Merge master into gold/2021

.github/workflows/conda-package.yml

@@ -29,11 +29,8 @@ jobs:
 
     strategy:
       matrix:
-        python: ['3.8', '3.9']
+        python: ['3.8', '3.9', '3.10']
         os: [ubuntu-20.04, windows-latest]
-        include:
-          - python: '3.10'
-            os: ubuntu-20.04
 
     runs-on: ${{ matrix.os }}
 
@@ -110,7 +107,7 @@ jobs:
 
     strategy:
       matrix:
-        python: ['3.8', '3.9']
+        python: ['3.8', '3.9', '3.10']
         os: [ubuntu-20.04, ubuntu-latest]
 
         experimental: [false]
@@ -215,7 +212,7 @@ jobs:
 
     strategy:
       matrix:
-        python: ['3.8', '3.9']
+        python: ['3.8', '3.9', '3.10']
         experimental: [false]
 
     continue-on-error: ${{ matrix.experimental }}
@@ -384,7 +381,7 @@ jobs:
 
     strategy:
       matrix:
-        python: ['3.8', '3.9']
+        python: ['3.8', '3.9', '3.10']
         os: [ubuntu-20.04, windows-latest]
 
     runs-on: ${{ matrix.os }}

0.build.sh

@@ -4,6 +4,14 @@ THEDIR=$(dirname $(readlink -e ${BASH_SOURCE[0]}))
 # . ${THEDIR}/0.env.sh
 cd ${THEDIR}
 
+# Assign $TMP env variable to a directory where the script locates.
+# The env variable is used by compiler as a path to temporary folder,
+# where it can store a temporary files generated during compilation and linkage phases.
+# By default the compiler uses /tmp folder, but it is limited by the size and
+# there might be not enough space to temporary keep all generated data.
+export TMP=${THEDIR}
+
+
 export DPNP_DEBUG=1
 
 python setup.py clean
@@ -17,7 +25,8 @@ CC=icpx python setup.py build_ext --inplace
 
 echo
 echo =========example3==============
-icpx -fsycl -g -fPIC dpnp/backend/examples/example3.cpp -Idpnp -Idpnp/backend/include -Ldpnp -Wl,-rpath='$ORIGIN'/dpnp -ldpnp_backend_c -o example3
+DPCTL_INCLUDES=$(python -m dpctl --includes)
+icpx -fsycl -g -O0 -ggdb3 -fPIC dpnp/backend/examples/example3.cpp $DPCTL_INCLUDES -Idpnp -Idpnp/backend/include -Ldpnp -Wl,-rpath='$ORIGIN'/dpnp -ldpnp_backend_c -o example3
 # LD_DEBUG=libs,bindings,symbols ./example3
 ./example3
 
@@ -39,7 +48,7 @@ icpx -fsycl -g -fPIC dpnp/backend/examples/example3.cpp -Idpnp -Idpnp/backend/in
 # strings /usr/share/miniconda/envs/dpnp*/lib/libstdc++.so | grep GLIBCXX | sort -n
 
 
-# echo
+echo
 echo =========example1==============
 # LD_DEBUG=libs,bindings,symbols python examples/example1.py
 # LD_DEBUG=libs python examples/example1.py

conda-recipe/build.sh

@@ -29,6 +29,11 @@ fi
 export CFLAGS="-Wl,-rpath,\$ORIGIN/../dpctl,-rpath,\$ORIGIN $CFLAGS"
 export LDFLAGS="-Wl,-rpath,\$ORIGIN/../dpctl,-rpath,\$ORIGIN $LDFLAGS"
 
+# Intel LLVM must cooperate with compiler and sysroot from conda
+echo "--gcc-toolchain=${BUILD_PREFIX} --sysroot=${BUILD_PREFIX}/${HOST}/sysroot -target ${HOST}" > icpx_for_conda.cfg
+export ICPXCFG="$(pwd)/icpx_for_conda.cfg"
+export ICXCFG="$(pwd)/icpx_for_conda.cfg"
+
 $PYTHON setup.py build_clib
 $PYTHON setup.py build_ext install
 

conda-recipe/meta.yaml

@@ -11,17 +11,17 @@ requirements:
       - numpy 1.19
       - cython
       - cmake >=3.19
-      - dpctl >=0.13
-      - mkl-devel-dpcpp {{ environ.get('MKL_VER', '>=2021.1.1') }}
+      - dpctl >=0.14
+      - mkl-devel-dpcpp {{ environ.get('MKL_VER', '>=2023.0.0') }}
       - onedpl-devel
       - tbb-devel
       - wheel
     build:
       - {{ compiler('cxx') }}
-      - {{ compiler('dpcpp') }}  >=2022.1  # [not osx]
+      - {{ compiler('dpcpp') }}  >=2023.0  # [not osx]
     run:
       - python
-      - dpctl >=0.13
+      - dpctl >=0.14
       - {{ pin_compatible('dpcpp-cpp-rt', min_pin='x.x', max_pin='x') }}
       - {{ pin_compatible('mkl-dpcpp', min_pin='x.x', max_pin='x') }}
       - {{ pin_compatible('numpy', min_pin='x.x', max_pin='x') }}

