 0x16 kTileModeThick_ThickPrt am=TiledThickPrt pipe/alt=P8_32x32_8x16 /P16 ts= 64 mtm=Thick ss=1
(D:$91000328), // GB_TILE_MODE23 0x17 kTileModeThick_2dThickPrt am=2dTiledThickPrt pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Thick ss=1
(D:$910002BC), // GB_TILE_MODE24 0x18 kTileModeThick_3dThickPrt am=3dTiledThickPrt pipe/alt=P8_32x32_8x16 /P16 ts= 64 mtm=Thick ss=1
(D:$91000320), // GB_TILE_MODE25 0x19 kTileModeThick_2dXThick am=2dTiledXThick pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Thick ss=1
(D:$910002B8), // GB_TILE_MODE26 0x1A kTileModeThick_3dXThick am=3dTiledXThick pipe/alt=P8_32x32_8x16 /P16 ts= 64 mtm=Thick ss=1
(D:$90C00308), // GB_TILE_MODE27 0x1B kTileModeRotated_1dThin am=1dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Rotated ss=1
(D:$92C00310), // GB_TILE_MODE28 0x1C kTileModeRotated_2dThin am=2dTiledThin pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Rotated ss=2
(D:$92C00294), // GB_TILE_MODE29 0x1D kTileModeRotated_ThinPrt am=TiledThinPrt pipe/alt=P8_32x32_8x16 /P16 ts= 64 mtm=Rotated ss=2
(D:$92C00318), // GB_TILE_MODE30 0x1E kTileModeRotated_2dThinPrt am=2dTiledThinPrt pipe/alt=P8_32x32_16x16/P16 ts= 64 mtm=Rotated ss=2
(D:$00000000) // GB_TILE_MODE31 0x1F kTileModeDisplay_LinearGeneral am=LinearGeneral pipe/alt=P2 / P2 ts= 64 mtm=Display ss=1
);
GB_MACROTILE_MODE:array[0..15] of TMACRO_TILE_MODE=(
(D:$26E8), // GB_MACROTILE_MODE0 0x00 kMacroTileMode_1x4_16 bankWidth=1 bankHeight=4 macroTileAspect=4 numBanks=16 altBankHeight=4 altNumBanks= 8 altMacroTileAspect=2
(D:$26D4), // GB_MACROTILE_MODE1 0x01 kMacroTileMode_1x2_16 bankWidth=1 bankHeight=2 macroTileAspect=2 numBanks=16 altBankHeight=4 altNumBanks= 8 altMacroTileAspect=2
(D:$21D0), // GB_MACROTILE_MODE2 0x02 kMacroTileMode_1x1_16 bankWidth=1 bankHeight=1 macroTileAspect=2 numBanks=16 altBankHeight=2 altNumBanks= 8 altMacroTileAspect=1
(D:$21D0), // GB_MACROTILE_MODE3 0x03 kMacroTileMode_1x1_16_dup bankWidth=1 bankHeight=1 macroTileAspect=2 numBanks=16 altBankHeight=2 altNumBanks= 8 altMacroTileAspect=1
(D:$2080), // GB_MACROTILE_MODE4 0x04 kMacroTileMode_1x1_8 bankWidth=1 bankHeight=1 macroTileAspect=1 numBanks= 8 altBankHeight=1 altNumBanks= 8 altMacroTileAspect=1
(D:$2040), // GB_MACROTILE_MODE5 0x05 kMacroTileMode_1x1_4 bankWidth=1 bankHeight=1 macroTileAspect=1 numBanks= 4 altBankHeight=1 altNumBanks= 8 altMacroTileAspect=1
(D:$1000), // GB_MACROTILE_MODE6 0x06 kMacroTileMode_1x1_2 bankWidth=1 bankHeight=1 macroTileAspect=1 numBanks= 2 altBankHeight=1 altNumBanks= 4 altMacroTileAspect=1
(D:$0000), // GB_MACROTILE_MODE7 0x07 kMacroTileMode_1x1_2_dup bankWidth=1 bankHeight=1 macroTileAspect=1 numBanks= 2 altBankHeight=1 altNumBanks= 2 altMacroTileAspect=1
(D:$36EC), // GB_MACROTILE_MODE8 0x08 kMacroTileMode_1x8_16 bankWidth=1 bankHeight=8 macroTileAspect=4 numBanks=16 altBankHeight=4 altNumBanks=16 altMacroTileAspect=2
(D:$26E8), // GB_MACROTILE_MODE9 0x09 kMacroTileMode_1x4_16_dup bankWidth=1 bankHeight=4 macroTileAspect=4 numBanks=16 altBankHeight=4 altNumBanks= 8 altMacroTileAspect=2
(D:$21D4), // GB_MACROTILE_MODE10 0x0A kMacroTileMode_1x2_16_dup bankWidth=1 bankHeight=2 macroTileAspect=2 numBanks=16 altBankHeight=2 altNumBanks= 8 altMacroTileAspect=1
(D:$20D0), // GB_MACROTILE_MODE11 0x0B kMacroTileMode_1x1_16_dup2 bankWidth=1 bankHeight=1 macroTileAspect=2 numBanks=16 altBankHeight=1 altNumBanks= 8 altMacroTileAspect=1
(D:$1080), // GB_MACROTILE_MODE12 0x0C kMacroTileMode_1x1_8_dup bankWidth=1 bankHeight=1 macroTileAspect=1 numBanks= 8 altBankHeight=1 altNumBanks= 4 altMacroTileAspect=1
(D:$1040), // GB_MACROTILE_MODE13 0x0D kMacroTileMode_1x1_4_dup bankWidth=1 bankHeight=1 macroTileAspect=1 numBanks= 4 altBankHeight=1 altNumBanks= 4 altMacroTileAspect=1
(D:$0000), // GB_MACROTILE_MODE14 0x0E kMacroTileMode_1x1_2_dup2 bankWidth=1 bankHeight=1 macroTileAspect=1 numBanks= 2 altBankHeight=1 altNumBanks= 2 altMacroTileAspect=1
(D:$0000) // GB_MACROTILE_MODE15 0x0F kMacroTileMode_1x1_2_dup3 bankWidth=1 bankHeight=1 macroTileAspect=1 numBanks= 2 altBankHeight=1 altNumBanks= 2 altMacroTileAspect=1
);
function getArrayMode(outArrayMode:PByte;tmode:Byte):Integer;
begin
Result:=-$7f2d0000;
if ((outArrayMode<>nil) and (tmode<$20)) then
begin
outArrayMode^:=GB_TILE_MODE[tmode].B.ARRAY_MODE;
Result:=0;
end;
end;
function getMicroTileMode(outMicroTileMode:PByte;tmode:Byte):Integer;
begin
Result:=-$7f2d0000;
if ((outMicroTileMode<>nil) and (tmode<$20)) then
begin
outMicroTileMode^:=GB_TILE_MODE[tmode].B.MICRO_TILE_MODE_NEW;
Result:=0;
end;
end;
function getSampleSplit(outSampleSplit:PByte;tmode:Byte):Integer;
begin
Result:=-$7f2d0000;
if ((outSampleSplit<>nil) and (tmode<$20)) then
begin
outSampleSplit^:=GB_TILE_MODE[tmode].B.SAMPLE_SPLIT;
Result:=0;
end;
end;
function getTileSplit(outTileSplit:PByte;tmode:Byte):Integer;
begin
Result:=-$7f2d0000;
if ((outTileSplit<>nil) and (tmode<$20)) then
begin
outTileSplit^:=GB_TILE_MODE[tmode].B.TILE_SPLIT;
Result:=0;
end;
end;
function TRENDER_TARGET.getTileMode:Byte; inline;
begin
Result:=ATTRIB.TILE_MODE_INDEX;
end;
function TRENDER_TARGET.getWidth:WORD; inline;
begin
Result:=Width;
end;
function TRENDER_TARGET.getHeight:WORD; inline;
begin
Result:=Height;
end;
function TRENDER_TARGET.getMinimumGpuMode:Byte; inline;
begin
Result:=INFO.RESERVED0;
end;
function TRENDER_TARGET.getNumFragments:Byte; inline;
begin
Result:=ATTRIB.NUM_FRAGMENTS;
end;
function TRENDER_TARGET.getPitchDiv8Minus1:Word; inline;
begin
Result:=PITCH.TILE_MAX;
end;
function TRENDER_TARGET.getPitch:DWORD; inline;
begin
Result:=(getPitchDiv8Minus1+1)*8;
end;
function TRENDER_TARGET.getDccCompressionEnable:Boolean; inline;
begin
Result:=INFO.DCC_ENABLE<>0;
end;
const
chY_1:array[0..2] of Integer=(5,0,4);
chX_1:array[0..2] of Integer=(4,4,5);
chW_1:array[0..2] of Integer=(1,5,1);
chY_2:array[0..2] of Integer=(5,5,6);
chZ_2:array[0..2] of Integer=(6,4,5);
chX_2:array[0..2] of Integer=(4,6,7);
chW_2:array[0..2] of Integer=(7,7,4);
chZ_3:array[0..2] of Integer=(6,0,4);
chX_3:array[0..2] of Integer=(4,4,6);
chW_3:array[0..2] of Integer=(1,6,1);
chZ_4:array[0..2] of Integer=(0,0,4);
chY_4:array[0..2] of Integer=(0,4,0);
chX_4:array[0..2] of Integer=(4,0,0);
function sce_Gnm_DataFormat_build(FORMAT,NUMBER_TYPE,COMP_SWAP:Byte):TDATA_FORMAT;
var
m_channelX:Integer;
m_channelY:Integer;
m_channelZ:Integer;
m_channelW:Integer;
m_channelType:Integer;
IS_SWAP_ALT_REV:Boolean;
label
_end,_zero;
begin
if (($1800004000000016 shr (FORMAT and $3f) and 1)=0) then
begin
if (($3fff08000700828 shr (FORMAT and $3f) and 1)<>0) then
begin
IS_SWAP_ALT_REV:=COMP_SWAP=3;
if (COMP_SWAP<3) then
begin
m_channelY:=chY_1[COMP_SWAP];
m_channelX:=chX_1[COMP_SWAP];
m_channelW:=chW_1[COMP_SWAP];
m_channelZ:=0;
end else
begin
m_channelY:=0;
m_channelX:=ord(IS_SWAP_ALT_REV)+ord(IS_SWAP_ALT_REV)*4;
m_channelW:=ord(IS_SWAP_ALT_REV)*3+1;
