Fixes and optimizations

.github/workflows/ci.windows.yml

@@ -112,3 +112,15 @@ jobs:
       artifact-out: test-windows-msvc17-avx512
       artifact-path: build_regression_tests/tests*.xml build_regression_tests/failed*
       artifact-on-failure: true
+
+  test-find-ospray:
+    needs: build-windows-msvc17
+    uses: intel-innersource/libraries.devops.renderkit.workflows/.github/workflows/windows.yml@main
+    with:
+      cmd: |
+        md buildtut
+        cd buildtut
+        $env:ospray_DIR=(Get-ChildItem $env:GITHUB_WORKSPACE\build\install\ospray\lib\cmake\ospray-* | Select-Object -Expand FullName)
+        cmake ..\apps\ospTutorial\ospTutorialFindospray
+        cmake --build .
+      artifact-in: build-windows-msvc17

apps/ospExamples/GLFWOSPRayWindow.cpp

@@ -164,7 +164,7 @@ GLFWOSPRayWindow::GLFWOSPRayWindow(
 
   glfwWindowHint(GLFW_SRGB_CAPABLE, GLFW_TRUE);
   glfwWindowHint(GLFW_CONTEXT_VERSION_MAJOR, 3);
-#ifdef __APPLE_
+#ifdef __APPLE__
   glfwWindowHint(GLFW_CONTEXT_VERSION_MINOR, 2);
   glfwWindowHint(GLFW_OPENGL_FORWARD_COMPAT, GL_TRUE);
   const char *glslVersion = "#version 150";

apps/ospTestSuite/test_enums.cpp

@@ -98,6 +98,7 @@ TEST(Enums, VKLUnstructuredCellType)
   ASSERT_EQ(OSP_HEXAHEDRON, VKL_HEXAHEDRON);
   ASSERT_EQ(OSP_WEDGE, VKL_WEDGE);
   ASSERT_EQ(OSP_PYRAMID, VKL_PYRAMID);
+  ASSERT_EQ(sizeof(OSPUnstructuredCellType), 1); // must fit in uint8
 }
 
 TEST(Enums, VKLAMRMethod)

cmake/compiler/msvc.cmake

@@ -2,6 +2,9 @@
 ## SPDX-License-Identifier: Apache-2.0
 
 set(COMMON_CXX_FLAGS "/EHsc /MP /GR /bigobj")
+if (MSVC_VERSION GREATER_EQUAL 1914)
+  string(APPEND COMMON_CXX_FLAGS " /Zc:__cplusplus")
+endif()
 
 set(CMAKE_CXX_FLAGS_DEBUG          "${CMAKE_CXX_FLAGS_DEBUG} ${COMMON_CXX_FLAGS}")
 set(CMAKE_CXX_FLAGS_RELEASE        "${CMAKE_CXX_FLAGS_RELEASE}        ${COMMON_CXX_FLAGS} /Ox /fp:fast /Oi /Gy ")

cmake/ospray_cmake_config/osprayConfig.cmake.in

@@ -196,6 +196,12 @@ set_target_properties(ospray::ospray PROPERTIES
 set_target_properties(ospray::ospray_module_cpu PROPERTIES
   INTERFACE_LINK_LIBRARIES "ospray::ospray;${OPENVKL_LIBRARY};${EMBREE_LIBRARY}")
 
+# https://gitlab.kitware.com/cmake/cmake/-/issues/18837
+if (MSVC AND (MSVC_VERSION GREATER_EQUAL 1914))
+  set_property(TARGET ospray::ospray APPEND PROPERTY
+    INTERFACE_COMPILE_OPTIONS "$<$<COMPILE_LANGUAGE:CXX>:/Zc:__cplusplus>")
+endif()
+
 # Reset CMake module path to its state when this script was called.
 set(CMAKE_MODULE_PATH ${OSPRAY_CALLERS_CMAKE_MODULE_PATH})
 

doc/api.md

@@ -832,11 +832,12 @@ the vertices and data values. Vertex ordering is the same as
 `VTK_PYRAMID`: four bottom vertices counterclockwise, then the top
 vertex.
 
