Merge branch 'release-1.3.x'

RenderKit · May 11, 2022 · ec551ea · ec551ea
2 parents 5da393e + c2151b5
commit ec551ea
Show file tree

Hide file tree

Showing 305 changed files with 2,973 additions and 1,459 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,6 +1,19 @@
 Version History
 ---------------
 
+### Open VKL 1.3.0
+
+-   Added AVX512 8-wide CPU device mode, enabled via the
+    `OPENVKL_ISA_AVX512SKX_8_WIDE` CMake option
+-   VDB volumes: added support for packed / contiguous data layouts for
+    temporally constant volumes, which can provide improved performance
+    (`nodesPackedDense`, `nodesPackedTile` parameters)
+-   VDB utility library: added `repackNodes` flag to toggle usage of packed data
+    layouts
+-   Particle volumes: general memory efficiency and performance improvements
+-   Superbuild updates to latest versions of dependencies
+-   Minimum ISPC version is now v1.18.0
+
 ### Open VKL 1.2.0
 
 -   Added `vklSetParam()` API function which can set parameters of any supported

diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -1,4 +1,4 @@
-## Copyright 2019-2021 Intel Corporation
+## Copyright 2019 Intel Corporation
 ## SPDX-License-Identifier: Apache-2.0
 
 ## Global CMake options ##
@@ -19,7 +19,7 @@ set(CMAKE_POSITION_INDEPENDENT_CODE ON)
 
 ## Establish project ##
 
-project(openvkl VERSION 1.2.0 LANGUAGES C CXX)
+project(openvkl VERSION 1.3.0 LANGUAGES C CXX)
 
 ## Add openvkl specific macros ##
 

diff --git a/README.md b/README.md
@@ -1,6 +1,6 @@
 # Intel® Open Volume Kernel Library
 
-This is release v1.2.0 of Intel® Open VKL. For changes and new features
+This is release v1.3.0 of Intel® Open VKL. For changes and new features
 see the [changelog](CHANGELOG.md). Visit http://www.openvkl.org for more
 information.
 
@@ -33,6 +33,20 @@ example renderers to demonstrate how to best use the Open VKL API.
 
 ## Version History
 
+### Open VKL 1.3.0
+
+  - Added AVX512 8-wide CPU device mode, enabled via the
+    `OPENVKL_ISA_AVX512SKX_8_WIDE` CMake option
+  - VDB volumes: added support for packed / contiguous data layouts for
+    temporally constant volumes, which can provide improved performance
+    (`nodesPackedDense`, `nodesPackedTile` parameters)
+  - VDB utility library: added `repackNodes` flag to toggle usage of
+    packed data layouts
+  - Particle volumes: general memory efficiency and performance
+    improvements
+  - Superbuild updates to latest versions of dependencies
+  - Minimum ISPC version is now v1.18.0
+
 ### Open VKL 1.2.0
 
   - Added `vklSetParam()` API function which can set parameters of any
@@ -550,7 +564,7 @@ variables described previously.
 