m_channelZ:=0;
end;
goto _end;
end;
if (($4000107000120c0 shr (FORMAT and $3f) and 1)=0) then
begin
if (($238000e5700 shr (FORMAT and $3f) and 1)=0) then
begin
m_channelW:=1;
goto _zero;
end;
if (COMP_SWAP<3) then
begin
m_channelY:=chY_2[COMP_SWAP];
m_channelZ:=chZ_2[COMP_SWAP];
m_channelX:=chX_2[COMP_SWAP];
m_channelW:=chW_2[COMP_SWAP];
goto _end;
end;
IS_SWAP_ALT_REV:=COMP_SWAP=3;
m_channelX:=ord(IS_SWAP_ALT_REV)*5;
m_channelY:=6;
m_channelZ:=7;
if (not IS_SWAP_ALT_REV) then
begin
m_channelZ:=0;
m_channelY:=0;
end;
end else
begin
if (COMP_SWAP<3) then
begin
m_channelZ:=chZ_3[COMP_SWAP];
m_channelX:=chX_3[COMP_SWAP];
m_channelW:=chW_3[COMP_SWAP];
m_channelY:=5;
goto _end;
end;
m_channelX:=6;
m_channelY:=ord(COMP_SWAP=3)*5;
m_channelZ:=0;
if (COMP_SWAP<>3) then
begin
m_channelX:=0;
end;
end;
m_channelW:=ord(COMP_SWAP=3)*3+1;
end else
begin
if (COMP_SWAP<3) then
begin
m_channelW:=1;
m_channelZ:=chZ_4[COMP_SWAP];
m_channelY:=chY_4[COMP_SWAP];
m_channelX:=chX_4[COMP_SWAP];
goto _end;
end;
m_channelW:=ord(COMP_SWAP=3)*3+1;
_zero:
m_channelZ:=0;
m_channelY:=0;
m_channelX:=0;
end;
_end:
m_channelType:=$9;
if (NUMBER_TYPE<>6) then
begin
m_channelType:=(NUMBER_TYPE and $F);
end;
Result.m_surfaceFormat:=FORMAT;
Result.m_channelType :=m_channelType;
Result.m_channelX :=m_channelX;
Result.m_channelY :=m_channelY;
Result.m_channelZ :=m_channelZ;
Result.m_channelW :=m_channelW;
Result.m_unused :=0;
end;
function TRENDER_TARGET.getDataFormat:TDATA_FORMAT;
begin
Result:=sce_Gnm_DataFormat_build(INFO.FORMAT,INFO.NUMBER_TYPE,INFO.COMP_SWAP);
end;
function isMacroTiled(tileMode:Byte):Boolean; inline;
begin
Result:=($7f7dcdf shr (tileMode and $3f) and 1)<>0;
end;
function isPartiallyResidentTexture(arrayMode:Byte):Boolean; inline;
begin
Result:=($8e60 shr (arrayMode and $3f) and 1)<>0;
end;
function isPowerOfTwo(x:DWORD):Boolean;
begin
if (x<>0) then
begin
Result:=((x-1) and x)=0;
end else
begin
Result:=False;
end;
end;
function getMicroTileThickness(arrayMode:Byte):Byte;
begin
Case arrayMode of
kArrayMode1dTiledThick ,
kArrayMode2dTiledThick ,
kArrayMode3dTiledThick ,
kArrayModeTiledThickPrt ,
kArrayMode2dTiledThickPrt,
kArrayMode3dTiledThickPrt:Result:=4;
kArrayMode2dTiledXThick ,
kArrayMode3dTiledXThick :Result:=8;
kArrayModeLinearGeneral ,
kArrayModeLinearAligned ,
kArrayMode1dTiledThin ,
kArrayMode2dTiledThin ,
kArrayModeTiledThinPrt ,
kArrayMode2dTiledThinPrt ,
kArrayMode3dTiledThinPrt ,
kArrayMode3dTiledThin :Result:=1;
else
Result:=0;
end;
end;
Function computeSurfaceMacroTileMode(outMacroTileMode:PByte;tileMode,bitsPerElement,numFragmentsPerPixel:Byte):Integer;
var
NumIsPower1:Boolean;
_MacroTiled:Boolean;
NumIsPower2:Boolean;
is_Thick_pow:Boolean;
isParRes:Boolean;
uVar1:DWORD;
uVar2:DWORD;
Thick:DWORD;
_TileSplit:Byte;
_SampleSplit:Byte;
_MicroTileMode:Byte;
_MicroTileThick:Byte;
_ArrayMode:Byte;
begin
Result:=-$7f2d0000;
if (outMacroTileMode <> nil) then
begin
NumIsPower1 := isPowerOfTwo(numFragmentsPerPixel);
if ((numFragmentsPerPixel < 9) and (NumIsPower1 <> false)) then
begin
Result := getArrayMode(@_ArrayMode,tileMode);
if (Result = 0) then
begin
_MacroTiled := isMacroTiled(_ArrayMode);
Result := -$7f2d0000;
if ((bitsPerElement - 1 < $80) and (_MacroTiled <> false)) then
begin
NumIsPower2 := isPowerOfTwo(numFragmentsPerPixel);
if ((numFragmentsPerPixel - 1 < $10) and (NumIsPower2 <> false)) then
begin
Result := getMicroTileMode(@_MicroTileMode,tileMode);
if (Result = 0) then
begin
Result := getSampleSplit(@_SampleSplit,tileMode);
if (Result = 0) then
begin
Result := getTileSplit(@_TileSplit,tileMode);
if (Result = 0) then
begin
_MicroTileThick := getMicroTileThickness(_ArrayMode);
Thick := bitsPerElement * _MicroTileThick * $40 shr 3;
uVar1 := Thick shl (_SampleSplit and $1f);
uVar2 := $100;
if ($100 < uVar1) then
begin
uVar2 := uVar1;
end;
uVar1 := ($40 shl (_TileSplit and $3f));
if (_MicroTileMode <> 2) then
begin
uVar1 := uVar2;
end;
uVar2 := $400;
if (uVar1 < $400) then
begin
uVar2 := uVar1;
end;
Thick := Thick * numFragmentsPerPixel;
if (uVar2 <= Thick) then
begin
Thick := uVar2;
end;
is_Thick_pow := isPowerOfTwo(Thick);
Result := -$7f2d0000;
if ((is_Thick_pow <> false) and (Thick - $40 < $fc1)) then
begin
uVar2 := 0;
Thick:=(Thick shr 6);
While ((Thick and $80000000) = 0) do
begin
uVar2 := uVar2 + 1;
Thick := (Thick or 1) shl 1;
end;
isParRes := isPartiallyResidentTexture(_ArrayMode);
Thick := (uVar2 xor $1f) + 8;
if (isParRes = false) then
begin
Thick := uVar2 xor $1f;
end;
Result := 0;
outMacroTileMode^ := Thick;
end;
end;
end;
end;
end;
end;
end;
end;
end;
end;
Function getAltNumBanks(outAltNumBanks:PByte;tileMode,bitsPerElement,numFragmentsPerPixel:Byte):Integer;
var
_MacroTileMode:Byte;
begin
Result:=-$7f2d0000;
if (outAltNumBanks<>nil) then
begin
Result := computeSurfaceMacroTileMode(@_MacroTileMode,tileMode,bitsPerElement,numFragmentsPerPixel);
if (Result = 0) then
begin
outAltNumBanks^:=GB_MACROTILE_MODE[_MacroTileMode].B.ALT_NUM_BANKS;
end;
end;
end;
Function getNumBanks(outNumBanks:PByte;tileMode,bitsPerElement,numFragmentsPerPixel:Byte):Integer;
var
_MacroTileMode:Byte;
begin
Result:=-$7f2d0000;
if (outNumBanks<>nil) then
begin
Result := computeSurfaceMacroTileMode(@_MacroTileMode,tileMode,bitsPerElement,numFragmentsPerPixel);
if (Result = 0) then
begin
outNumBanks^:=GB_MACROTILE_MODE[_MacroTileMode].B.NUM_BANKS
end;
end;
end;
function TRENDER_TARGET.getTileSwizzleMask:Byte;
var
_isMacroTiled:Boolean;
dataFormat:TDATA_FORMAT;
m_bitsPerFragment:Byte;
sVar1:Integer;
_NumBanks:Byte;
begin
if (Integer(INFO) < 0) then
begin
dataFormat:=getDataFormat;
if (dataFormat.m_surfaceFormat <> 0) then
begin
_isMacroTiled:=isMacroTiled(ATTRIB.TILE_MODE_INDEX);
if (_isMacroTiled <> false) then
begin
m_bitsPerFragment:=getTotalBitsPerElement(dataFormat);
if (Integer(INFO) < 0) then
begin
getAltNumBanks(@_NumBanks,ATTRIB.TILE_MODE_INDEX,m_bitsPerFragment,ATTRIB.NUM_FRAGMENTS);
sVar1:=4;
end else
begin
getNumBanks(@_NumBanks,ATTRIB.TILE_MODE_INDEX,m_bitsPerFragment,ATTRIB.NUM_FRAGMENTS);
sVar1:=3;
end;
Result := ((((1 shl ((_NumBanks + 1) and $1f)) -1) shl sVar1) and BASE) shr 4;
Exit;
end;
end;
end;
Result:=0;
end;
function getFlagsForSurfaceType(minGpuMode:Byte;outFlags:PSurfaceFlags;surfaceType:Byte;mipmapMode:Byte):Integer;
begin
if (outFlags=nil) then Exit(-$7f2d0000);
DWORD(outFlags^):=0;
case surfaceType of
kSurfaceTypeColorTargetDisplayable:;
kSurfaceTypeColorTarget:;
kSurfaceTypeDepthTarget:
begin
outFlags^.m_depthTarget :=1;
outFlags^.m_stencilTarget:=1;
end;
kSurfaceTypeDepthOnlyTarget:
begin
outFlags^.m_depthTarget:=1;
end;
kSurfaceTypeStencilOnlyTarget:
begin
outFlags^.m_stencilTarget:=1;
end;
kSurfaceTypeFmaskBuffer:
begin
outFlags^.m_fmask:=1;
end;
kSurfaceTypeTextureFlat,
kSurfaceTypeRwTextureFlat:
begin
outFlags^.m_pow2Pad := (mipmapMode and 1); // Must be set for textures w/mipmaps.