-To maintain VTK data compatibility, the `index` array may be specified
-with cell sizes interleaved with vertex indices in the following format:
-$n, id_1, ..., id_n, m, id_1, ..., id_m$. This alternative `index` array
-layout can be enabled through the `indexPrefixed` flag (in which case,
-the `cell.type` parameter must be omitted).
+For backward compatibility with legacy vtkCellArray the `index` array
+can also be specified with cell sizes interleaved with vertex indices in
+the following format: $n, id_1, ..., id_n, m, id_1, ..., id_m$. This
+alternative `index` array layout can be enabled through the
+`indexPrefixed` flag (in which case, the `cell.type` parameter must be
+omitted).
 
   ------------------- ------------------ --------  ---------------------------------------
   Type                Name                Default  Description
@@ -850,9 +851,9 @@ the `cell.type` parameter must be omitted).
                                                    vertex array(s)) that form cells
 
   bool                indexPrefixed          false indicates that the `index` array is
-                                                   compatible to VTK, where the indices of
-                                                   each cell are prefixed with the number
-                                                   of vertices
+                                                   compatible to legacy vtkCellArray,
+                                                   where the indices of each cell are
+                                                   prefixed with the number of vertices
 
   uint32[] / uint64[] cell.index                   [data] array of locations (into the
                                                    index array), specifying the first index
@@ -3387,8 +3388,8 @@ The CPU module is implicitly loaded and the `cpu` device is
 automatically used if no other options are specified.
 
 
-GPU (Beta)
----------
+GPU
+----
 
 To use the GPU for rendering load the `gpu` module and select the `gpu`
 device:
@@ -3441,11 +3442,9 @@ correctly on the GPU device:
 
 - Multiple volumes in the scene
 - Clipping
-- Motion blur
 - Subdivision surfaces
 - Progress reporting via `ospGetProgress` or canceling the frame via `ospCancel`
 - Picking via `ospPick`
-- Adaptive accumulation via `OSP_FB_VARIANCE` and `varianceThreshold`
 - Framebuffer channels `OSP_FB_ID_*` (id buffers)
 - Experimental support for shared device-only data, works only for
   `structuredRegular` volume

modules/cpu/CMakeLists.txt

@@ -6,7 +6,7 @@ mark_as_advanced(OSPRAY_MODULE_CPU)
 
 cmake_dependent_option(
   OSPRAY_MODULE_GPU
-  "Build the GPU module (beta)"
+  "Build the GPU module"
   OFF
   EMBREE_SYCL_SUPPORT
   OFF

modules/cpu/geometry/Curves.cpp

@@ -189,6 +189,12 @@ void Curves::createEmbreeGeometry()
     setEmbreeGeometryBuffer(
         embreeGeometry, RTC_BUFFER_TYPE_VERTEX_ATTRIBUTE, texcoordData, 1);
   }
+  capData.clear();
+  if (curveType == OSP_DISJOINT) {
+    // disable caps to get always open cones/cylinders
+    capData.resize(numPrimitives(), RTC_CURVE_FLAG_NEIGHBOR_LEFT | RTC_CURVE_FLAG_NEIGHBOR_RIGHT);
+    setEmbreeGeometryBuffer(embreeGeometry, RTC_BUFFER_TYPE_FLAGS, capData);
+  }
 
   rtcCommitGeometry(embreeGeometry);
 }

modules/cpu/geometry/Curves.h

@@ -28,6 +28,7 @@ struct OSPRAY_SDK_INTERFACE Curves
   Ref<const DataT<vec4f>> tangentData;
   Ref<const DataT<vec4f>> colorData;
   Ref<const DataT<vec2f>> texcoordData;
+  std::vector<uint8_t> capData;
 