doc/conf.py

@@ -33,7 +33,7 @@
 # The short X.Y version
 version = '0.11'
 # The full version, including alpha/beta/rc tags
-release = '0.11.0'
+release = '0.11.1'
 
 
 # -- General configuration ---------------------------------------------------

dpnp/backend/CMakeLists.txt

@@ -1,5 +1,5 @@
 # *****************************************************************************
-# Copyright (c) 2016-2022, Intel Corporation
+# Copyright (c) 2016-2023, Intel Corporation
 # All rights reserved.
 #
 # Redistribution and use in source and binary forms, with or without
@@ -27,7 +27,7 @@
 
 cmake_minimum_required(VERSION 3.10 FATAL_ERROR)
 
-# set(DPNP_VERSION 0.11.0)
+# set(DPNP_VERSION 0.11.1)
 # set(DPNP_API_VERSION 0.11)
 
 # set directory where the custom finders live
@@ -93,6 +93,7 @@ string(CONCAT COMMON_COMPILE_FLAGS
   "-fsycl "
   "-fsycl-device-code-split=per_kernel "
   "-fno-approx-func "
+  "-fno-finite-math-only "
 )
 string(CONCAT COMMON_LINK_FLAGS
   "-fsycl "
@@ -111,7 +112,7 @@ elseif(WIN32)
   # set(CMAKE_RANLIB "llvm-ranlib")
   # set(CMAKE_CXX_FLAGS "/EHsc")
 
-  string(APPEND COMMON_COMPILER_FLAGS
+  string(APPEND COMMON_COMPILE_FLAGS
     "/EHsc "
 #    "/Ox "
 #    "/W3 "
@@ -133,23 +134,29 @@ string(CONCAT DPNP_WARNING_FLAGS
   "-Wextra "
   "-Wshadow "
   "-Wall "
-  "-Wstring-prototypes "
+  "-Wstrict-prototypes "
   "-Wformat "
   "-Wformat-security "
 )
-string(APPEND COMMON_COMPILER_FLAGS
+string(APPEND COMMON_COMPILE_FLAGS
   "${DPNP_WARNING_FLAGS}"
 )
 
 # debug/release compile definitions
 if(DPNP_DEBUG_ENABLE)
   set(CMAKE_BUILD_TYPE "Debug")
-  string(APPEND COMMON_COMPILER_FLAGS
+  string(APPEND COMMON_COMPILE_FLAGS
     "-O0 "
+    "-ggdb3 "
+  )
+  string(APPEND COMMON_LINK_FLAGS
+    "-O0 "
+    "-ggdb3 "
+    "-fsycl-link-huge-device-code "
   )
 else()
   set(CMAKE_BUILD_TYPE "Release")
-  string(APPEND COMMON_COMPILER_FLAGS
+  string(APPEND COMMON_COMPILE_FLAGS
     "-O3 "
   )
 endif()
@@ -162,7 +169,7 @@ string(CONCAT DPNP_DEFS
   "-D_FORTIFY_SOURCE=2 "
 )
 if(NOT WIN32)
-  string(APPEND COMMON_COMPILER_FLAGS
+  string(APPEND COMMON_COMPILE_FLAGS
     "-fno-delete-null-pointer-checks "
     "-fstack-protector-strong "
     "-fno-strict-overflow "

dpnp/backend/doc/Doxyfile

@@ -38,7 +38,7 @@ PROJECT_NAME           = "DPNP C++ backend kernel library"
 # could be handy for archiving the generated documentation or if some version
 # control system is used.
 