outFlags^.m_texCompatible:= (minGpuMode and 1);
end;
kSurfaceTypeTextureVolume,
kSurfaceTypeRwTextureVolume:
begin
outFlags^.m_volume:= 1;
outFlags^.m_pow2Pad := (mipmapMode and 1); // Must be set for textures w/mipmaps.
outFlags^.m_texCompatible:= (minGpuMode and 1);
end;
kSurfaceTypeTextureCubemap,
kSurfaceTypeRwTextureCubemap:
begin
outFlags^.m_cube:= 1;
outFlags^.m_pow2Pad := (mipmapMode and 1); // Must be set for textures w/mipmaps.
outFlags^.m_texCompatible:= (minGpuMode and 1);
end;
else
Exit(-$7f2d0000);
end;
Result:=0;
end;
{
int32_t sce::GpuAddress::TilingParameters::initFromTexture(const Gnm::Texture *texture, uint32_t mipLevel, uint32_t arraySlice)
SCE_GNM_ASSERT_MSG_RETURN(texture != 0, kStatusInvalidArgument, "texture must not be NULL.");
SCE_GNM_ASSERT_MSG_RETURN(mipLevel <= texture->getLastMipLevel(), kStatusInvalidArgument, "mipLevel (%u) is out of range for texture; last level is %u", mipLevel, texture->getLastMipLevel());
bool isCubemap = (texture->getTextureType() == Gnm::kTextureTypeCubemap);
bool isVolume = (texture->getTextureType() == Gnm::kTextureType3d);
// Building surface flags manually is error-prone, but we don't know exactly what type of texture this is.
m_surfaceFlags.m_value = 0;
Gnm::MicroTileMode microTileMode;
int32_t status = getMicroTileMode(&microTileMode, texture->getTileMode());
if (status != kStatusSuccess)
return status;
m_surfaceFlags.m_depthTarget = (!isVolume && (microTileMode == Gnm::kMicroTileModeDepth) && (texture->getDataFormat().getZFormat() != Gnm::kZFormatInvalid)) ? 1 : 0;
m_surfaceFlags.m_stencilTarget = (!isVolume && (microTileMode == Gnm::kMicroTileModeDepth) && (texture->getDataFormat().getStencilFormat() != Gnm::kStencilInvalid)) ? 1 : 0;
m_surfaceFlags.m_cube = isCubemap ? 1 : 0;
m_surfaceFlags.m_volume = isVolume ? 1 : 0;
m_surfaceFlags.m_pow2Pad = texture->isPaddedToPow2() ? 1 : 0;
if (texture->getMinimumGpuMode() == Gnm::kGpuModeNeo)
{
m_surfaceFlags.m_texCompatible = 1;
}
m_tileMode = texture->getTileMode(); // see below, though
m_minGpuMode = texture->getMinimumGpuMode();
Gnm::DataFormat dataFormat = texture->getDataFormat();
m_bitsPerFragment = dataFormat.getTotalBitsPerElement() / dataFormat.getTexelsPerElement();
m_isBlockCompressed = (dataFormat.getTexelsPerElement() > 1);
m_tileSwizzleMask = texture->getTileSwizzleMask();
m_linearWidth = std::max(texture->getWidth() >> mipLevel, 1U);
m_linearHeight = std::max(texture->getHeight() >> mipLevel, 1U);
m_linearDepth = m_surfaceFlags.m_volume ? std::max(texture->getDepth() >> mipLevel, 1U) : 1;
m_numFragmentsPerPixel = 1 << texture->getNumFragments();
m_baseTiledPitch = texture->getPitch();
m_mipLevel = mipLevel;
SCE_GNM_ASSERT_MSG_RETURN(arraySlice == 0 || !m_surfaceFlags.m_volume, kStatusInvalidArgument, "for volume textures, arraySlice must be 0."); // volume textures can't be arrays
uint32_t arraySliceCount = texture->getTotalArraySliceCount();
if (isCubemap)
arraySliceCount *= 6; // Cube maps store 6 faces per array slice
else if (isVolume)
arraySliceCount = 1;
if (texture->isPaddedToPow2())
arraySliceCount = nextPowerOfTwo(arraySliceCount); // array slice counts are padded to a power of two as well
SCE_GNM_ASSERT_MSG_RETURN(arraySlice < arraySliceCount, kStatusInvalidArgument, "arraySlice (%u) is out of range for texture (0x%p) with %u slices.", arraySlice, texture, arraySliceCount);
m_arraySlice = arraySlice;
// Use computeSurfaceInfo() to determine what array mode we REALLY need to use, since it's occasionally not the one the Texture uses.
// (e.g. for a 2D-tiled texture, the smaller mip levels will implicitly use a 1D array mode to cut down on wasted padding space)
SurfaceInfo surfInfoOut = {0};
status = computeSurfaceInfo(&surfInfoOut, this);
if (status != kStatusSuccess)
return status;
status = adjustTileMode(m_minGpuMode, &m_tileMode, m_tileMode, surfInfoOut.m_arrayMode);
if (status != kStatusSuccess)
return status;
return kStatusSuccess;
}
function TilingParameters.initFromRenderTarget(var target:TRENDER_TARGET;arraySlice:DWORD):Integer;
var
dataFormat:TDATA_FORMAT;
status:Integer;
maxUncompressedBlockSize:Integer;
st:Byte;
microTileMode:Byte;
independentDccBlocks:Boolean;
isDccEnabled:Boolean;
begin
m_tileMode:=target.getTileMode(); // see below, though
m_minGpuMode:=target.getMinimumGpuMode();
dataFormat:=target.getDataFormat;
m_bitsPerFragment:=getTotalBitsPerElement(dataFormat) div getTexelsPerElement(dataFormat);
m_isBlockCompressed:=(getTexelsPerElement(dataFormat)>1);
m_tileSwizzleMask:= target.getTileSwizzleMask();
m_linearWidth :=target.getWidth();
m_linearHeight:=target.getHeight();
m_linearDepth :=1;
m_numFragmentsPerPixel:=1 shl target.getNumFragments();
m_baseTiledPitch:=target.getPitch();
m_mipLevel :=0; // unused for render targets
m_arraySlice:=arraySlice;
status:=getMicroTileMode(@microTileMode,target.getTileMode);
if (status<>0) then Exit(status);
if (microTileMode=kMicroTileModeDisplay) then
st:=kSurfaceTypeColorTargetDisplayable
else
st:=kSurfaceTypeColorTarget;
status:=getFlagsForSurfaceType(m_minGpuMode,@m_surfaceFlags,st,kSurfaceMipmapDisable);
if (status<>0) then Exit(status);
if (m_minGpuMode=kGpuModeNeo) then
begin
independentDccBlocks :=target.DCC_CONTROL.INDEPENDENT_64B_BLOCKS<>0;
maxUncompressedBlockSize:=target.DCC_CONTROL.MAX_COMPRESSED_BLOCK_SIZE;
isDccEnabled:=target.getDccCompressionEnable();
if (isDccEnabled and (not independentDccBlocks or (maxUncompressedBlockSize>kDccBlockSize64))) then
m_surfaceFlags.m_texCompatible:=0
else
m_surfaceFlags.m_texCompatible:=1;
end;
//
Result:=0;
end;
//int32_t sce::GpuAddress::TilingParameters::initFromRenderTarget(const Gnm::RenderTarget *target, uint32_t arraySlice)
//{
// SurfaceInfo surfInfoOut = {0};
// status = computeSurfaceInfo(&surfInfoOut, this);
// if (status != kStatusSuccess)
// return status;
// status = adjustTileMode(m_minGpuMode, &m_tileMode, m_tileMode, surfInfoOut.m_arrayMode);
// if (status != kStatusSuccess)
// return status;
// return kStatusSuccess;
//}
function TilingParameters.initFromRenderTargetSpec(var target:RenderTargetSpec;arraySlice:DWORD):Integer;
var
status:Integer;
maxUncompressedBlockSize:Integer;
dFormat:TDATA_FORMAT;
microTileMode:Byte;
st:Byte;
independentDccBlocks:Boolean;
isDccEnabled:Boolean;
begin
m_tileMode := target.m_colorTileModeHint; // see below, though
m_minGpuMode := target.m_minGpuMode;
dFormat := target.m_colorFormat;
m_bitsPerFragment := getTotalBitsPerElement(dFormat) div getTexelsPerElement(dFormat);
m_isBlockCompressed := (getTexelsPerElement(dFormat) > 1);
m_tileSwizzleMask := target.m_regs.getTileSwizzleMask();
m_linearWidth := target.m_width;
m_linearHeight := target.m_height;
m_linearDepth := 1;
m_numFragmentsPerPixel := 1 shl target.m_numFragments;
m_baseTiledPitch := target.m_pitch;
m_mipLevel := 0; // unused for render targets
m_arraySlice := arraySlice;
status:=getMicroTileMode(@microTileMode,target.m_regs.getTileMode);
if (status<>0) then Exit(status);
if (microTileMode=kMicroTileModeDisplay) then
st:=kSurfaceTypeColorTargetDisplayable
else
st:=kSurfaceTypeColorTarget;
status:=getFlagsForSurfaceType(m_minGpuMode,@m_surfaceFlags,st,kSurfaceMipmapDisable);
if (status<>0) then Exit(status);
if (m_minGpuMode=kGpuModeNeo) then
begin
independentDccBlocks :=target.m_regs.DCC_CONTROL.INDEPENDENT_64B_BLOCKS<>0;
maxUncompressedBlockSize:=target.m_regs.DCC_CONTROL.MAX_COMPRESSED_BLOCK_SIZE;
isDccEnabled:=target.m_regs.getDccCompressionEnable();
if (isDccEnabled and (not independentDccBlocks or (maxUncompressedBlockSize>kDccBlockSize64))) then
m_surfaceFlags.m_texCompatible:=0
else
m_surfaceFlags.m_texCompatible:=1;
end;
{
// Use computeSurfaceInfo() to determine what array mode we REALLY need to use, since it's occasionally not the one the Texture uses.