   RTCGeometryType embreeCurveType{(RTCGeometryType)-1};
 

modules/cpu/geometry/SpheresShared.h

@@ -15,11 +15,10 @@ struct Spheres
   Data1D sphere;
   Data1D texcoord;
   Data1D normalData;
-  float global_radius;
   OSPSphereType sphereType;
 
 #ifdef __cplusplus
-  Spheres() : global_radius(.01f)
+  Spheres()
   {
     super.type = GEOMETRY_TYPE_SPHERES;
   }

modules/cpu/math/sobol.ih

@@ -30,19 +30,30 @@ OSPRAY_BEGIN_ISPC_NAMESPACE
 #ifdef OSPRAY_TARGET_SYCL
 #include "sobol.inl"
 #else
-extern const uniform unsigned int Sobol_revMatrices[4 * 32];
+extern const uniform unsigned int Sobol_revMatrices[3 * 32];
 extern const uniform unsigned int Sobol_matrices[5 * 52];
 #endif
 
+// compute component 1 of the Sobol'-sequence (count starting from 0)
+// Ahmed, "An Implementation Algorithm of 2D Sobol Sequence Fast, Elegant, and
+// Compact", EGSR 2024, https://doi.org/10.2312/sr.20241147
+inline uint32 Sobol_revSample1(uint32 v)
+{
+  v ^= v << 16;
+  v ^= (v & 0x00FF00FFu) << 8;
+  v ^= (v & 0x0F0F0F0Fu) << 4;
+  v ^= (v & 0x33333333u) << 2;
+  v ^= (v & 0x55555555u) << 1;
+  return reverseBits(v);
+}
+
 // compute components 1 and 2 of the Sobol'-sequence (count starting from 0)
 inline vec2ui Sobol_revSample2(uint32 revIndex)
 {
-  vec2ui result = make_vec2ui(0);
-  for (uniform uint32 i = 0; revIndex; revIndex <<= 1, i += 4) {
-    if (revIndex & 0x80000000u) {
-      result.x ^= Sobol_revMatrices[i];
-      result.y ^= Sobol_revMatrices[i + 1];
-    }
+  vec2ui result = make_vec2ui(Sobol_revSample1(revIndex), 0u);
+  for (uniform uint32 i = 0; revIndex; revIndex <<= 1, i += 3) {
+    if (revIndex & 0x80000000u)
+      result.y ^= Sobol_revMatrices[i];
   }
 
   return result;
@@ -51,12 +62,11 @@ inline vec2ui Sobol_revSample2(uint32 revIndex)
 // compute components 1 to 3 of the Sobol'-sequence (count starting from 0)
 inline vec3ui Sobol_revSample3(uint32 revIndex)
 {
-  vec3ui result = make_vec3ui(0);
-  for (uniform uint32 i = 0; revIndex; revIndex <<= 1, i += 4) {
+  vec3ui result = make_vec3ui(Sobol_revSample1(revIndex), 0u, 0u);
+  for (uniform uint32 i = 0; revIndex; revIndex <<= 1, i += 3) {
     if (revIndex & 0x80000000u) {
-      result.x ^= Sobol_revMatrices[i];
-      result.y ^= Sobol_revMatrices[i + 1];
-      result.z ^= Sobol_revMatrices[i + 2];
+      result.y ^= Sobol_revMatrices[i];
+      result.z ^= Sobol_revMatrices[i + 1];
     }
   }
 
@@ -66,13 +76,12 @@ inline vec3ui Sobol_revSample3(uint32 revIndex)
 // compute components 1 to 4 of the Sobol'-sequence (count starting from 0)
 inline vec4ui Sobol_revSample4(uint32 revIndex)
 {
-  vec4ui result = make_vec4ui(0);
-  for (uniform uint32 i = 0; revIndex; revIndex <<= 1, i += 4) {
+  vec4ui result = make_vec4ui(Sobol_revSample1(revIndex), 0u, 0u, 0u);
+  for (uniform uint32 i = 0; revIndex; revIndex <<= 1, i += 3) {
     if (revIndex & 0x80000000u) {
-      result.x ^= Sobol_revMatrices[i];
-      result.y ^= Sobol_revMatrices[i + 1];
-      result.z ^= Sobol_revMatrices[i + 2];
-      result.w ^= Sobol_revMatrices[i + 3];
+      result.y ^= Sobol_revMatrices[i];
+      result.z ^= Sobol_revMatrices[i + 1];
+      result.w ^= Sobol_revMatrices[i + 2];
     }
   }
 