 ## Basic data types
 
-Open VKL defines 3-component vectors of integer and vector types:
+Open VKL defines 3-component vectors of integer and float types:
 
 ``` cpp
 typedef struct
@@ -600,8 +614,8 @@ typedef struct
 
 Objects in Open VKL are exposed to the APIs as handles with internal
 reference counting for lifetime determination. Objects are created with
-particular type’s `vklNew...` API entry point. For example, `vklNewData`
-and `vklNewVolume`.
+each particular type’s `vklNew...` API entry point. For example,
+`vklNewData` and `vklNewVolume`.
 
 In general, modifiable parameters to objects are modified using
 `vklSet...` functions based on the type of the parameter being set. The
@@ -775,7 +789,7 @@ called concurrently on the same object.
 
 Open VKL currently supports structured volumes on regular and spherical
 grids; unstructured volumes with tetrahedral, wedge, pyramid, and
-hexaderal primitive types; adaptive mesh refinement (AMR) volumes;
+hexahedral primitive types; adaptive mesh refinement (AMR) volumes;
 sparse VDB volumes; and particle volumes. Volumes are created with
 `vklNewVolume` with a device and appropriate type string:
 
@@ -1120,13 +1134,18 @@ VDB volumes are created by passing the type string `"vdb"` to
 | int\[\]            | node.temporallyStructuredNumTimesteps |                                    | For temporally structured variation, number of timesteps per voxel. Only valid if `temporalFormat` is `VKL_TEMPORAL_FORMAT_STRUCTURED`.                                                                                                                                                                                                                                                                                                                                                                                               |
 | VKLData\[\]        | node.temporallyUnstructuredIndices    |                                    | For temporally unstructured variation, beginning per voxel. Supported data types for each node are `VKL_UINT` and `VKL_ULONG`. Only valid if `temporalFormat` is `VKL_TEMPORAL_FORMAT_UNSTRUCTURED`.                                                                                                                                                                                                                                                                                                                                  |
 | VKLData\[\]        | node.temporallyUnstructuredTimes      |                                    | For temporally unstructured variation, time values corresponding to values in `node.data`. For each node, the data must be of type `VKL_FLOAT`. Only valid if `temporalFormat` is `VKL_TEMPORAL_FORMAT_UNSTRUCTURED`.                                                                                                                                                                                                                                                                                                                 |
+| VKLData\[\]        | nodesPackedDense                      |                                    | Optionally provided instead of `node.data`, for each attribute a single array of all dense node data (`VKL_FORMAT_DENSE_ZYX` only) in a contiguous layout, provided in the same order as the corresponding `node.*` parameters. This packed layout may be more performant. Supported for temporally constant data only.                                                                                                                                                                                                               |
+| VKLData\[\]        | nodesPackedTile                       |                                    | Optionally provided instead of `node.data`, for each attribute a single array of all tile node data (`VKL_FORMAT_TILE` only) in a contiguous layout, provided in the same order as the corresponding `node.*` parameters. This packed layout may be more performant. Supported for temporally constant data only.                                                                                                                                                                                                                     |
 | float\[\]          | background                            | `VKL_BACKGROUND_UNDEFINED`         | For each attribute, the value that is returned when sampling an undefined region outside the volume domain.                                                                                                                                                                                                                                                                                                                                                                                                                           |
