Sync README, fix links and spelling

CHANGELOG.md

@@ -3,17 +3,6 @@ Version History
 
 ### Changes in v3.2.0:
 
--   Clarify the size of `OSP_BOOL` to be 1 byte
--   Support half-precision (16\ bit float) texture formats
-    `OSP_TEXTURE_[RGBA16F|RGB16F|RA16F|R16F]` and two-channel 32\ bit
-    float textures `OSP_TEXTURE_RA32F`
--   Implement MIP Mapping for better texture filtering. If the
-    additional memory per texture needed cannot be spared, applications
-    can disable the generation of MIP maps with device parameter
-    `disableMipMapGeneration`
--   New parameter `limitIndirectLightSamples` for the `pathtracer` which
-    limits the number of light samples after the first non-specular
-    (i.e., diffuse and glossy) bounce to at most one
 -   Sampling improvements:
       - Better performance (lower rendering time and faster convergence)
       - More pleasing blue noise enabled when the total number of frames
@@ -24,18 +13,32 @@ Version History
       - User-controlled quality levels via parameter `quality`
       - Optionally denoise alpha channel as well, enabled via
         parameter `denoiseAlpha`
+-   Support half-precision (16\ bit float) texture formats
+    `OSP_TEXTURE_[RGBA16F|RGB16F|RA16F|R16F]` and two-channel 32\ bit
+    float textures `OSP_TEXTURE_RA32F`
+-   New parameter `limitIndirectLightSamples` for the `pathtracer` which
+    limits the number of light samples after the first non-specular
+    (i.e., diffuse and glossy) bounce to at most one
+-   Implement MIP Mapping for better texture filtering. If the
+    additional memory per texture needed cannot be spared, applications
+    can disable the generation of MIP maps with device parameter
+    `disableMipMapGeneration`
+-   The backplate (background texture) is now always sampled at the pixel
+    center and thus not blurred by the pixel filter anymore
+-   Avoid color bleeding across eye-subimages when stereo rendering
+-   Superbuild uses binary packages of Open VKL
 -   Removed Intel ISPCRT dependency (ISPC compiler is still needed):
     -   oneAPI Level Zero Loader is no longer necessary
     -   `zeContext` and `zeDevice`device parameters are no longer supported
     -   `ispcrtContext` and `ispcrtDevice`device parameters are no longer
         supported
--   Superbuild uses binary packages of Open VKL
--   Avoid color bleeding across eye-subimages when stereo rendering
--   The backplate (background texture) is now always sampled at the pixel
-    center and thus not blurred by the pixel filter anymore
--   Fix artifacts occassionally appearing with `gpu` device
+-   Clarify the size of `OSP_BOOL` to be 1 byte
+-   Fix artifacts occasionally appearing with `gpu` device
 -   The new minimum versions of dependencies:
     -    Embree v4.3.3 (better error reporting)
+    -    Open Image Denoise v2.3 (better image quality with `HIGH`
+         quality mode, added `FAST` quality mode)
+    -    rkcommon v1.14.0
 
 ### Changes in v3.1.0:
 

STYLEGUIDE.md

@@ -7,7 +7,7 @@ be overly strict. However, deviations from this document should be
 supported with technical reasoning.
 
 This document is relatively simple and meant as a "quick overview" of
-expecations for OSPRay. A more complete set of guidelines for all of C++
+expectations for OSPRay. A more complete set of guidelines for all of C++
 can be found in the [ISO C++ Core
 Guidelines](https://github.com/isocpp/CppCoreGuidelines/blob/master/CppCoreGuidelines.md).
 

cmake/compiler/clang.cmake

@@ -15,7 +15,7 @@ if(OSPRAY_STRICT_BUILD)
   set(OSPRAY_CXX_FLAGS "-Wno-header-hygiene ${OSPRAY_CXX_FLAGS}")
   set(OSPRAY_CXX_FLAGS "-Wno-covered-switch-default ${OSPRAY_CXX_FLAGS}")
   set(OSPRAY_CXX_FLAGS "-Wno-date-time ${OSPRAY_CXX_FLAGS}")
-  set(OSPRAY_CXX_FLAGS "-Wno-unsafe-buffer-usage ${OSPRAY_CXX_FLAGS}") #we use pointer arithmetics on buffers
+  set(OSPRAY_CXX_FLAGS "-Wno-unsafe-buffer-usage ${OSPRAY_CXX_FLAGS}") #we use pointer arithmetic on buffers
   set(OSPRAY_CXX_FLAGS "-Wno-reserved-identifier ${OSPRAY_CXX_FLAGS}") #used in autogenerated ISPC headers
 