// (e.g. for a 2D-tiled texture, the smaller mip levels will implicitly use a 1D array mode to cut down on wasted padding space)
SurfaceInfo surfInfoOut = {0};
status = computeSurfaceInfo(&surfInfoOut, this);
if (status != kStatusSuccess)
return status;
status = adjustTileMode(m_minGpuMode, &m_tileMode, m_tileMode, surfInfoOut.m_arrayMode);
if (status != kStatusSuccess)
return status;
return kStatusSuccess;}
Result:=0;
end;
function getElementIndex(x,y,z,bitsPerElement,microTileMode,arrayMode:DWORD):DWORD;
var
elem:DWORD;
begin
elem:=0;
case microTileMode of
kMicroTileModeDisplay:
begin
case bitsPerElement of
8:
begin
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (x shr 1) and $1 ) shl 1;
elem:=elem or ( (x shr 2) and $1 ) shl 2;
elem:=elem or ( (y shr 1) and $1 ) shl 3;
elem:=elem or ( (y shr 0) and $1 ) shl 4;
elem:=elem or ( (y shr 2) and $1 ) shl 5;
end;
16:
begin
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (x shr 1) and $1 ) shl 1;
elem:=elem or ( (x shr 2) and $1 ) shl 2;
elem:=elem or ( (y shr 0) and $1 ) shl 3;
elem:=elem or ( (y shr 1) and $1 ) shl 4;
elem:=elem or ( (y shr 2) and $1 ) shl 5;
end;
32:
begin
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (x shr 1) and $1 ) shl 1;
elem:=elem or ( (y shr 0) and $1 ) shl 2;
elem:=elem or ( (x shr 2) and $1 ) shl 3;
elem:=elem or ( (y shr 1) and $1 ) shl 4;
elem:=elem or ( (y shr 2) and $1 ) shl 5;
end;
64:
begin
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (y shr 0) and $1 ) shl 1;
elem:=elem or ( (x shr 1) and $1 ) shl 2;
elem:=elem or ( (x shr 2) and $1 ) shl 3;
elem:=elem or ( (y shr 1) and $1 ) shl 4;
elem:=elem or ( (y shr 2) and $1 ) shl 5;
end;
else
Assert(false,'Unsupported bitsPerElement (%u) for displayable surface.');
end;
end;
kMicroTileModeThin,
kMicroTileModeDepth:
begin
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (y shr 0) and $1 ) shl 1;
elem:=elem or ( (x shr 1) and $1 ) shl 2;
elem:=elem or ( (y shr 1) and $1 ) shl 3;
elem:=elem or ( (x shr 2) and $1 ) shl 4;
elem:=elem or ( (y shr 2) and $1 ) shl 5;
case arrayMode of
kArrayMode2dTiledXThick,
kArrayMode3dTiledXThick:
begin
elem:=elem or ( (z shr 2) and $1 ) shl 8;
end;
kArrayMode1dTiledThick,
kArrayMode2dTiledThick,
kArrayMode3dTiledThick,
kArrayModeTiledThickPrt,
kArrayMode2dTiledThickPrt,
kArrayMode3dTiledThickPrt:
begin
elem:=elem or ( (z shr 0) and $1 ) shl 6;
elem:=elem or ( (z shr 1) and $1 ) shl 7;
end;
end;
end;
kMicroTileModeThick:
begin
case arrayMode of
kArrayMode2dTiledXThick,
kArrayMode3dTiledXThick:
begin
elem:=elem or ( (z shr 2) and $1 ) shl 8;
end;
kArrayMode1dTiledThick,
kArrayMode2dTiledThick,
kArrayMode3dTiledThick,
kArrayModeTiledThickPrt,
kArrayMode2dTiledThickPrt,
kArrayMode3dTiledThickPrt:
case bitsPerElement of
8,16:
begin
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (y shr 0) and $1 ) shl 1;
elem:=elem or ( (x shr 1) and $1 ) shl 2;
elem:=elem or ( (y shr 1) and $1 ) shl 3;
elem:=elem or ( (z shr 0) and $1 ) shl 4;
elem:=elem or ( (z shr 1) and $1 ) shl 5;
elem:=elem or ( (x shr 2) and $1 ) shl 6;
elem:=elem or ( (y shr 2) and $1 ) shl 7;
end;
32:
begin
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (y shr 0) and $1 ) shl 1;
elem:=elem or ( (x shr 1) and $1 ) shl 2;
elem:=elem or ( (z shr 0) and $1 ) shl 3;
elem:=elem or ( (y shr 1) and $1 ) shl 4;
elem:=elem or ( (z shr 1) and $1 ) shl 5;
elem:=elem or ( (x shr 2) and $1 ) shl 6;
elem:=elem or ( (y shr 2) and $1 ) shl 7;
end;
64,128:
begin
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (y shr 0) and $1 ) shl 1;
elem:=elem or ( (z shr 0) and $1 ) shl 2;
elem:=elem or ( (x shr 1) and $1 ) shl 3;
elem:=elem or ( (y shr 1) and $1 ) shl 4;
elem:=elem or ( (z shr 1) and $1 ) shl 5;
elem:=elem or ( (x shr 2) and $1 ) shl 6;
elem:=elem or ( (y shr 2) and $1 ) shl 7;
end;
else
Assert(false,'Invalid bitsPerElement (%u) for microTileMode=kMicroTileModeThick.');
end;
else
Assert(false,'Invalid arrayMode (0x%02X) for thick/xthick microTileMode=kMicroTileModeThick.');
end;
end;
end;
Result:=elem;
end;
function getPipeIndex(x,y,pipeCfg:DWORD):DWORD;
var
pipe:DWORD;
begin
pipe:= 0;
case pipeCfg of
kPipeConfigP8_32x32_8x16:
begin
pipe:=pipe or ( ((x shr 4) xor (y shr 3) xor (x shr 5)) and $1 ) shl 0;
pipe:=pipe or ( ((x shr 3) xor (y shr 4)) and $1 ) shl 1;
pipe:=pipe or ( ((x shr 5) xor (y shr 5)) and $1 ) shl 2;
end;
kPipeConfigP8_32x32_16x16:
begin
pipe:=pipe or ( ((x shr 3) xor (y shr 3) xor (x shr 4)) and $1 ) shl 0;
pipe:=pipe or ( ((x shr 4) xor (y shr 4)) and $1 ) shl 1;
pipe:=pipe or ( ((x shr 5) xor (y shr 5)) and $1 ) shl 2;
end;
kPipeConfigP16:
begin
pipe:=pipe or ( ((x shr 3) xor (y shr 3) xor (x shr 4)) and $1 ) shl 0;
pipe:=pipe or ( ((x shr 4) xor (y shr 4)) and $1 ) shl 1;
pipe:=pipe or ( ((x shr 5) xor (y shr 5)) and $1 ) shl 2;
pipe:=pipe or ( ((x shr 6) xor (y shr 5)) and $1 ) shl 3;
end;
else
Assert(false,'Unsupported pipeCfg (0x%02X).');
end;
Result:=pipe;
end;
function fastIntLog2(i:DWORD):DWORD; inline;
begin
Result:=BsrDWord(i or 1);
end;
function getBankIndex(x,y,bank_width,bank_height,num_banks,num_pipes:DWORD):DWORD;
var
x_shift_offset,
y_shift_offset,
xs,ys:DWORD;
bank:DWORD;
begin
x_shift_offset := fastIntLog2(bank_width * num_pipes);
y_shift_offset := fastIntLog2(bank_height);
xs := x shr x_shift_offset;
ys := y shr y_shift_offset;
bank:= 0;
case num_banks of
2:
begin
bank :=bank or ( ((xs shr 3) xor (ys shr 3)) and $1 ) shl 0;
end;
4:
begin
bank :=bank or ( ((xs shr 3) xor (ys shr 4)) and $1 ) shl 0;
bank :=bank or ( ((xs shr 4) xor (ys shr 3)) and $1 ) shl 1;
end;
8:
begin
bank :=bank or ( ((xs shr 3) xor (ys shr 5)) and $1 ) shl 0;
bank :=bank or ( ((xs shr 4) xor (ys shr 4) xor (ys shr 5)) and $1 ) shl 1;
bank :=bank or ( ((xs shr 5) xor (ys shr 3)) and $1 ) shl 2;
end;
16:
begin
bank :=bank or ( ((xs shr 3) xor (ys shr 6)) and $1 ) shl 0;
bank :=bank or ( ((xs shr 4) xor (ys shr 5) xor (ys shr 6)) and $1 ) shl 1;
bank :=bank or ( ((xs shr 5) xor (ys shr 4)) and $1 ) shl 2;
bank :=bank or ( ((xs shr 6) xor (ys shr 3)) and $1 ) shl 3;
end;
else
Assert(false,'invalid num_banks (%u) -- must be 2, 4, 8, or 16.');
end;
Result:=bank;
end;
function Tiler1d.getTiledElementBitOffset(var outTiledBitOffset:QWORD;x,y,z:DWORD):integer;
var
element_index:QWORD;
slice_offset:QWORD;
tile_row_index:QWORD;
tile_column_index:QWORD;
tile_offset:QWORD;
element_offset:QWORD;
final_offset:QWORD;
begin
element_index := getElementIndex(x, y, z, m_bitsPerElement, m_microTileMode, m_arrayMode);
slice_offset := (z div m_tileThickness) * m_tilesPerSlice * m_tileBytes;
tile_row_index := y div kMicroTileHeight;
tile_column_index := x div kMicroTileWidth;
tile_offset := ((tile_row_index * m_tilesPerRow) + tile_column_index) * m_tileBytes;
element_offset := element_index * m_bitsPerElement;
final_offset := (slice_offset + tile_offset)*8 + element_offset;
outTiledBitOffset := final_offset;
Result:=0;
end;
function Tiler2d.init(var tp:TilingParameters):integer;
begin
if @tp=nil then Exit(-$7f2d0000);
//// Use gpu_addr to come up with actual legal/padded surface parameters
//SurfaceInfo surfInfoOut = {0};
//int32_t status = computeSurfaceInfo(&surfInfoOut, tp);
//SCE_GNM_ASSERT_MSG_RETURN(status == kStatusSuccess, status, "computeSurfaceInfo() failed: %d", status);
// derived inputs
m_minGpuMode:=tp.m_minGpuMode;
m_tileMode :=tp.m_tileMode;
//m_arrayMode :=surfInfoOut.m_arrayMode;
getArrayMode(@m_arrayMode,m_tileMode);
getMicroTileMode(@m_microTileMode,m_tileMode);
//// other constants
Case m_arrayMode of
kArrayMode2dTiledThin,
kArrayMode3dTiledThin,
kArrayModeTiledThinPrt,
kArrayMode2dTiledThinPrt,
kArrayMode3dTiledThinPrt:
m_tileThickness:=1;
kArrayMode2dTiledThick,
kArrayMode3dTiledThick,
kArrayModeTiledThickPrt,
kArrayMode2dTiledThickPrt,
kArrayMode3dTiledThickPrt:
m_tileThickness:=4;
kArrayMode2dTiledXThick,
kArrayMode3dTiledXThick:
m_tileThickness:=8;
end;
m_linearWidth :=tp.m_linearWidth; // unpadded
m_linearHeight :=tp.m_linearHeight; // unpadded
m_linearDepth :=tp.m_linearDepth; // unpadded
m_bitsPerElement:=tp.m_bitsPerFragment;
//m_paddedWidth :=surfInfoOut.m_pitch; // padded
//m_paddedHeight :=surfInfoOut.m_height;
//m_paddedDepth :=surfInfoOut.m_depth;
m_numFragmentsPerPixel:=tp.m_numFragmentsPerPixel;
end;
{
// For multi-texel-per-element formats, each block is treated as an element for tiling purposes.