 | box3i              | indexClippingBounds                   |                                    | Clips the volume to the specified index-space bounding box. This is useful for volumes with dimensions that are not even multiples of the leaf node dimensions, or .vdb files with restrictive active voxel bounding boxes.                                                                                                                                                                                                                                                                                                           |
 
 Configuration parameters for VDB (`"vdb"`) volumes.
 
 The level, origin, format, and data parameters must have the same size,
-and there must be at least one valid node or `commit()` will fail.
+and there must be at least one valid node or `commit()` will fail. The
+`nodesPackedDense` and `nodesPackedTile` parameters may be provided
+instead of `node.data`; this packed data layout may provide better
+performance.
 
 VDB volumes support temporally structured and temporally unstructured
 temporal variation. See section ‘Temporal Variation’ for more detail.
@@ -2476,7 +2495,7 @@ before you can build Open VKL you need the following prerequisites:
     examples, you should also have some version of OpenGL.
 
   - Additionally you require a copy of the [Intel® Implicit SPMD Program
-    Compiler (Intel® ISPC)](http://ispc.github.io), version 1.16.0 or
+    Compiler (Intel® ISPC)](http://ispc.github.io), version 1.18.0 or
     later. Please obtain a release of ISPC from the [ISPC downloads
     page](https://ispc.github.io/downloads.html).
 

diff --git a/cmake/openvklConfig.cmake.in b/cmake/openvklConfig.cmake.in
@@ -1,4 +1,4 @@
-## Copyright 2019-2021 Intel Corporation
+## Copyright 2019 Intel Corporation
 ## SPDX-License-Identifier: Apache-2.0
 
 @PACKAGE_INIT@

diff --git a/cmake/openvkl_ispc.cmake b/cmake/openvkl_ispc.cmake
@@ -1,10 +1,10 @@
-## Copyright 2019-2021 Intel Corporation
+## Copyright 2019 Intel Corporation
 ## SPDX-License-Identifier: Apache-2.0
 
 option(OPENVKL_ISPC_FAST_MATH "enable ISPC fast-math optimizations" OFF)
 
 # ISPC versions to look for, in decending order (newest first)
-set(ISPC_VERSION_WORKING "1.16.0")
+set(ISPC_VERSION_WORKING "1.18.0")
 list(GET ISPC_VERSION_WORKING -1 ISPC_VERSION_REQUIRED)
 
 if (NOT ISPC_EXECUTABLE)
@@ -89,9 +89,11 @@ macro(openvkl_configure_ispc_isa)
     option(OPENVKL_ISA_AVX2 "Enables AVX2 ISA." ON)
     option(OPENVKL_ISA_AVX512KNL "Enables AVX512 ISA for Knights Landing." OFF)
     option(OPENVKL_ISA_AVX512SKX "Enables AVX512 ISA for Skylake." ON)
+    option(OPENVKL_ISA_AVX512SKX_8_WIDE "Enables AVX512 ISA for Skylake (8 wide)." OFF)
 
-    if (OPENVKL_ISA_AVX512KNL AND OPENVKL_ISA_AVX512SKX)
-      message(FATAL_ERROR "Only one AVX512 ISA may be enabled; choose either AVX512KNL or AVX512SKX")
+    if ((OPENVKL_ISA_AVX512KNL AND (OPENVKL_ISA_AVX512SKX OR OPENVKL_ISA_AVX512SKX_8_WIDE)) OR
+        (OPENVKL_ISA_AVX512SKX AND OPENVKL_ISA_AVX512SKX_8_WIDE))
+      message(FATAL_ERROR "Only one AVX512 ISA may be enabled; choose either AVX512KNL, AVX512SKX or AVX512SKX_8_WIDE")
     endif()
   else()
     option(OPENVKL_ISA_NEON "Enables NEON ISA." ON)
@@ -140,6 +142,12 @@ macro(openvkl_configure_ispc_isa)
     message(STATUS "OpenVKL AVX512SKX ISA target enabled.")
   endif()
 
+  if (OPENVKL_ISA_AVX512SKX_8_WIDE)
+    set(OPENVKL_ISPC_TARGET_LIST ${OPENVKL_ISPC_TARGET_LIST} avx512skx-i32x8)
+    set(OPENVKL_ISPC_TARGET_LIST_8 ${OPENVKL_ISPC_TARGET_LIST_8} avx512skx-i32x8)
+    message(STATUS "OpenVKL AVX512SKX_8_WIDE ISA target enabled.")
+  endif()
+
   # if only one target is specified for a given width, add a second target to
   # force ISPC name mangling. this avoids global name conflicts between devices
   # of different widths.
@@ -250,7 +258,8 @@ macro (OPENVKL_ISPC_COMPILE)
     list(LENGTH ISPC_TARGETS NUM_TARGETS)
     if (NUM_TARGETS GREATER 1)
       foreach(target ${ISPC_TARGETS})
-        string(REPLACE "-i32x16" "" target ${target}) # strip avx512(knl|skx)-i32x16
+        string(REPLACE "-i32x8" "" target ${target}) # strip "-i32x8"
+        string(REPLACE "-i32x16" "" target ${target}) # strip "-i32x16"
         set(results ${results} "${outdir}/${fname}.dev_${target}${ISPC_TARGET_EXT}")
       endforeach()
     endif()

diff --git a/cmake/openvkl_macros.cmake b/cmake/openvkl_macros.cmake
@@ -1,4 +1,4 @@
-## Copyright 2019-2022 Intel Corporation
+## Copyright 2019 Intel Corporation
 ## SPDX-License-Identifier: Apache-2.0
 
 macro(openvkl_add_library_ispc name type)
@@ -56,6 +56,10 @@ macro(openvkl_configure_global_build_flags)
 
      set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-strict-aliasing")
 
+     if(CMAKE_CXX_COMPILER_ID STREQUAL "IntelLLVM")
+       set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fp-model=precise")
+     endif()
+
   elseif(CMAKE_CXX_COMPILER_ID STREQUAL "Intel")
     if(NOT WIN32)
       set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -no-ansi-alias")