   # Should try to fix and remove...

cmake/compiler/ispc.cmake

@@ -172,7 +172,7 @@ macro (ispc_compile)
     message(FATAL_ERROR "CMAKE_BUILD_TYPE (${CMAKE_BUILD_TYPE}) allows only the following values: Debug;Release;RelWithDebInfo")
   endif()
 
-  # turn space sparated list into ';' separated list
+  # turn space separated list into ';' separated list
   string(REPLACE " " ";" ISPC_OPT_FLAGS "${ISPC_OPT_FLAGS}")
 
   if (NOT WIN32)
@@ -225,8 +225,8 @@ macro (ispc_compile)
       endforeach()
     endif()
 
-    # use relative path for genereated header file, because the path is
-    # included as commment and the header will be distributed
+    # use relative path for generated header file, because the path is
+    # included as comment and the header will be distributed
     set(headerfile "${outdir}/${fname}_ispc.h")
     file(RELATIVE_PATH headerfile ${CMAKE_CURRENT_BINARY_DIR} ${headerfile})
 

doc/api.md

@@ -3368,8 +3368,8 @@ memory with OSPRay or to control which device should be used
 (e.g., in case multiple GPUs are present). If neither parameter is set,
 the `gpu` device will automatically create a context internally
 and select a GPU (that selection can be influenced via environment
-variable `ONEAPI_DEVICE_SELECTOR`, see [Intel oneAPI DPC++ Compiler documentation]
-(https://intel.github.io/llvm-docs/EnvironmentVariables.html#oneapi-device-selector)).
+variable `ONEAPI_DEVICE_SELECTOR`, see [Intel oneAPI DPC++ Compiler
+documentation](https://intel.github.io/llvm-docs/EnvironmentVariables.html#oneapi-device-selector)).
 
 Compile times for just in time compilation (JIT compilation) can be
 large. To resolve this issue we recommend enabling persistent JIT
@@ -3382,8 +3382,8 @@ To reduce GPU memory allocation overhead when rendering scenes with many
 objects (geometries, instances, etc.), memory pooling should be enabled
 by setting the environment variable
 `SYCL_PI_LEVEL_ZERO_USM_ALLOCATOR="1;0;shared:1M,0,2M"`.
-See [Intel oneAPI DPC++ Compiler documentation]
-(https://intel.github.io/llvm-docs/EnvironmentVariables.html#debugging-variables-for-level-zero-plugin)
+See [Intel oneAPI DPC++ Compiler
+documentation](https://intel.github.io/llvm-docs/EnvironmentVariables.html#debugging-variables-for-level-zero-plugin)
 for more details.
 
 ### Known Issues {-}

modules/cpu/ISPCDevice.h

@@ -162,7 +162,7 @@ struct OSPRAY_SDK_INTERFACE ISPCDevice : public Device
 
 #ifdef OSPRAY_TARGET_SYCL
   /* Compute the rounded dispatch global size for the given work group size.
-   * SYCL requires that globalSize % workgroupSize == 0, ths function will
+   * SYCL requires that globalSize % workgroupSize == 0, this function will
    * round up globalSize and return nd_range(roundedSize, workgroupSize).
    * The kernel being launched must discard tasks that are out of bounds
    * bounds due to this rounding

modules/cpu/common/World.ispc

@@ -50,7 +50,7 @@ unmasked void clippingIntersectionFilterV(
           instance, NULL, dg, *ray, DG_NG | DG_NS, true, *context->super.ffh);
 