// This affects a few of the above variables.
if (tp->m_isBlockCompressed)
{
switch(tp->m_bitsPerFragment)
{
case 1:
SCE_GNM_ASSERT_MSG_RETURN(m_microTileMode == Gnm::kMicroTileModeDisplay, kStatusInvalidArgument, "1bpp surfaces must use Gnm::kMicroTileModeDisplay");
m_bitsPerElement *= 8;
m_linearWidth = std::max((m_linearWidth+7)/8, 1U);
m_paddedWidth = std::max((m_paddedWidth+7)/8, 1U);
break;
case 4:
case 8:
m_bitsPerElement *= 16;
m_linearWidth = std::max((m_linearWidth+3)/4, 1U);
m_linearHeight = std::max((m_linearHeight+3)/4, 1U);
m_paddedWidth = std::max((m_paddedWidth+3)/4, 1U);
m_paddedHeight = std::max((m_paddedHeight+3)/4, 1U);
break;
case 16:
// TODO
break;
default:
SCE_GNM_ASSERT_MSG_RETURN(!tp->m_isBlockCompressed, kStatusInvalidArgument, "Unknown bit depth %u for block-compressed format", m_bitsPerElement);
break;
}
}
m_linearSizeBytes = (m_linearWidth * m_linearHeight * m_linearDepth * m_bitsPerElement * m_numFragmentsPerPixel + 7) / 8;
m_tiledSizeBytes = surfInfoOut.m_surfaceSize;
Gnm::BankWidth bankWidthHW;
Gnm::BankHeight bankHeightHW;
Gnm::MacroTileAspect macroAspectHW;
Gnm::NumBanks numBanksHW;
if (tp->m_minGpuMode == Gnm::kGpuModeNeo)
{
getAltPipeConfig(&m_pipeConfig, m_tileMode);
getAllAltMacroTileData(m_tileMode, m_bitsPerElement, m_numFragmentsPerPixel,
&bankWidthHW, &bankHeightHW, &macroAspectHW, &numBanksHW);
}
else
{
getPipeConfig(&m_pipeConfig, m_tileMode);
getAllMacroTileData(m_tileMode, m_bitsPerElement, m_numFragmentsPerPixel,
&bankWidthHW, &bankHeightHW, &macroAspectHW, &numBanksHW);
}
m_bankWidth = 1 << bankWidthHW;
m_bankHeight = 1 << bankHeightHW;
m_numBanks = 2 << numBanksHW;
m_macroTileAspect = 1 << macroAspectHW;
uint32_t tileBytes1x = (m_tileThickness*m_bitsPerElement*kMicroTileWidth*kMicroTileHeight + 7)/8;
Gnm::SampleSplit sampleSplitHw;
Gnm::TileSplit tileSplitHw;
GpuAddress::getSampleSplit(&sampleSplitHw, tp->m_tileMode);
GpuAddress::getTileSplit(&tileSplitHw, tp->m_tileMode);
uint32_t sampleSplit = 1 << sampleSplitHw;
uint32_t tileSplitC = (m_microTileMode == Gnm::kMicroTileModeDepth)
? (64 << tileSplitHw) // depth modes store tile split directly
: std::max(256U, tileBytes1x*sampleSplit); // other modes store a sample split multiplier
m_tileSplitBytes = std::min(kDramRowSize, tileSplitC);
// Hardware constants -- see GB_ADDR_CONFIG register
m_pipeInterleaveBytes = kPipeInterleaveBytes;
m_numPipes = getPipeCount(m_pipeConfig);
m_pipeInterleaveBits = fastIntLog2(m_pipeInterleaveBytes);
m_pipeInterleaveMask = (1 << (m_pipeInterleaveBits)) - 1;
m_pipeBits = fastIntLog2(m_numPipes);
m_bankBits = fastIntLog2(m_numBanks);
m_pipeMask = (m_numPipes-1) << m_pipeInterleaveBits;
m_bankSwizzleMask = tp->m_tileSwizzleMask;
m_pipeSwizzleMask = 0; // not currently used
m_macroTileWidth = (kMicroTileWidth * m_bankWidth * m_numPipes) * m_macroTileAspect;
m_macroTileHeight = (kMicroTileHeight * m_bankHeight * m_numBanks) / m_macroTileAspect;
m_arraySlice = tp->m_arraySlice;
// Verify 2D tiled addressing restrictions
// These restrictions should be addressed by the computeSurfaceInfo() function. If any of these
// asserts fire, it probably means computeSurfaceInfo() isn't doing its job correctly.
SCE_GNM_ASSERT_MSG_RETURN(m_paddedWidth % m_macroTileWidth == 0, kStatusInternalTilingError, "internal consistency check failed.");
SCE_GNM_ASSERT_MSG_RETURN(m_paddedHeight % m_macroTileHeight == 0, kStatusInternalTilingError, "internal consistency check failed.");
SCE_GNM_ASSERT_MSG_RETURN(m_numBanks * m_numPipes >= 4, kStatusInternalTilingError, "internal consistency check failed.");
return kStatusSuccess;
}
function Tiler2d.getTiledElementBitOffset(var outTiledBitOffset:QWORD;x,y,z,fragmentIndex:DWORD):integer;
var
element_index,xh,yh:DWORD;
tile_bytes:DWORD;
slice:DWORD;
pipe,bank:QWORD;
element_offset:QWORD;
pixel_offset:QWORD;
fragment_offset:QWORD;
slices_per_tile:QWORD;
tile_split_slice:QWORD;
macro_tile_bytes:QWORD;
macro_tiles_per_row:QWORD;
macro_tile_row_index:QWORD;
macro_tile_column_index:QWORD;
macro_tile_index:QWORD;
macro_tile_offset:QWORD;
macro_tiles_per_slice:QWORD;
slice_bytes:QWORD;
slice_offset:QWORD;
tile_row_index:QWORD;
tile_column_index:QWORD;
tile_index:QWORD;
tile_offset:QWORD;
bank_swizzle:QWORD;
pipe_swizzle:QWORD;
pipe_slice_rotation:QWORD;
slice_rotation:QWORD;
tile_split_slice_rotation:QWORD;
total_offset:QWORD;
bitOffset:QWORD;
pipe_interleave_offset:QWORD;
offset:QWORD;
finalByteOffset:QWORD;
begin
element_index:=getElementIndex(x, y, z, m_bitsPerElement, m_microTileMode, m_arrayMode);
xh := x;
yh := y;
if (m_arrayMode=kArrayModeTiledThinPrt) or (m_arrayMode = kArrayModeTiledThickPrt) then
begin
xh := xh mod m_macroTileWidth;
yh := yh mod m_macroTileHeight;
end;
pipe := getPipeIndex(xh, yh, m_pipeConfig);
bank := getBankIndex(xh, yh, m_bankWidth, m_bankHeight, m_numBanks, m_numPipes);
tile_bytes := (kMicroTileWidth * kMicroTileHeight * m_tileThickness * m_bitsPerElement * m_numFragmentsPerPixel + 7) div 8;
element_offset:=0;
if (m_microTileMode=kMicroTileModeDepth) then
begin
pixel_offset := element_index * m_bitsPerElement * m_numFragmentsPerPixel;
element_offset := pixel_offset + (fragmentIndex * m_bitsPerElement);
end else
begin
fragment_offset := fragmentIndex * (tile_bytes div m_numFragmentsPerPixel) * 8;
element_offset := fragment_offset + (element_index * m_bitsPerElement);
end;
slices_per_tile := 1;
tile_split_slice := 0;
if (tile_bytes > m_tileSplitBytes) and (m_tileThickness = 1) then
begin
slices_per_tile := tile_bytes div m_tileSplitBytes;
tile_split_slice := element_offset div (m_tileSplitBytes*8);
element_offset:=element_offset mod (m_tileSplitBytes*8);
tile_bytes := m_tileSplitBytes;
end;
macro_tile_bytes := (m_macroTileWidth div kMicroTileWidth) * (m_macroTileHeight div kMicroTileHeight) * tile_bytes div (m_numPipes * m_numBanks);
macro_tiles_per_row := m_paddedWidth div m_macroTileWidth;
macro_tile_row_index := y div m_macroTileHeight;
macro_tile_column_index := x div m_macroTileWidth;
macro_tile_index := (macro_tile_row_index * macro_tiles_per_row) + macro_tile_column_index;
macro_tile_offset := macro_tile_index * macro_tile_bytes;
macro_tiles_per_slice := macro_tiles_per_row * (m_paddedHeight div m_macroTileHeight);
slice_bytes := macro_tiles_per_slice * macro_tile_bytes;
slice := z;
slice_offset := (tile_split_slice + slices_per_tile * slice div m_tileThickness) * slice_bytes;
if (m_arraySlice<>0) then slice := m_arraySlice;
tile_row_index := (y div kMicroTileHeight) mod m_bankHeight;
tile_column_index := ((x div kMicroTileWidth) div m_numPipes) mod m_bankWidth;
