Clean up dpctl.tensor code

dpctl_ext/tensor/libtensor/include/kernels/constructors.hpp

@@ -33,8 +33,8 @@
 //===----------------------------------------------------------------------===//
 
 #pragma once
+
 #include <array>
-#include <complex>
 #include <cstddef>
 #include <vector>
 

dpctl_ext/tensor/libtensor/include/kernels/elementwise_functions/common.hpp

@@ -33,11 +33,8 @@
 #pragma once
 
 #include <algorithm>
-#include <cmath>
-#include <complex>
 #include <cstddef>
 #include <cstdint>
-#include <limits>
 #include <type_traits>
 #include <vector>
 

dpctl_ext/tensor/libtensor/include/kernels/elementwise_functions/logaddexp.hpp

@@ -46,7 +46,6 @@
 #include "vec_size_util.hpp"
 
 #include "utils/math_utils.hpp"
-#include "utils/offset_utils.hpp"
 #include "utils/type_dispatch_building.hpp"
 #include "utils/type_utils.hpp"
 

dpctl_ext/tensor/libtensor/include/kernels/elementwise_functions/maximum.hpp

@@ -45,7 +45,6 @@
 #include "vec_size_util.hpp"
 
 #include "utils/math_utils.hpp"
-#include "utils/offset_utils.hpp"
 #include "utils/type_dispatch_building.hpp"
 #include "utils/type_utils.hpp"
 

dpctl_ext/tensor/libtensor/include/kernels/elementwise_functions/minimum.hpp

@@ -44,7 +44,6 @@
 #include "vec_size_util.hpp"
 
 #include "utils/math_utils.hpp"
-#include "utils/offset_utils.hpp"
 #include "utils/type_dispatch_building.hpp"
 #include "utils/type_utils.hpp"
 

dpctl_ext/tensor/libtensor/include/kernels/integer_advanced_indexing.hpp

@@ -33,7 +33,7 @@
 //===----------------------------------------------------------------------===//
 
 #pragma once
-#include <complex>
+
 #include <cstddef>
 #include <type_traits>
 #include <vector>

dpctl_ext/tensor/libtensor/source/accumulators.cpp

@@ -32,17 +32,18 @@
 /// This file defines functions of dpctl.tensor._tensor_impl extensions
 //===----------------------------------------------------------------------===//
 
-#include <algorithm>
 #include <cstddef>
 #include <cstdint>
 #include <stdexcept>
+#include <string>
+#include <tuple>
+#include <utility>
 #include <vector>
 
 #include <sycl/sycl.hpp>
 
 #include "dpnp4pybind11.hpp"
 #include <pybind11/pybind11.h>
-#include <pybind11/stl.h>
 
 #include "kernels/accumulators.hpp"
 #include "simplify_iteration_space.hpp"
@@ -196,8 +197,8 @@ std::size_t py_mask_positions(const dpctl::tensor::usm_ndarray &mask,
     int mask_nd = mask.get_ndim();
     int nd = mask_nd;
 
-    dpctl::tensor::py_internal::compact_iteration_space(
-        nd, shape, strides_vector, compact_shape, compact_strides);
+    compact_iteration_space(nd, shape, strides_vector, compact_shape,
+                            compact_strides);
 
     // Strided implementation
     auto strided_fn =
@@ -351,8 +352,8 @@ std::size_t py_cumsum_1d(const dpctl::tensor::usm_ndarray &src,
     int src_nd = src.get_ndim();
     int nd = src_nd;
 
-    dpctl::tensor::py_internal::compact_iteration_space(
-        nd, shape, strides_vector, compact_shape, compact_strides);
+    compact_iteration_space(nd, shape, strides_vector, compact_shape,
+                            compact_strides);
 
     // Strided implementation
     auto strided_fn = cumsum_1d_strided_dispatch_vector[src_typeid];

dpctl_ext/tensor/libtensor/source/accumulators.hpp

@@ -39,7 +39,6 @@
 #include <sycl/sycl.hpp>
 
 #include "dpnp4pybind11.hpp"
-#include <pybind11/pybind11.h>
 
 namespace dpctl::tensor::py_internal
 {

dpctl_ext/tensor/libtensor/source/boolean_advanced_indexing.cpp

@@ -336,21 +336,19 @@ std::pair<sycl::event, sycl::event>
         shT masked_src_shape;
         shT ortho_src_strides;
         shT masked_src_strides;
-        dpctl::tensor::py_internal::split_iteration_space(
-            src_shape_vec, src_strides_vec, axis_start, axis_end,
-            ortho_src_shape,
-            masked_src_shape, // 4 vectors modified
-            ortho_src_strides, masked_src_strides);
+        split_iteration_space(src_shape_vec, src_strides_vec, axis_start,
+                              axis_end, ortho_src_shape,
+                              masked_src_shape, // 4 vectors modified
+                              ortho_src_strides, masked_src_strides);
 