diff --git a/doc/api.md b/doc/api.md
@@ -181,7 +181,7 @@ previously.
 Basic data types
 ----------------
 
-Open VKL defines 3-component vectors of integer and vector types:
+Open VKL defines 3-component vectors of integer and float types:
 
     typedef struct
     {
@@ -225,8 +225,8 @@ Object model
 ------------
 
 Objects in Open VKL are exposed to the APIs as handles with internal reference
-counting for lifetime determination.  Objects are created with particular type's
-`vklNew...` API entry point. For example, `vklNewData` and `vklNewVolume`.
+counting for lifetime determination.  Objects are created with each particular
+type's `vklNew...` API entry point. For example, `vklNewData` and `vklNewVolume`.
 
 In general, modifiable parameters to objects are modified using `vklSet...`
 functions based on the type of the parameter being set. The parameter name is
@@ -417,7 +417,7 @@ Volume types
 ------------
 
 Open VKL currently supports structured volumes on regular and spherical grids;
-unstructured volumes with tetrahedral, wedge, pyramid, and hexaderal primitive
+unstructured volumes with tetrahedral, wedge, pyramid, and hexahedral primitive
 types; adaptive mesh refinement (AMR) volumes; sparse VDB volumes; and particle
 volumes.  Volumes are created with `vklNewVolume` with a device and appropriate
 type string:
@@ -928,6 +928,28 @@ following parameters:
                                                                                        Only valid if `temporalFormat` is
                                                                                        `VKL_TEMPORAL_FORMAT_UNSTRUCTURED`.
 
+  VKLData[]     nodesPackedDense                                                       Optionally provided instead of
+                                                                                       `node.data`, for each attribute a
+                                                                                       single array of all dense node data
+                                                                                       (`VKL_FORMAT_DENSE_ZYX` only) in a
+                                                                                       contiguous layout, provided in the
+                                                                                       same order as the corresponding
+                                                                                       `node.*` parameters. This packed
+                                                                                       layout may be more performant.
+                                                                                       Supported for temporally constant data
+                                                                                       only.
+
+  VKLData[]     nodesPackedTile                                                        Optionally provided instead of
+                                                                                       `node.data`, for each attribute a
+                                                                                       single array of all tile node data
+                                                                                       (`VKL_FORMAT_TILE` only) in a
+                                                                                       contiguous layout, provided in the
+                                                                                       same order as the corresponding
+                                                                                       `node.*` parameters. This packed
+                                                                                       layout may be more performant.
+                                                                                       Supported for temporally constant data
+                                                                                       only.
+
   float[]       background                             `VKL_BACKGROUND_UNDEFINED`      For each attribute, the value that is
                                                                                        returned when sampling an undefined
                                                                                        region outside the volume domain.
@@ -943,7 +965,9 @@ following parameters:
   : Configuration parameters for VDB (`"vdb"`) volumes.
 
 The level, origin, format, and data parameters must have the same size, and there must
-be at least one valid node or `commit()` will fail.
+be at least one valid node or `commit()` will fail. The `nodesPackedDense` and
+`nodesPackedTile` parameters may be provided instead of `node.data`; this packed data
+layout may provide better performance.
 
 VDB volumes support temporally structured and temporally unstructured temporal
 variation. See section 'Temporal Variation' for more detail.

diff --git a/doc/compilation.md b/doc/compilation.md