       // Use geometry normal for clipping
-      // but use shading normal to check if invertion is needed
+      // but use shading normal to check if inversion is needed
       Nc = (dot(dg.Ns, dg.Ng) > 0.f) ? ray->Ng : neg(ray->Ng);
     }
 
@@ -60,7 +60,7 @@ unmasked void clippingIntersectionFilterV(
   }
 
   // Based on clipping normal vector decide if the ray is
-  // entering clipping geometry (set positive hit value) or exitting clipping
+  // entering clipping geometry (set positive hit value) or exiting clipping
   // geometry (set negative hit value)
   const bool exitsClipping = (dot(ray->dir, Nc) > 0.f);
   ClippingHit cHit;

modules/cpu/fb/FrameBuffer.h

@@ -59,10 +59,8 @@ struct OSPRAY_SDK_INTERFACE FrameBuffer
   virtual utility::ArrayView<uint32_t> getRenderTaskIDs(
       const float errorThreshold, const uint32_t spp) = 0;
 
-  // get number of pixels in x and y diretion
   vec2i getNumPixels() const;
 
-  // get the color format type for this Buffer
   ColorBufferFormat getColorBufferFormat() const;
 
   virtual float getVariance() const;
@@ -139,7 +137,7 @@ struct OSPRAY_SDK_INTERFACE FrameBuffer
   bool accumulationFinished() const;
 
  private:
-  // for consistent reproducability of variance accumulation
+  // for consistent reproducibility of variance accumulation
   constexpr static uint32_t mtSeed = 43;
   std::mt19937 mtGen{mtSeed};
   bool pickOdd{false};

modules/cpu/fb/LocalFB.cpp

@@ -423,7 +423,7 @@ vec2i LocalFrameBuffer::getTaskStartPos(const uint32_t taskID) const
 devicert::AsyncEvent LocalFrameBuffer::postProcess()
 {
   // Calculate per-task variance if any samples accumulated into variance
-  // buffer, skip it if frameVariance overriden in writeTiles()
+  // buffer, skip it if frameVariance overridden in writeTiles()
   devicert::AsyncEvent event;
   const bool oddFrame = (getSh()->super.frameID & 1) == 1;
   if (varianceFrameOp && oddFrame && (frameVariance == float(inf)))

modules/cpu/lights/SunSkyLight.h

@@ -59,7 +59,7 @@ struct OSPRAY_SDK_INTERFACE SunSkyLight : public Light
   float intensityScale{0.025f};
 };
 
-// Inlined defintions /////////////////////////////////////////////////////////
+// Inlined definitions /////////////////////////////////////////////////////////
 
 inline uint32_t SunSkyLight::getShCount() const
 {

modules/cpu/render/RendererRenderTaskFn.inl

@@ -90,7 +90,7 @@ static void Renderer_default_renderTask(const uniform vec3ui itemIndex,
           make_vec3f(Renderer_getBackground(self, screenSample.pos, ffh));
       screenSample.normal = make_vec3f(0.f);
 
-      // The proper sample rendering function name is substitued here via macro
+      // The proper sample rendering function name is substituted here via macro
       renderSampleFn(self, fb, world, screenSample, ffh);
 
       col = col + screenSample.rgb;

modules/cpu/render/pathtracer/NextEventEstimation.ispc

@@ -55,7 +55,8 @@ SYCL_EXTERNAL vec3f nextEventEstimation(const PathContext &pathContext,
     foreach_unique (l in light)
       ls = Light_dispatch_sample(l, pathVertex.dg, s, pathState.time, ffh);
 
-    // adjust the contibution with the probabiltiy of selecting the light source
+    // adjust the contribution with the probabiltiy of selecting the light
+    // source
     ls.weight = ls.weight / lightSelectionProb;
     ls.pdf *= lightSelectionProb;
 

modules/cpu/render/pathtracer/TransparentShadow.ispc

@@ -26,7 +26,7 @@ SYCL_EXTERNAL __noinline vec3f transparentShadow(
     const uniform FeatureFlagsHandler &ffh)
 {
   const uniform FeatureFlags ff = getFeatureFlags(ffh);
-  // It's not possible to have transpareny shadows if we don't have any geometry
+  // It's not possible to have transparent shadows if we don't have any geometry
   if (!ff.geometry) {
     return lightContrib;
   }

