/****************************************************************************** * The MIT License (MIT) * * Copyright (c) 2019-2020 Baldur Karlsson * * Permission is hereby granted, free of charge, to any person obtaining a copy * of this software and associated documentation files (the "Software"), to deal * in the Software without restriction, including without limitation the rights * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell * copies of the Software, and to permit persons to whom the Software is * furnished to do so, subject to the following conditions: * * The above copyright notice and this permission notice shall be included in * all copies or substantial portions of the Software. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN * THE SOFTWARE. ******************************************************************************/ #include "d3d12_test.h" RD_TEST(D3D12_CBuffer_Zoo, D3D12GraphicsTest) { static constexpr const char *Description = "Tests every kind of constant that can be in a cbuffer to make sure it's decoded " "correctly"; std::string pixel = R"EOSHADER( struct float3_1 { float3 a; float b; }; struct nested { float3_1 a; float4 b[4]; float3_1 c[4]; }; struct float2_struct { float x; float y; }; struct nested_with_padding { float a; // 0, <1, 2, 3> float4 b; // {4, 5, 6, 7} float c; // 8, <9, 10, 11> float3 d[4]; // [0]: {12, 13, 14}, <15> // [1]: {16, 17, 18}, <19> // [2]: {20, 21, 22}, <23> // [3]: {24, 25, 26}, <27> }; struct empty_struct { }; struct nested_with_empty { float3 a; // 0, 1, 2, 3 empty_struct b; // float2 c; // 4, 5, <6, 7> }; struct misaligned_struct { float4 a; float2 b; }; cbuffer consts : register(b7) { // dummy* entries are just to 'reset' packing to avoid pollution between tests float4 a; // basic float4 = {0, 1, 2, 3} float3 b; // should have a padding word at the end = {4, 5, 6}, <7> float2 c; float2 d; // should be packed together = {8, 9}, {10, 11} float e; float3 f; // should be packed together = 12, {13, 14, 15} float g; float2 h; float i; // should be packed together = 16, {17, 18}, 19 float j; float2 k; // should have a padding word at the end = 20, {21, 22}, <23> float2 l; float m; // should have a padding word at the end = {24, 25}, 26, <27> float n[4]; // should cover 4 float4s = 28, <29..31>, 32, <33..35>, 36, <37..39>, 40 float4 dummy1; float o[4]; // should cover 4 float4s = 48, <..>, 52, <..>, 56, <..>, 60 float p; // should be packed in with above array, with two padding words = 61, <62, 63> float4 dummy2; float4 dummygl1; // padding to match GL so matrices start on same values float4 dummygl2; column_major float4x4 q; // should cover 4 float4s. // row0: {76, 80, 84, 88} // row1: {77, 81, 85, 89} // row2: {78, 82, 86, 90} // row3: {79, 83, 87, 91} row_major float4x4 r; // should cover 4 float4s // row0: {92, 93, 94, 95} // row1: {96, 97, 98, 99} // row2: {100, 101, 102, 103} // row3: {104, 105, 106, 107} column_major float3x4 s; // covers 4 float4s with padding at end of each column // row0: {108, 112, 116, 120} // row1: {109, 113, 117, 121} // row2: {110, 114, 118, 122} // <111, 115, 119, 123> float4 dummy3; row_major float3x4 t; // covers 3 float4s with no padding // row0: {128, 129, 130, 131} // row1: {132, 133, 134, 135} // row2: {136, 137, 138, 139} float4 dummy4; column_major float2x3 u; // covers 3 float4s with padding at end of each column (but not row) // row0: {144, 148, 152} // row1: {145, 149, 153} // <146, 150, 154> // <147, 151, 155> float4 dummy5; row_major float2x3 v; // covers 2 float4s with padding at end of each row (but not column) // row0: {160, 161, 162}, <163> // row1: {164, 165, 166}, <167> float4 dummy6; column_major float2x2 w; // covers 2 float4s with padding at end of each column (but not row) // row0: {172, 176} // row1: {173, 177} // <174, 178> // <175, 179> float4 dummy7; row_major float2x2 x; // covers 2 float4s with padding at end of each row (but not column) // row1: {184, 