tile_index := (tile_row_index * m_bankWidth) + tile_column_index;
tile_offset := tile_index * tile_bytes;
bank_swizzle := m_bankSwizzleMask;
pipe_swizzle := m_pipeSwizzleMask;
pipe_slice_rotation:=0;
case m_arrayMode of
kArrayMode3dTiledThin,
kArrayMode3dTiledThick,
kArrayMode3dTiledXThick:
begin
pipe_slice_rotation := max(1, (m_numPipes div 2)-1) * (slice div m_tileThickness);
end;
end;
pipe_swizzle:=pipe_swizzle+pipe_slice_rotation;
pipe_swizzle:=pipe_swizzle and (m_numPipes - 1);
pipe := pipe xor pipe_swizzle;
slice_rotation:=0;
case m_arrayMode of
kArrayMode2dTiledThin,
kArrayMode2dTiledThick,
kArrayMode2dTiledXThick:
begin
slice_rotation := ((m_numBanks div 2)-1) * (slice div m_tileThickness);
end;
kArrayMode3dTiledThin,
kArrayMode3dTiledThick,
kArrayMode3dTiledXThick:
begin
slice_rotation := max(1, (m_numPipes div 2)-1) * (slice div m_tileThickness) div m_numPipes;
end;
end;
tile_split_slice_rotation:= 0;
case m_arrayMode of
kArrayMode2dTiledThin,
kArrayMode3dTiledThin,
kArrayMode2dTiledThinPrt,
kArrayMode3dTiledThinPrt:
begin
tile_split_slice_rotation := ((m_numBanks div 2)+1) * tile_split_slice;
end;
end;
bank:=bank xor bank_swizzle + slice_rotation;
bank:=bank xor tile_split_slice_rotation;
bank:=bank and (m_numBanks - 1);
total_offset := (slice_offset + macro_tile_offset + tile_offset)*8 + element_offset;
bitOffset := total_offset and $7;
total_offset:=total_offset div 8;
pipe_interleave_offset := total_offset and m_pipeInterleaveMask;
offset := total_offset shr m_pipeInterleaveBits;
finalByteOffset := pipe_interleave_offset or
(pipe shl (m_pipeInterleaveBits)) or
(bank shl (m_pipeInterleaveBits + m_pipeBits)) or
(offset shl (m_pipeInterleaveBits + m_pipeBits + m_bankBits));
outTiledBitOffset := (finalByteOffset shl 3) or bitOffset;
Result:=0;
end;
////////
procedure Tiler2d.getTiledElementByteOffset_2d_32(var outTiledByteOffset:QWORD;x,y,z:DWORD);
var
element_index:DWORD;
pipe,bank:QWORD;
macro_tiles_per_row:QWORD;
macro_tile_row_index:QWORD;
macro_tile_column_index:QWORD;
macro_tile_index:QWORD;
macro_tiles_per_slice:QWORD;
tile_row_index:QWORD;
tile_column_index:QWORD;
tile_index:QWORD;
total_offset:QWORD;
offset:QWORD;
function getBankIndex16(x,y:DWORD):DWORD; inline;
begin
x := x shr 3;
Result:= 0;
Result:=Result or ( ((x shr 3) xor (y shr 6)) and $1 ) shl 0;
Result:=Result or ( ((x shr 4) xor (y shr 5) xor (y shr 6)) and $1 ) shl 1;
Result:=Result or ( ((x shr 5) xor (y shr 4)) and $1 ) shl 2;
Result:=Result or ( ((x shr 6) xor (y shr 3)) and $1 ) shl 3;
end;
begin
element_index:=0;
element_index:=element_index or ( (x shr 0) and $1 ) shl 0;
element_index:=element_index or ( (x shr 1) and $1 ) shl 1;
element_index:=element_index or ( (y shr 0) and $1 ) shl 2;
element_index:=element_index or ( (x shr 2) and $1 ) shl 3;
element_index:=element_index or ( (y shr 1) and $1 ) shl 4;
element_index:=element_index or ( (y shr 2) and $1 ) shl 5;
pipe:=0;
pipe:=pipe or ( ((x shr 3) xor (y shr 3) xor (x shr 4)) and $1 ) shl 0;
pipe:=pipe or ( ((x shr 4) xor (y shr 4)) and $1 ) shl 1;
pipe:=pipe or ( ((x shr 5) xor (y shr 5)) and $1 ) shl 2;
bank := getBankIndex16(x,y);
macro_tiles_per_row := m_paddedWidth div 128;
macro_tile_row_index := y div 64;
macro_tile_column_index := x div 128;
macro_tile_index := (macro_tile_row_index * macro_tiles_per_row) + macro_tile_column_index;
macro_tiles_per_slice := macro_tiles_per_row * (m_paddedHeight div 64);
tile_row_index := (y div 8) mod 1;
tile_column_index := ((x div 8) div 8) mod 1;
tile_index := tile_row_index + tile_column_index;
bank:=(bank+7*z) and 15;
total_offset:=((z*macro_tiles_per_slice)+macro_tile_index+tile_index)*256+(element_index*4);
offset := total_offset shr 8;
outTiledByteOffset := (total_offset and 255) or
(pipe shl (8)) or
(bank shl (11)) or
(offset shl (15));
end;
{
function fastIntLog2(i:DWORD):DWORD; inline;
begin
Result:=BsrDWord(i or 1);
end;
function getElementIndex(x,y:DWORD):DWORD;
var
elem:DWORD;
begin
elem:=0;
elem:=elem or ( (x shr 0) and $1 ) shl 0;
elem:=elem or ( (y shr 0) and $1 ) shl 2;
elem:=elem or ( (x shr 1) and $1 ) shl 1;
elem:=elem or ( (x shr 2) and $1 ) shl 3;
elem:=elem or ( (y shr 1) and $1 ) shl 4;
elem:=elem or ( (y shr 2) and $1 ) shl 5;
Result:=elem;
end;
function getPipeIndex(x,y:DWORD):DWORD;
var
pipe:DWORD;
begin
pipe:= 0;
pipe:=pipe or ( ((x shr 3) xor (y shr 3) xor (x shr 4)) and $1 ) shl 0;
pipe:=pipe or ( ((x shr 4) xor (y shr 4)) and $1 ) shl 1;
pipe:=pipe or ( ((x shr 5) xor (y shr 5)) and $1 ) shl 2;
pipe:=pipe or ( ((x shr 6) xor (y shr 5)) and $1 ) shl 3;
Result:=pipe;
end;
function getBankIndex(x,y:DWORD):DWORD;
var
xs,ys:DWORD;
bank:DWORD;
begin
xs := (x shr 4);
ys := (y shr 1);
bank:= 0;
bank :=bank or ( ((xs shr 3) xor (ys shr 5)) and $1 ) shl 0;
bank :=bank or ( ((xs shr 4) xor (ys shr 4) xor (ys shr 5)) and $1 ) shl 1;
bank :=bank or ( ((xs shr 5) xor (ys shr 3)) and $1 ) shl 2;
Result:=bank;
end;
const
m_paddedWidth=1920;
m_paddedHeight=1152;
function getTiledElementBitOffset(var outTiledBitOffset:QWORD;x,y:DWORD):integer;
var
element_index,xh,yh:DWORD;
pipe,bank:QWORD;
//element_offset:QWORD;
//macro_tiles_per_row:QWORD;
//macro_tile_row_index:QWORD;
//macro_tile_column_index:QWORD;
//macro_tile_index:QWORD;
macro_tile_offset:QWORD;
//macro_tiles_per_slice:QWORD;
//slice_offset:QWORD;
tile_row_index:QWORD;
tile_column_index:QWORD;
tile_index:QWORD;
//tile_offset:QWORD;
total_offset:QWORD;
bitOffset:QWORD;
pipe_interleave_offset:QWORD;
offset:QWORD;
finalByteOffset:QWORD;
begin
element_index:=getElementIndex(x, y);
xh := x;
yh := y;
pipe := getPipeIndex(xh, yh);
bank := getBankIndex(xh, yh);
//element_offset := (element_index * 32);
//macro_tiles_per_row := (m_paddedWidth div 128);
//macro_tile_row_index := (y div 128);
//macro_tile_column_index := (x div 128);
//macro_tile_index := ((y div 128) * (m_paddedWidth div 128)) + (x div 128);
macro_tile_offset := (((y div 128) * (m_paddedWidth div 128)) + (x div 128)) shl 9;
//macro_tiles_per_slice := (m_paddedWidth div 128) * (m_paddedHeight div 128);
tile_row_index := (y div 8) mod 2;
tile_column_index := (x div 128) mod 1;
tile_index := tile_row_index + tile_column_index;
//tile_offset := (tile_index * 256);
bank:=0;
total_offset := (((((y div 128) * (m_paddedWidth div 128)) + (x div 128)) shl 9) + (tile_index shl 8)) + (element_index shl 2);
//bitOffset := total_offset and $7;
//total_offset:=total_offset div 8;
pipe_interleave_offset := total_offset and 255;
offset := total_offset shr 8;
finalByteOffset := pipe_interleave_offset or
(pipe shl (8)) or
(bank shl (12)) or
(offset shl (15));