-PROJECT_NUMBER         = 0.11.0
+PROJECT_NUMBER         = 0.11.1
 
 # Using the PROJECT_BRIEF tag one can provide an optional one line description
 # for a project that appears at the top of each page and should give viewer a

dpnp/backend/include/dpnp_gen_1arg_1type_tbl.hpp

@@ -1,5 +1,5 @@
 //*****************************************************************************
-// Copyright (c) 2016-2020, Intel Corporation
+// Copyright (c) 2016-2023, Intel Corporation
 // All rights reserved.
 //
 // Redistribution and use in source and binary forms, with or without
@@ -23,6 +23,8 @@
 // THE POSSIBILITY OF SUCH DAMAGE.
 //*****************************************************************************
 
+#if defined(MACRO_1ARG_1TYPE_OP)
+
 /*
  * This header file contains single argument element wise functions definitions
  *
@@ -35,10 +37,6 @@
  *
  */
 
-#ifndef MACRO_1ARG_1TYPE_OP
-#error "MACRO_1ARG_1TYPE_OP is not defined"
-#endif
-
 #ifdef _SECTION_DOCUMENTATION_GENERATION_
 
 #define MACRO_1ARG_1TYPE_OP(__name__, __operation1__, __operation2__)                                                   \
@@ -88,7 +86,7 @@
                   const shape_elem_type* input1_strides,                                                                \
                   const size_t* where);
 
-#endif
+#endif // _SECTION_DOCUMENTATION_GENERATION_
 
 MACRO_1ARG_1TYPE_OP(dpnp_conjugate_c, std::conj(input_elem), q.submit(kernel_func))
 MACRO_1ARG_1TYPE_OP(dpnp_copy_c, input_elem, q.submit(kernel_func))
@@ -107,3 +105,62 @@ MACRO_1ARG_1TYPE_OP(dpnp_square_c,
                     oneapi::mkl::vm::sqr(q, input1_size, input1_data, result))
 
 #undef MACRO_1ARG_1TYPE_OP
+
+#elif defined(MACRO_1ARG_1TYPE_LOGIC_OP)
+
+/*
+ * This header file contains single argument element wise functions definitions
+ *
+ * Macro `MACRO_1ARG_1TYPE_LOGIC_OP` must be defined before usage
+ *
+ * Parameters:
+ * - public name of the function and kernel name
+ * - operation used to calculate the result
+ *
+ */
+
+#ifdef _SECTION_DOCUMENTATION_GENERATION_
+
+#define MACRO_1ARG_1TYPE_LOGIC_OP(__name__, __operation__)                                                             \
+    /** @ingroup BACKEND_API                                                                                         */ \
+    /** @brief Per element operation function __name__                                                               */ \
+    /**                                                                                                              */ \
+    /** Function "__name__" executes operator "__operation__" over corresponding elements of input array             */ \
+    /**                                                                                                              */ \
+    /** @param[in]  q_ref              Reference to SYCL queue.                                                      */ \
+    /** @param[out] result_out         Output array.                                                                 */ \
+    /** @param[in]  result_size        Output array size.                                                            */ \
+    /** @param[in]  result_ndim        Number of output array dimensions.                                            */ \
+    /** @param[in]  result_shape       Output array shape.                                                           */ \
+    /** @param[in]  result_strides     Output array strides.                                                         */ \
+    /** @param[in]  input1_in          Input array 1.                                                                */ \
+    /** @param[in]  input1_size        Input array 1 size.                                                           */ \
+    /** @param[in]  input1_ndim        Number of input array 1 dimensions.                                           */ \
+    /** @param[in]  input1_shape       Input array 1 shape.                                                          */ \
+    /** @param[in]  input1_strides     Input array 1 strides.                                                        */ \
+    /** @param[in]  where              Where condition.                                                              */ \
+    /** @param[in]  dep_event_vec_ref  Reference to vector of SYCL events.                                           */ \
+    template <typename _DataType_input1>                                                                                \
+    DPCTLSyclEventRef __name__(DPCTLSyclQueueRef q_ref,                                                                 \
+                               void* result_out,                                                                        \
+                               const size_t result_size,                                                                \
+                               const size_t result_ndim,                                                                \
+                               const shape_elem_type* result_shape,                                                     \
+                               const shape_elem_type* result_strides,                                                   \
+                               const void* input1_in,                                                                   \
+                               const size_t input1_size,                                                                \
+                               const size_t input1_ndim,                                                                \
+                               const shape_elem_type* input1_shape,                                                     \
+                               const shape_elem_type* input1_strides,                                                   \
+                               const size_t* where,                                                                     \
+                               const DPCTLEventVectorRef dep_event_vec_ref);
+
+#endif // _SECTION_DOCUMENTATION_GENERATION_
+
+MACRO_1ARG_1TYPE_LOGIC_OP(dpnp_logical_not_c, !input1_elem)
+
+#undef MACRO_1ARG_1TYPE_LOGIC_OP
+
+#else
+#error "MACRO_1ARG_1TYPE_OP or MACRO_1ARG_1TYPE_LOGIC_OP is not defined"
+#endif // MACRO_1ARG_1TYPE_OP || MACRO_1ARG_1TYPE_LOGIC_OP