185}, <186, 187> // row1: {188, 189}, <190, 191> float4 dummy8; row_major float2x2 y; // covers the same as above, but z overlaps // row0: {196, 197}, <198, 199> // row1: {200, 201}, <202, 203> float z; // overlaps after padding in final row = 202 float4 gldummy3; // account for z not overlapping in GL/VK row_major float4x1 aa; // covers 4 vec4s with maximum padding // row0: {208}, <209, 210, 211> // row1: {212}, <213, 214, 215> // row2: {216}, <217, 218, 219> // row3: {220}, <221, 222, 223> column_major float4x1 ab; // covers 1 vec4 (equivalent to a plain vec4) // row0: {224} // row1: {225} // row2: {226} // row3: {227} float4 multiarray[3][2]; // [0][0] = {228, 229, 230, 231} // [0][1] = {232, 233, 234, 235} // [1][0] = {236, 237, 238, 239} // [1][1] = {240, 241, 242, 243} // [2][0] = {244, 245, 246, 247} // [2][1] = {248, 249, 250, 251} nested structa[2]; // [0] = { // .a = { { 252, 253, 254 }, 255 } // .b[0] = { 256, 257, 258, 259 } // .b[1] = { 260, 261, 262, 263 } // .b[2] = { 264, 265, 266, 267 } // .b[3] = { 268, 269, 270, 271 } // .c[0] = { { 272, 273, 274 }, 275 } // .c[1] = { { 276, 277, 278 }, 279 } // .c[2] = { { 280, 281, 282 }, 283 } // .c[3] = { { 284, 285, 286 }, 287 } // } // [1] = { // .a = { { 288, 289, 290 }, 291 } // .b[0] = { 292, 293, 294, 295 } // .b[1] = { 296, 297, 298, 299 } // .b[2] = { 300, 301, 302, 303 } // .b[3] = { 304, 305, 306, 307 } // .c[0] = { { 308, 309, 310 }, 311 } // .c[1] = { { 312, 313, 314 }, 315 } // .c[2] = { { 316, 317, 318 }, 319 } // .c[3] = { { 320, 321, 322 }, 323 } // } column_major float3x2 ac; // covers 2 float4s with padding at end of each column (but not row) // row0: {324, 328} // row1: {325, 329} // row2: {326, 330} // <327, 331> row_major float3x2 ad; // covers 3 float4s with padding at end of each row (but not column) // row0: {332, 333}, <334, 335> // row1: {336, 337}, <338, 339> // row2: {340, 341}, <342, 343> column_major float3x2 ae[2]; // covers 2 float4s with padding at end of each column (but not row) // [0] = { // row0: {344, 348} // row1: {345, 349} // row2: {346, 350} // <347, 351> // } // [1] = { // row0: {352, 356} // row1: {353, 357} // row2: {354, 358} // <355, 359> // } row_major float3x2 af[2]; // covers 3 float4s with padding at end of each row (but not column) // [0] = { // row0: {360, 361}, <362, 363> // row1: {364, 365}, <366, 367> // row2: {368, 369}, <370, 371> // } // [1] = { // row0: {372, 373}, <374, 375> // row1: {376, 377}, <378, 379> // row2: {380, 381}, // } float2 dummy9; // consumes leftovers from above array = {382, 383} float2 dummy10; // should have padding at the end = {384, 385}, <386, 387> row_major float2x2 ag; // each row is aligned to float4: // row0: {388, 389}, <390, 391> // row1: {392, 393}, float2 dummy11; // consumes leftovers from above matrix = {394, 395} float2 dummy12; // should have padding at the end = {396, 397}, <398, 399> column_major float2x2 ah; // each column is aligned to float4: // row0: {400, 404} // row1: {401, 405} // <402, 406> // <403, 407> row_major float2x2 ai[2]; // [0] = { // row0: {408, 409}, <410, 411> // row1: {412, 413}, <414, 415> // } // [1] = { // row0: {416, 417}, <418, 419> // row1: {420, 421}, <422, 423> // } column_major float2x2 aj[2]; // [0] = { // row0: {424, 428} // row1: {425, 429} // <426, 430> // <427, 431> // } // [1] = { // row0: {432, 436} // row1: {433, 437} // <434, 438> // <435, 439> // } nested_with_padding ak[2]; // 440 - 467, 468 - 495 float4 dummy13; // forces no trailing overlap with ak float al; // {500}, <501, 502, 503> // struct is always float4 aligned, can't be packed with al float2_struct am; // {504, 505}, <506, 507> // struct allows trailing things into padding float an; // {506} float4 gldummy4; // account for an not overlapping in GL/VK empty_struct empty; // completely omitted nested_with_empty nested_empty; // empty struct will take up a float4 misaligned_struct ao[2]; // [0] = { // .a = { 520, 521, 522, 523 } // .b = { 524, 525 } <526, 527> // } // [1] = { // .a = { 528, 529, 530, 531 } // .b = { 532, 533 } <534, 535> // } float4 test; // {536, 537, 538, 539} }; // this comes from root signature constants cbuffer rootconsts : register(b1) { float4 root_zero; float4 root_a; float2 root_b, root_c; float3_1 root_d; }; cbuffer hugespace : register(b0, space999999999) { float4 huge_val; }; float4 main() : SV_Target0 { return test + root_zero + huge_val * 1e-30f + float4(0.1f, 0.0f, 0.0f, 0.0f); } )EOSHADER"; struct float3_1 { float a[3]; float b; }; struct RootData { float root_zero[4]; float root_a[4]; float root_b[2], root_c[2]; float3_1 root_d; }; int main() { // initialise, create window, create device, etc if(!Init()) return 3; ID3DBlobPtr vsblob = Compile(D3DDefaultVertex, "main", "vs_5_0"); ID3DBlobPtr psblob = Compile(pixel, "main", "ps_5_1"); const size_t bindOffset = 16; Vec4f cbufferdata[bindOffset + 512]; for(int i = 0; i < bindOffset; i++) cbufferdata[i] = Vec4f(-99.9f, -88.8f, -77.7f, -66.6f); for(int i = 0; i < 512; i++) cbufferdata[bindOffset + i] = Vec4f(float(i * 4 + 0), float(i * 4 + 1), float(i * 4 + 2), float(i * 4 + 3)); RootData rootData = {}; rootData.root_a[0] = 10.0f; rootData.root_a[1] = 20.0f; rootData.root_a[2] = 30.0f; rootData.root_a[3] = 40.0f; rootData.root_b[0] = 50.0f; rootData.root_b[1] = 60.0f; rootData.root_c[0] = 70.0f; rootData.root_c[1] = 80.0f; rootData.root_d.a[0] = 90.0f; rootData.root_d.a[1] = 100.0f; rootData.root_d.a[2] = 110.0f; rootData.root_d.b = 120.0f; static_assert(sizeof(rootData) == 64, "Root data is mis-sized"); ID3D12ResourcePtr vb = MakeBuffer().Data(DefaultTri); ID3D12ResourcePtr cb = MakeBuffer().Data(cbufferdata); ID3D12RootSignaturePtr sig = MakeSig({ cbvParam(D3D12_SHADER_VISIBILITY_PIXEL, 0, 7), constParam(D3D12_SHADER_VISIBILITY_PIXEL, 0, 1, sizeof(rootData) / sizeof(uint32_t)), cbvParam(D3D12_SHADER_VISIBILITY_PIXEL, 999999999, 0), }); ID3D12PipelineStatePtr pso = MakePSO().RootSig(sig).InputLayout().VS(vsblob).PS(psblob).RTVs( {DXGI_FORMAT_R32G32B32A32_FLOAT}); ResourceBarrier(vb, D3D12_RESOURCE_STATE_COMMON, D3D12_RESOURCE_STATE_VERTEX_AND_CONSTANT_BUFFER); ResourceBarrier(cb, D3D12_RESOURCE_STATE_COMMON, D3D12_RESOURCE_STATE_VERTEX_AND_CONSTANT_BUFFER); ID3D12ResourcePtr rtvtex = MakeTexture(DXGI_FORMAT_R32G32B32A32_FLOAT, screenWidth, screenHeight) .RTV() .InitialState(D3D12_RESOURCE_STATE_RENDER_TARGET); while(Running()) { ID3D12GraphicsCommandListPtr cmd = GetCommandBuffer(); Reset(cmd); ID3D12ResourcePtr bb = StartUsingBackbuffer(cmd, D3D12_RESOURCE_STATE_RENDER_TARGET); D3D12_CPU_DESCRIPTOR_HANDLE bbrtv = MakeRTV(bb).Format(DXGI_FORMAT_R8G8B8A8_UNORM_SRGB).CreateCPU(0); ClearRenderTargetView(cmd, bbrtv, {0.2f, 0.2f, 0.2f, 1.0f}); D3D12_CPU_DESCRIPTOR_HANDLE offrtv = MakeRTV(rtvtex).CreateCPU(0); ClearRenderTargetView(cmd, offrtv, {0.2f, 0.2f, 0.2f, 1.0f}); cmd->IASetPrimitiveTopology(D3D_PRIMITIVE_TOPOLOGY_TRIANGLELIST); IASetVertexBuffer(cmd, vb, sizeof(DefaultA2V), 0); cmd->SetPipelineState(pso); cmd->SetGraphicsRootSignature(sig); cmd->SetGraphicsRootConstantBufferView( 0, cb->GetGPUVirtualAddress() + bindOffset * sizeof(Vec4f)); cmd->SetGraphicsRoot32BitConstants(1, sizeof(rootData) / sizeof(uint32_t), &rootData, 0); cmd->SetGraphicsRootConstantBufferView( 2, cb->GetGPUVirtualAddress() + bindOffset * sizeof(Vec4f) + 256); RSSetViewport(cmd, {0.0f, 0.0f, (float)screenWidth, (float)screenHeight, 0.0f, 1.0f}); RSSetScissorRect(cmd, {0, 0, screenWidth, screenHeight}); OMSetRenderTargets(cmd, {offrtv}, {}); cmd->DrawInstanced(3, 1, 0, 0); ResourceBarrier(cmd, rtvtex, D3D12_RESOURCE_STATE_RENDER_TARGET, D3D12_RESOURCE_STATE_PIXEL_SHADER_RESOURCE); blitToSwap(cmd, rtvtex, bb); ResourceBarrier(cmd, rtvtex, D3D12_RESOURCE_STATE_PIXEL_SHADER_RESOURCE, D3D12_RESOURCE_STATE_RENDER_TARGET); FinishUsingBackbuffer(cmd, D3D12_RESOURCE_STATE_RENDER_TARGET); cmd->Close(); Submit({cmd}); Present(); } return 0; } }; REGISTER_TEST();