@@ -30,7 +30,7 @@ you can build Open VKL you need the following prerequisites:
     have some version of OpenGL.
 
 -   Additionally you require a copy of the [Intel® Implicit SPMD Program
-    Compiler (Intel® ISPC)](http://ispc.github.io), version 1.16.0 or later.
+    Compiler (Intel® ISPC)](http://ispc.github.io), version 1.18.0 or later.
     Please obtain a release of ISPC from the [ISPC downloads
     page](https://ispc.github.io/downloads.html).
 

diff --git a/examples/CMakeLists.txt b/examples/CMakeLists.txt
@@ -1,4 +1,4 @@
-## Copyright 2019-2022 Intel Corporation
+## Copyright 2019 Intel Corporation
 ## SPDX-License-Identifier: Apache-2.0
 
 ## "Hello world" VKL tutorials ##

diff --git a/examples/from_openvkl_install/CMakeLists.txt b/examples/from_openvkl_install/CMakeLists.txt
@@ -1,4 +1,4 @@
-## Copyright 2019-2021 Intel Corporation
+## Copyright 2019 Intel Corporation
 ## SPDX-License-Identifier: Apache-2.0
 
 # NOTE: This CMakeLists.txt is intended to be used to exercise an OpenVKL

diff --git a/examples/interactive/CMakeLists.txt b/examples/interactive/CMakeLists.txt
@@ -1,4 +1,4 @@
-## Copyright 2019-2021 Intel Corporation
+## Copyright 2019 Intel Corporation
 ## SPDX-License-Identifier: Apache-2.0
 
 add_subdirectory(renderer)

diff --git a/examples/interactive/ParameterGui.cpp b/examples/interactive/ParameterGui.cpp
@@ -1,4 +1,4 @@
-// Copyright 2021-2022 Intel Corporation
+// Copyright 2021 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #include "ParameterGui.h"

diff --git a/examples/interactive/TransferFunctionWidget.cpp b/examples/interactive/TransferFunctionWidget.cpp
@@ -1,4 +1,4 @@
-// Copyright 2019-2021 Intel Corporation
+// Copyright 2019 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #include "TransferFunctionWidget.h"

diff --git a/examples/interactive/TransferFunctionWidget.h b/examples/interactive/TransferFunctionWidget.h
@@ -1,4 +1,4 @@
-// Copyright 2019-2021 Intel Corporation
+// Copyright 2019 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #pragma once

diff --git a/examples/interactive/renderer/CMakeLists.txt b/examples/interactive/renderer/CMakeLists.txt
@@ -1,4 +1,4 @@
-## Copyright 2019-2021 Intel Corporation
+## Copyright 2019 Intel Corporation
 ## SPDX-License-Identifier: Apache-2.0
 
 include_directories_ispc(

diff --git a/examples/interactive/renderer/DensityPathTracer.ispc b/examples/interactive/renderer/DensityPathTracer.ispc
@@ -1,4 +1,4 @@
-// Copyright 2019-2021 Intel Corporation
+// Copyright 2019 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #include "Random.ih"

diff --git a/examples/interactive/renderer/HitIteratorRenderer.ispc b/examples/interactive/renderer/HitIteratorRenderer.ispc
@@ -1,4 +1,4 @@
-// Copyright 2019-2021 Intel Corporation
+// Copyright 2019 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #include "Ray.ih"

diff --git a/examples/interactive/renderer/IntervalIteratorDebug.cpp b/examples/interactive/renderer/IntervalIteratorDebug.cpp
@@ -1,4 +1,4 @@
-// Copyright 2021-2022 Intel Corporation
+// Copyright 2021 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #include "IntervalIteratorDebug.h"

diff --git a/examples/interactive/renderer/IntervalIteratorDebug.h b/examples/interactive/renderer/IntervalIteratorDebug.h
@@ -1,4 +1,4 @@
-// Copyright 2021-2022 Intel Corporation
+// Copyright 2021 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #pragma once

diff --git a/examples/interactive/renderer/IntervalIteratorDebug.ispc b/examples/interactive/renderer/IntervalIteratorDebug.ispc
@@ -1,4 +1,4 @@
-// Copyright 2019-2021 Intel Corporation
+// Copyright 2019 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #include "Ray.ih"

diff --git a/examples/interactive/renderer/Random.ih b/examples/interactive/renderer/Random.ih
@@ -1,4 +1,4 @@
-// Copyright 2019-2021 Intel Corporation
+// Copyright 2019 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #pragma once

diff --git a/examples/interactive/renderer/Ray.ih b/examples/interactive/renderer/Ray.ih
@@ -1,4 +1,4 @@
-// Copyright 2019-2021 Intel Corporation
+// Copyright 2019 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #pragma once

diff --git a/examples/interactive/renderer/RayMarchIteratorRenderer.cpp b/examples/interactive/renderer/RayMarchIteratorRenderer.cpp
@@ -1,4 +1,4 @@
-// Copyright 2021-2022 Intel Corporation
+// Copyright 2021 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #include "RayMarchIteratorRenderer.h"

diff --git a/examples/interactive/renderer/RayMarchIteratorRenderer.h b/examples/interactive/renderer/RayMarchIteratorRenderer.h
@@ -1,4 +1,4 @@
-// Copyright 2021-2022 Intel Corporation
+// Copyright 2021 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #pragma once

diff --git a/examples/interactive/renderer/RayMarchIteratorRenderer.ispc b/examples/interactive/renderer/RayMarchIteratorRenderer.ispc
@@ -1,4 +1,4 @@
-// Copyright 2019-2021 Intel Corporation
+// Copyright 2019 Intel Corporation
 // SPDX-License-Identifier: Apache-2.0
 
 #include "Ray.ih"