outTiledBitOffset := (finalByteOffset shl 3) or bitOffset;
Result:=0;
end;
}
//xorl %r8d, %r8d 3
//xorl %edx, %edx 2
//xorl %ecx, %ecx 1
procedure detile32bppDisplaySse2(dst,src:Pointer;destPitch:DWORD); assembler; nostackframe; MS_ABI_CDecl;
asm
//subq $40, %rsp //unsafe
//movaps %xmm6, (%rsp)
//movaps %xmm7, 16(%rsp)
movdqa 16(%rdx), %xmm5
movdqa 32(%rdx), %xmm6
movdqa 48(%rdx), %xmm4
movdqa 64(%rdx), %xmm3
movdqa 80(%rdx), %xmm1
leal (%r8,%r8,2), %eax
movdqa 96(%rdx), %xmm2
leal 0(,%r8,8), %r9d
sall $2, %eax
movdqa 112(%rdx), %xmm0
leal 0(,%r8,4), %r10d
sall $4, %r8d
movdqa (%rdx), %xmm7
movups %xmm6, 16(%rcx)
movups %xmm7, (%rcx)
movups %xmm5, (%rcx,%r10)
movups %xmm4, 16(%rcx,%r10)
movups %xmm3, (%rcx,%r9)
movups %xmm2, 16(%rcx,%r9)
movups %xmm1, (%rcx,%rax)
movups %xmm0, 16(%rcx,%rax)
movdqa 128(%rdx), %xmm7
addq %r8, %rcx
movdqa 144(%rdx), %xmm5
movdqa 160(%rdx), %xmm6
movdqa 176(%rdx), %xmm4
movdqa 192(%rdx), %xmm3
movdqa 208(%rdx), %xmm1
movdqa 224(%rdx), %xmm2
movdqa 240(%rdx), %xmm0
movups %xmm7, (%rcx)
movups %xmm6, 16(%rcx)
movups %xmm5, (%rcx,%r10)
movups %xmm4, 16(%rcx,%r10)
movups %xmm3, (%rcx,%r9)
movups %xmm2, 16(%rcx,%r9)
movups %xmm1, (%rcx,%rax)
movups %xmm0, 16(%rcx,%rax)
//movaps (%rsp), %xmm6
//movaps 16(%rsp), %xmm7
//addq $40, %rsp
end;
procedure detile32bppDisplayAvx(dst,src:Pointer;destPitch:DWORD); assembler; nostackframe; MS_ABI_CDecl;
asm
vmovdqa 32(%rdx), %ymm2
vmovdqa 64(%rdx), %ymm1
vmovdqa 96(%rdx), %ymm0
vmovaps (%rdx), %ymm3
leal (%r8,%r8,2), %eax
leal 0(,%r8,8), %r9d
sall $2, %eax
vmovups %xmm3, (%rcx)
leal 0(,%r8,4), %r10d
sall $4, %r8d
vextractf128 $0x1, %ymm3, (%rcx,%r10)
vmovups %xmm2, 16(%rcx)
vextractf128 $0x1, %ymm2, 16(%rcx,%r10)
vmovups %xmm1, (%rcx,%r9)
vextractf128 $0x1, %ymm1, (%rcx,%rax)
vmovups %xmm0, 16(%rcx,%r9)
vextractf128 $0x1, %ymm0, 16(%rcx,%rax)
vmovdqa 160(%rdx), %ymm2
addq %r8, %rcx
vmovdqa 192(%rdx), %ymm1
vmovdqa 224(%rdx), %ymm0
vmovaps 128(%rdx), %ymm3
vmovups %xmm3, (%rcx)
vextractf128 $0x1, %ymm3, (%rcx,%r10)
vmovups %xmm2, 16(%rcx)
vextractf128 $0x1, %ymm2, 16(%rcx,%r10)
vmovups %xmm1, (%rcx,%r9)
vextractf128 $0x1, %ymm1, (%rcx,%rax)
vmovups %xmm0, 16(%rcx,%r9)
vextractf128 $0x1, %ymm0, 16(%rcx,%rax)
//vzeroupper
end;
//xorl %r8 , %r8 3 destPitch
//xorl %rdx, %rdx 2 src
//xorl %rcx, %rcx 1 dst
//[3] ymm0 = ymm4 [7]
//[2] ymm1 = ymm5 [6]
//[1] ymm2 = ymm6 [5]
//[0] ymm3 = ymm7 [4]
procedure detile32bppDisplayAvx_cached(dst,src:Pointer;destPitch:DWORD); assembler; nostackframe; MS_ABI_CDecl;
asm
leal (%r8,%r8,2), %eax
leal 0(,%r8,8), %r9d
sall $2, %eax
vmovups %xmm3, (%rcx)
leal 0(,%r8,4), %r10d
sall $4, %r8d
vextractf128 $0x1, %ymm3, (%rcx,%r10)
vmovups %xmm2, 16(%rcx)
vextractf128 $0x1, %ymm2, 16(%rcx,%r10)
vmovups %xmm1, (%rcx,%r9)
vextractf128 $0x1, %ymm1, (%rcx,%rax)
vmovups %xmm0, 16(%rcx,%r9)
vextractf128 $0x1, %ymm0, 16(%rcx,%rax)
addq %r8, %rcx
vmovups %xmm7, (%rcx)
vextractf128 $0x1, %ymm7, (%rcx,%r10)
vmovups %xmm6, 16(%rcx)
vextractf128 $0x1, %ymm6, 16(%rcx,%r10)
vmovups %xmm5, (%rcx,%r9)
vextractf128 $0x1, %ymm5, (%rcx,%rax)
vmovups %xmm4, 16(%rcx,%r9)
vextractf128 $0x1, %ymm4, 16(%rcx,%rax)
end;
procedure move64_sse(dst,src:Pointer); assembler; nostackframe; MS_ABI_CDecl;
asm
movdqa 0(%rdx), %xmm0
movdqa 16(%rdx), %xmm1
movdqa 32(%rdx), %xmm2
movdqa 48(%rdx), %xmm3
movdqa %xmm0, 0(%rcx)
movdqa %xmm1, 16(%rcx)
movdqa %xmm2, 32(%rcx)
movdqa %xmm3, 48(%rcx)
end;
procedure move64_avx(dst,src:Pointer); assembler; nostackframe; MS_ABI_CDecl;
asm
vmovdqa 0(%rdx), %ymm0
vmovdqa 32(%rdx), %ymm1
vmovdqa %ymm0, 0(%rcx)
vmovdqa %ymm1, 32(%rcx)
end;
//[3] ymm0 = ymm4 [7] 11 15
//[2] ymm1 = ymm5 [6] 10 14
//[1] ymm2 = ymm6 [5] 9 13
//[0] ymm3 = ymm7 [4] 8 12
//xorl %r8 , %r8 3 destPitch
//xorl %rdx, %rdx 2 src
//xorl %rcx, %rcx 1 dst
procedure move64_avx_cached(dst,src:Pointer;id:Byte); MS_ABI_CDecl;
begin
Case id of
0:asm
vmovdqa 0(%rdx), %ymm3
vmovdqa 32(%rdx), %ymm2
end;
1:asm
vmovdqa 0(%rdx), %ymm1
vmovdqa 32(%rdx), %ymm0
end;
2:asm
vmovdqa 0(%rdx), %ymm7
vmovdqa 32(%rdx), %ymm6
end;
3:asm
vmovdqa 0(%rdx), %ymm5
vmovdqa 32(%rdx), %ymm4
end;
4:asm
vmovdqa 0(%rdx), %ymm11
vmovdqa 32(%rdx), %ymm10
end;
5:asm
vmovdqa 0(%rdx), %ymm9
vmovdqa 32(%rdx), %ymm8
end;
6:asm
vmovdqa 0(%rdx), %ymm15
vmovdqa 32(%rdx), %ymm14
end;
7..15:
begin
src:=src+(id-7)*64;
asm
vmovdqa 0(%rdx), %ymm13
vmovdqa 32(%rdx), %ymm12
vmovdqa %ymm13, 0(%rcx)
vmovdqa %ymm12, 32(%rcx)
end;
end;
end;
end;
//vmovaps 0(%rdx), %ymm3
//vmovdqa 32(%rdx), %ymm2
//vmovdqa 64(%rdx), %ymm1
//vmovdqa 96(%rdx), %ymm0
//
//
//vmovaps 128(%rdx), %ymm3
//vmovdqa 160(%rdx), %ymm2
//vmovdqa 192(%rdx), %ymm1
//vmovdqa 224(%rdx), %ymm0
type
TOffset=packed record
x:Byte;
y:Byte;
//m_z:Byte;
//m_w:Byte;
end;
TOffsets=packed record
cl:Byte;
off:array[0..15] of TOffset;
end;
Const
g_offsetOfCacheLine:array[0..2,0..4] of TOffsets=(
( // DISPLAY
(cl: 1; off:((x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 8bpp
(cl: 2; off:((x:0;y:0),(x:0;y:4),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 16bpp
(cl: 4; off:((x:0;y:0),(x:0;y:2),(x:0;y:4),(x:0;y:6),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 32bpp
(cl: 8; off:((x:0;y:0),(x:4;y:0),(x:0;y:2),(x:4;y:2),
(x:0;y:4),(x:4;y:4),(x:0;y:6),(x:4;y:6),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 64bpp
(cl: 0; off:((x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) )
),
( // THIN
(cl: 1; off:((x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 8bpp
(cl: 2; off:((x:0;y:0),(x:0;y:4),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 16bpp
(cl: 4; off:((x:0;y:0),(x:4;y:0),(x:0;y:4),(x:4;y:4),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 32bpp
(cl: 8; off:((x:0;y:0),(x:0;y:2),(x:4;y:0),(x:4;y:2),
(x:0;y:4),(x:0;y:6),(x:4;y:4),(x:4;y:6),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 64bpp
(cl:16; off:((x:0;y:0),(x:2;y:0),(x:0;y:2),(x:2;y:2),
(x:4;y:0),(x:6;y:0),(x:4;y:2),(x:6;y:2),
(x:0;y:4),(x:2;y:4),(x:0;y:6),(x:2;y:6),
(x:4;y:4),(x:6;y:4),(x:4;y:6),(x:6;y:6)) ) // 128bpp
),
( // DEPTH
(cl: 1; off:((x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 8bpp
(cl: 2; off:((x:0;y:0),(x:0;y:4),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 16bpp
(cl: 4; off:((x:0;y:0),(x:4;y:0),(x:0;y:4),(x:4;y:4),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 32bpp
(cl: 8; off:((x:0;y:0),(x:0;y:2),(x:4;y:0),(x:4;y:2),
(x:0;y:4),(x:0;y:6),(x:4;y:4),(x:4;y:6),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0),
(x:0;y:0),(x:0;y:0),(x:0;y:0),(x:0;y:0)) ), // 64bpp
(cl:16; off:((x:0;y:0),(x:2;y:0),(x:0;y:2),(x:2;y:2),
(x:4;y:0),(x:6;y:0),(x:4;y:2),(x:6;y:2),
