fabi/Xenia-Canary
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[D3D12] Use typed buffers for texture untiling, calculate guest address once per thread

Browse Source
This commit is contained in:
Triang3l
2020-07-03 20:03:53 +03:00
parent 1d35ba9ee7
commit 713f26b5c8
169 changed files with 49015 additions and 22864 deletions
Download Patch File Download Diff File Expand all files Collapse all files

91
src/xenia/gpu/d3d12/shaders/texture_load_dxt3aas1111.cs.hlsl
View File

@@ -1,56 +1,55 @@
#include "pixel_formats.hlsli"
#include "texture_load.hlsli"
uint4 XeDXT3AAs1111TwoBlocksRowToBGRA4(uint2 halfblocks) {
// Only 16 bits of half-blocks are used. X contains pixels 0123, Y - 4567 (in
// the image, halfblocks.y is halfblocks.x + 8).
// In the row, X contains pixels 01, Y - 23, Z - 45, W - 67.
// Assuming alpha in LSB and red in MSB, because it's consistent with how
// DXT1/DXT3/DXT5 color components and CTX1 X/Y are ordered in:
// http://fileadmin.cs.lth.se/cs/Personal/Michael_Doggett/talks/unc-xenos-doggett.pdf
// (LSB on the right, MSB on the left.)
// TODO(Triang3l): Investigate this better, Halo: Reach is the only known game
// that uses it (for lighting in certain places - one of easy to notice usages
// is the T-shaped (or somewhat H-shaped) metal beams in the beginning of
// Winter Contingency), however the contents don't say anything about the
// channel order.
uint4 row = (((halfblocks.xxyy >> uint2(3u, 11u).xyxy) & 1u) << 8u) |
(((halfblocks.xxyy >> uint2(7u, 15u).xyxy) & 1u) << 24u) |
(((halfblocks.xxyy >> uint2(2u, 10u).xyxy) & 1u) << 4u) |
(((halfblocks.xxyy >> uint2(6u, 14u).xyxy) & 1u) << 20u) |
((halfblocks.xxyy >> uint2(1u, 9u).xyxy) & 1u) |
(((halfblocks.xxyy >> uint2(5u, 13u).xyxy) & 1u) << 16u) |
(((halfblocks.xxyy >> uint2(0u, 8u).xyxy) & 1u) << 12u) |
(((halfblocks.xxyy >> uint2(4u, 12u).xyxy) & 1u) << 28u);
row |= row << 1u;
row |= row << 2u;
return row;
}
Buffer<uint4> xe_texture_load_source : register(t0);
RWBuffer<uint4> xe_texture_load_dest : register(u0);
[numthreads(16, 32, 1)]
[numthreads(8, 32, 1)]
void main(uint3 xe_thread_id : SV_DispatchThreadID) {
// 1 thread = 2 DXT3A-as-1111 blocks to 8x4 B4G4R4A4 texels.
uint3 block_index = xe_thread_id;
block_index.x <<= 1u;
// 1 thread = 4 DXT3A-as-1111 blocks to 16x4 B4G4R4A4 texels.
uint3 block_index = xe_thread_id << uint3(2, 0, 0);
[branch] if (any(block_index >= xe_texture_load_size_blocks)) {
return;
}
uint4 block_offsets_guest =
XeTextureLoadGuestBlockOffsets(block_index, 8u, 3u);
uint4 blocks = uint4(xe_texture_load_source.Load2(block_offsets_guest.x),
xe_texture_load_source.Load2(block_offsets_guest.y));
blocks = XeByteSwap(blocks, xe_texture_load_endianness);
uint3 texel_index_host = block_index << uint3(2u, 2u, 0u);
uint texel_offset_host = XeTextureHostLinearOffset(
texel_index_host, xe_texture_load_height_texels,
xe_texture_load_host_pitch, 2u) + xe_texture_load_host_base;
for (uint i = 0u; i < 4u; ++i) {
xe_texture_load_dest.Store4(
texel_offset_host,
XeDXT3AAs1111TwoBlocksRowToBGRA4(
(i >= 2u ? blocks.yw : blocks.xz) >> ((i & 1u) << 4u)));
if (++texel_index_host.y >= xe_texture_load_height_texels) {
return;
int3 texel_index_host = int3(block_index) << int3(2, 2, 0);
int block_offset_host =
(XeTextureHostLinearOffset(texel_index_host,
xe_texture_load_height_texels,
xe_texture_load_host_pitch, 2u) +
xe_texture_load_host_base) >> 4;
int elements_pitch_host = xe_texture_load_host_pitch >> 4;
int block_offset_guest =
XeTextureLoadGuestBlockOffset(int3(block_index), 8u, 3u) >> 4;
uint endian = XeTextureLoadEndian();
uint4 blocks_01 = XeByteSwap(xe_texture_load_source[block_offset_guest],
endian);
// Odd 2 blocks = even 2 blocks + 32 bytes when tiled.
block_offset_guest += XeTextureLoadIsTiled() ? 2 : 1;
uint4 blocks_23 = XeByteSwap(xe_texture_load_source[block_offset_guest],
endian);
xe_texture_load_dest[block_offset_host] =
XeDXT3AAs1111TwoBlocksRowToBGRA4(blocks_01.xz);
xe_texture_load_dest[block_offset_host + 1] =
XeDXT3AAs1111TwoBlocksRowToBGRA4(blocks_23.xz);
[branch] if (++texel_index_host.y < int(xe_texture_load_height_texels)) {
block_offset_host += elements_pitch_host;
xe_texture_load_dest[block_offset_host] =
XeDXT3AAs1111TwoBlocksRowToBGRA4(blocks_01.xz >> 16u);
xe_texture_load_dest[block_offset_host + 1] =
XeDXT3AAs1111TwoBlocksRowToBGRA4(blocks_23.xz >> 16u);
[branch] if (++texel_index_host.y < int(xe_texture_load_height_texels)) {
block_offset_host += elements_pitch_host;
xe_texture_load_dest[block_offset_host] =
XeDXT3AAs1111TwoBlocksRowToBGRA4(blocks_01.yw);
xe_texture_load_dest[block_offset_host + 1] =
XeDXT3AAs1111TwoBlocksRowToBGRA4(blocks_23.yw);
[branch] if (++texel_index_host.y < int(xe_texture_load_height_texels)) {
block_offset_host += elements_pitch_host;
xe_texture_load_dest[block_offset_host] =
XeDXT3AAs1111TwoBlocksRowToBGRA4(blocks_01.yw >> 16u);
xe_texture_load_dest[block_offset_host + 1] =
XeDXT3AAs1111TwoBlocksRowToBGRA4(blocks_23.yw >> 16u);
}
}
texel_offset_host += xe_texture_load_host_pitch;
}
}