modules/cpu/math/sobol.inl

@@ -30,163 +30,131 @@
 // http://web.maths.unsw.edu.au/~fkuo/sobol/new-joe-kuo-6.21201
 
 // bit-reversed and interleaved
-OSPRAY_GLOBAL const unsigned int Sobol_revMatrices[4 * 32] = {
-    0x1u,
+OSPRAY_GLOBAL const unsigned int Sobol_revMatrices[3 * 32] = {
     0x1u,
     0x1u,
     0x1u,
 
-    0x3u,
     0x3u,
     0x3u,
     0x2u,
 
-    0x5u,
     0x6u,
     0x4u,
     0x4u,
 
-    0xfu,
     0x9u,
     0xau,
     0xdu,
 
-    0x11u,
     0x17u,
     0x1fu,
     0x1fu,
 
-    0x33u,
     0x3au,
     0x2eu,
     0x3bu,
 
-    0x55u,
     0x71u,
     0x45u,
     0x5eu,
 
-    0xffu,
     0xa3u,
     0xc9u,
     0xb9u,
 
-    0x101u,
     0x116u,
     0x11bu,
     0x15au,
 
-    0x303u,
     0x339u,
     0x2a4u,
     0x3f4u,
 
-    0x505u,
     0x677u,
     0x79au,
     0x685u,
 
-    0xf0fu,
     0x9aau,
     0xb67u,
     0xd0fu,
 
-    0x1111u,
     0x1601u,
     0x101eu,
     0x115bu,
 
-    0x3333u,
     0x3903u,
     0x302du,
     0x23f6u,
 
-    0x5555u,
     0x7706u,
     0x4041u,
     0x4681u,
 
-    0xffffu,
     0xaa09u,
     0xa0c3u,
     0xdd02u,
 
-    0x10001u,
     0x10117u,
     0x1f104u,
     0x1e144u,
 
-    0x30003u,
     0x3033au,
     0x2e28au,
     0x393cdu,
 
-    0x50005u,
     0x60671u,
     0x457dfu,
     0x5a6dfu,
 
-    0xf000fu,
     0x909a3u,
     0xc9baeu,
     0xb4dbbu,
 
-    0x110011u,
     0x171616u,
     0x11a105u,
     0x14401eu,
 
-    0x330033u,
     0x3a3939u,
     0x2a7289u,
     0x3cd039u,
 
-    0x550055u,
     0x717777u,
     0x79e7dbu,
     0x6df05au,
 
-    0xff00ffu,
     0xa3aaaau,
     0xb6dba4u,
     0xdbb0b4u,
 
-    0x1010101u,
     0x1170001u,
     0x100011au,
     0x101e145u,
 
-    0x3030303u,
     0x33a0003u,
     0x30002a7u,
     0x20393cfu,
 
-    0x5050505u,
     0x6710006u,
     0x400079eu,
     0x405a6dbu,
 
-    0xf0f0f0fu,
     0x9a30009u,
     0xa000b6du,
     0xd0b4db6u,
 
-    0x11111111u,
     0x16160017u,
     0x1f001001u,
     0x1f144001u,
 
-    0x33333333u,
     0x3939003au,
     0x2e003003u,
     0x3b3cd002u,
 
-    0x55555555u,
     0x77770071u,
     0x45004004u,
     0x5e6df004u,
 
-    0xffffffffu,
     0xaaaa00a3u,
     0xc900a00au,
     0xb9dbb00du,