(x:0;y:4),(x:2;y:4),(x:0;y:6),(x:2;y:6),
(x:4;y:4),(x:6;y:4),(x:4;y:6),(x:6;y:6)) ) // 128bpp
)
);
procedure detile32bppBuf(var T:Tiler2d;src,dst:Pointer);
type
PcontiguousCache=^TcontiguousCache;
TcontiguousCache=array[0..15,0..63] of Byte;
var
x,y,z,bytesPerElement,cacheLine:Ptrint;
tiled_offset,linear_offset:Ptrint;
offsetOfCacheLine:TOffsets;
contiguous_data:packed record
data:TcontiguousCache;
align:array[0..{15}31] of Byte;
end;
contiguous:PcontiguousCache;
cacheLineX:Ptrint;
cacheLineY:Ptrint;
cacheLineZ:Ptrint;
begin
linear_offset:=0;
bytesPerElement:=T.m_bitsPerElement div 8;
offsetOfCacheLine:=g_offsetOfCacheLine[T.m_microTileMode][fastIntLog2(bytesPerElement)];
//Writeln(HexStr(@contiguous_data));
//Writeln(HexStr(Align(@contiguous_data,16)));
contiguous:=Align(@contiguous_data,{16}32);
//FillChar(contiguous^,SizeOf(TcontiguousCache),0);
z:=0;
While (z+T.m_tileThickness<=T.m_linearDepth) do
begin
y:=0;
While (y+kMicroTileHeight<=T.m_linearHeight) do
begin
x:=0;
While (x+kMicroTileWidth<=T.m_linearWidth) do
begin
For cacheLine:=0 to offsetOfCacheLine.cl-1 do //4*64/32=8
begin
cacheLineX:=x + offsetOfCacheLine.off[cacheLine].x;
cacheLineY:=y + offsetOfCacheLine.off[cacheLine].y;
cacheLineZ:=z + 0;
tiled_offset:=0;
T.getTiledElementByteOffset_2d_32(qword(tiled_offset),cacheLineX,cacheLineY,cacheLineZ);
move64_avx_cached(@contiguous^[0][0],(src+tiled_offset),cacheLine);
//move64_avx(@contiguous^[cacheLine][0],(src + tiled_offset));
//Move((src + tiled_offset)^,contiguous^[cacheLine][0], 64);
end;
//Writeln(HexStr(dst + linear_offset));
//Writeln(HexStr(contiguous));
//Writeln(T.m_linearWidth);
linear_offset:=(x*bytesPerElement)+(y*bytesPerElement*T.m_linearWidth)+(z*bytesPerElement*T.m_linearWidth*T.m_linearHeight);
detile32bppDisplayAvx_cached(dst + linear_offset,contiguous, T.m_linearWidth);
//detile32bppDisplayAvx(dst + linear_offset, contiguous, T.m_linearWidth);
//detile32bppDisplaySse2(dst + linear_offset, contiguous, T.m_linearWidth);
//linear_offset:=linear_offset+bytesPerElement*kMicroTileWidth;
x:=x+kMicroTileWidth;
end;
{if (x<T.m_linearWidth) then //slow
begin
For x:=x to T.m_linearWidth-1 do
begin
tiled_offset:=0;
T.getTiledElementBitOffset(qword(tiled_offset),x,y,z,0);
tiled_offset:=tiled_offset div 8;
linear_offset:=(x*bytesPerElement)+(y*bytesPerElement*T.m_linearWidth)+(z*bytesPerElement*T.m_linearWidth*T.m_linearHeight);
PDWORD(dst + linear_offset)^:=PDWORD(src+tiled_offset)^;
end;
end;}
y:=y+kMicroTileHeight;
end;
if (y<T.m_linearHeight) then //slow
begin
For y:=y to T.m_linearHeight-1 do
begin
For x:=0 to T.m_linearWidth-1 do
begin
tiled_offset:=0;
T.getTiledElementByteOffset_2d_32(qword(tiled_offset),x,y,z);
linear_offset:=(x*bytesPerElement)+(y*bytesPerElement*T.m_linearWidth)+(z*bytesPerElement*T.m_linearWidth*T.m_linearHeight);
PDWORD(dst + linear_offset)^:=PDWORD(src+tiled_offset)^;
end;
end;
end;
z:=z+T.m_tileThickness;
end;
end;
//int32_t sce::GpuAddress::Tiler2d::detileSurfaceRegionOneFragment(
//void *outUntiledPixels,
//const void *inTiledPixels,
//const SurfaceRegion *srcRegion,
// uint32_t destPitch,
// uint32_t destSlicePitch,
// uint32_t fragment)
{
const auto region = *srcRegion;
const auto in_bytes = static_cast<const uint8_t*>(inTiledPixels);
const auto out_bytes = static_cast<uint8_t*>(outUntiledPixels);
const auto bytesPerElement = m_bitsPerElement / 8;
if(m_microTileMode == Gnm::kMicroTileModeDepth && m_numFragmentsPerPixel > 1)
{
for(auto z = 0; z < depth(region); ++z)
for(auto y = 0; y < height(region); ++y)
{
uint64_t linear_offset;
computeLinearElementByteOffset(&linear_offset, 0, y, z, 0, destPitch, destSlicePitch, m_bitsPerElement, 1);
for(auto x = 0; x < width(region); ++x)
{
uint64_t tiled_offset;
getTiledElementByteOffset(&tiled_offset, region.m_left + x, region.m_top + y, region.m_front + z, fragment);
small_memcpy(out_bytes + linear_offset, in_bytes + tiled_offset, bytesPerElement);
linear_offset += bytesPerElement;
}
}
return kStatusSuccess;
}
bool canTakeFastPath = true;
if(m_microTileMode >= sizeof(g_offsetOfCacheLine)/sizeof(g_offsetOfCacheLine[0]))
canTakeFastPath = false;
if(canTakeFastPath)
{
Regions regions;
regions.Init(region, m_tileThickness);
if(hasTexels(regions.m_aligned))
{
const auto microTileFunc = getDetileFuncSse2(m_microTileMode, m_bitsPerElement);
SCE_GNM_ASSERT_MSG_RETURN(nullptr != microTileFunc, kStatusInvalidArgument, "Can't find SSE2 detiling function for micro tilemode %d.", m_microTileMode);
const auto offsetOfCacheLine = &g_offsetOfCacheLine[m_microTileMode][fastIntLog2(bytesPerElement)];
const int dx = regions.m_aligned.m_left - region.m_left;
const int dy = regions.m_aligned.m_top - region.m_top;
const int dz = regions.m_aligned.m_front - region.m_front;
for(auto z = 0; z < depth(regions.m_aligned); z += m_tileThickness)
for(auto y = 0; y < height(regions.m_aligned); y += kMicroTileHeight)
for(auto x = 0; x < width(regions.m_aligned); x += kMicroTileWidth)
{
// Due to tile split, the cache lines of a microtile may be stored non-contiguously.
// But to use the optimized microtile detiler, all cache lines of a microtile must be stored contiguously.
// Therefore, here we gather all the cache lines together into a temporary buffer before proceeding...
uint8_t contiguous[16][64];
for(auto cacheLine = 0U; cacheLine < offsetOfCacheLine->m_cacheLinesPerFragment; ++cacheLine)
{
const auto cacheLineX = regions.m_aligned.m_left + x + offsetOfCacheLine->m_offset[cacheLine].m_x;
const auto cacheLineY = regions.m_aligned.m_top + y + offsetOfCacheLine->m_offset[cacheLine].m_y;
const auto cacheLineZ = regions.m_aligned.m_front + z + offsetOfCacheLine->m_offset[cacheLine].m_z;
uint64_t tiled_offset;
getTiledElementByteOffset(&tiled_offset, cacheLineX, cacheLineY, cacheLineZ, fragment);
memcpy(contiguous[cacheLine], in_bytes + tiled_offset, 64);
}
// Now that we have one contiguous microtile, we can pass it to the optimized microtile detiler...
uint64_t linear_offset;
computeLinearElementByteOffset(&linear_offset, dx + x, dy + y, dz + z, 0, destPitch, destSlicePitch, m_bitsPerElement, 1);
microTileFunc(out_bytes + linear_offset, contiguous, destPitch, destSlicePitch);
}
for(auto i = 0; i < regions.m_unaligneds; ++i)
slowDetileOneFragment<Tiler2d>(this, region, regions.m_unaligned[i], fragment, destPitch, destSlicePitch, out_bytes, in_bytes, bytesPerElement);
return kStatusSuccess;
}
}
slowDetileOneFragment<Tiler2d>(this, region, region, fragment, destPitch, destSlicePitch, out_bytes, in_bytes, bytesPerElement);
return kStatusSuccess;
}
end.
Reference in New Issue View Git Blame Copy Permalink