modules/mpi/ospray/fb/DistributedFrameBuffer.cpp

@@ -439,7 +439,7 @@ void DistributedFrameBuffer::processMessage(MasterTileMessage_FB *msg)
     }
     // All IDs are sent if any were requested, however we need to write only the
     // ones that actually exist in the local FB since it doesn't allocate
-    // buffers for the non-existant channels
+    // buffers for the non-existent channels
     if (hasPrimitiveIDBuffer) {
       pidBuf = reinterpret_cast<uint32 *>(data);
     }

ospray/common/DeviceRT.h

@@ -273,7 +273,7 @@ inline Device &DeviceAllocator<T>::getDevice() const
 // Use it for data that needs the highest read/write performance on the device.
 // The buffer supports initialization from system (non-USM) memory. Preferably
 // should be used for all scene data that does not need to be updated from or
-// transfered back to the host.
+// transferred back to the host.
 template <typename T>
 struct BufferDevice
 {
@@ -319,7 +319,7 @@ BufferDevice<T>::BufferDevice(Device &device, const std::vector<T> &v)
       ptr(static_cast<T *>(device.deviceMalloc(v.size() * sizeof(T))))
 {
   // Schedule buffer copying and wait till completed so the input resource can
-  // be released immediatelly
+  // be released immediately
   AsyncEvent event = device.memcpy(ptr, v.data(), v.size() * sizeof(T));
   event.wait();
 }
@@ -331,7 +331,7 @@ BufferDevice<T>::BufferDevice(Device &device, T *data, std::size_t count)
       ptr(static_cast<T *>(device.deviceMalloc(count * sizeof(T))))
 {
   // Schedule buffer copying and wait till completed so the input resource can
-  // be released immediatelly
+  // be released immediately
   AsyncEvent event = device.memcpy(ptr, data, count * sizeof(T));
   event.wait();
 }
@@ -378,7 +378,7 @@ inline std::unique_ptr<BufferDevice<T>> make_buffer_device_unique(
 // Buffer in the device memory with host memory shadow
 
 // For usages that need the highest read/write efficiency on the device, similar
-// to `BufferDevice`, but additionally the data has to be transfered from/to
+// to `BufferDevice`, but additionally the data has to be transferred from/to
 // host periodically, e.g. framebuffers. After construction and initialization
 // of the host/shadow buffer, an explicit call to copyToDevice() is required so
 // the data will be visible for the device. Beware, it may use pinned (not

ospray/include/ospray/ospray_util.h

@@ -86,7 +86,7 @@ OSPRAY_INTERFACE void ospSetObjectAsData(OSPObject target,
 
 // Rendering helpers //////////////////////////////////////////////////////////
 
-// Start a frame task and immediately wait on it, return frame buffer varaince
+// Start a frame task and immediately wait on it, return frame buffer variance
 OSPRAY_INTERFACE float ospRenderFrameBlocking(
     OSPFrameBuffer, OSPRenderer, OSPCamera, OSPWorld);
 

scripts/files-to-md5.sh

@@ -46,7 +46,7 @@ set -e
 for FILE in $BASELINE_INPUT_IMAGES_DIR/*; do
     # Copy file to local tmp directory with new name (based on md5 from this img)
     cp $FILE $LOCAL_TMP_DIR/`md5 $FILE`
-    # Create/Filll metadata file in ospray repo
+    # Create/Fill metadata file in ospray repo
     # so we can link img in remote repo by this md5 string
     md5 $FILE > $BASELINE_MD5_OUTPUT_DIR/`basename $FILE`.md5
 done

scripts/superbuild/README.md

@@ -14,7 +14,7 @@ cmake [<OSPRAY_SOURCE_DIR>/scripts/superbuild]
 cmake --build .
 ```
 
-On Windows make sure to select a 64bit generator, e.g.
+On Windows make sure to select a 64\ bit generator, e.g.
 
 ```sh
 cmake -G "Visual Studio 17 2022" [<OSPRAY_SOURCE_DIR>/scripts/superbuild]