         shT ortho_dst_shape;
         shT masked_dst_shape;
         shT ortho_dst_strides;
         shT masked_dst_strides;
-        dpctl::tensor::py_internal::split_iteration_space(
-            dst_shape_vec, dst_strides_vec, axis_start, axis_start + 1,
-            ortho_dst_shape,
-            masked_dst_shape, // 4 vectors modified
-            ortho_dst_strides, masked_dst_strides);
+        split_iteration_space(dst_shape_vec, dst_strides_vec, axis_start,
+                              axis_start + 1, ortho_dst_shape,
+                              masked_dst_shape, // 4 vectors modified
+                              ortho_dst_strides, masked_dst_strides);
 
         assert(ortho_src_shape.size() == static_cast<std::size_t>(ortho_nd));
         assert(ortho_dst_shape.size() == static_cast<std::size_t>(ortho_nd));
@@ -366,7 +364,7 @@ std::pair<sycl::event, sycl::event>
         py::ssize_t ortho_src_offset(0);
         py::ssize_t ortho_dst_offset(0);
 
-        dpctl::tensor::py_internal::simplify_iteration_space(
+        simplify_iteration_space(
             ortho_nd, _shape, ortho_src_strides, ortho_dst_strides,
             // output
             simplified_ortho_shape, simplified_ortho_src_strides,
@@ -646,21 +644,19 @@ std::pair<sycl::event, sycl::event>
         shT masked_dst_shape;
         shT ortho_dst_strides;
         shT masked_dst_strides;
-        dpctl::tensor::py_internal::split_iteration_space(
-            dst_shape_vec, dst_strides_vec, axis_start, axis_end,
-            ortho_dst_shape,
-            masked_dst_shape, // 4 vectors modified
-            ortho_dst_strides, masked_dst_strides);
+        split_iteration_space(dst_shape_vec, dst_strides_vec, axis_start,
+                              axis_end, ortho_dst_shape,
+                              masked_dst_shape, // 4 vectors modified
+                              ortho_dst_strides, masked_dst_strides);
 
         shT ortho_rhs_shape;
         shT masked_rhs_shape;
         shT ortho_rhs_strides;
         shT masked_rhs_strides;
-        dpctl::tensor::py_internal::split_iteration_space(
-            rhs_shape_vec, rhs_strides_vec, axis_start, axis_start + 1,
-            ortho_rhs_shape,
-            masked_rhs_shape, // 4 vectors modified
-            ortho_rhs_strides, masked_rhs_strides);
+        split_iteration_space(rhs_shape_vec, rhs_strides_vec, axis_start,
+                              axis_start + 1, ortho_rhs_shape,
+                              masked_rhs_shape, // 4 vectors modified
+                              ortho_rhs_strides, masked_rhs_strides);
 
         assert(ortho_dst_shape.size() == static_cast<std::size_t>(ortho_nd));
         assert(ortho_rhs_shape.size() == static_cast<std::size_t>(ortho_nd));
@@ -676,7 +672,7 @@ std::pair<sycl::event, sycl::event>
         py::ssize_t ortho_dst_offset(0);
         py::ssize_t ortho_rhs_offset(0);
 
-        dpctl::tensor::py_internal::simplify_iteration_space(
+        simplify_iteration_space(
             ortho_nd, _shape, ortho_dst_strides, ortho_rhs_strides,
             simplified_ortho_shape, simplified_ortho_dst_strides,
             simplified_ortho_rhs_strides, ortho_dst_offset, ortho_rhs_offset);

dpctl_ext/tensor/libtensor/source/copy_and_cast_usm_to_usm.cpp

@@ -188,11 +188,10 @@ std::pair<sycl::event, sycl::event> copy_usm_ndarray_into_usm_ndarray(
     const py::ssize_t *shape = src_shape;
 
     // nd, simplified_* and *_offset are modified by reference
-    dpctl::tensor::py_internal::simplify_iteration_space(
-        nd, shape, src_strides, dst_strides,
-        // output
-        simplified_shape, simplified_src_strides, simplified_dst_strides,
-        src_offset, dst_offset);
+    simplify_iteration_space(nd, shape, src_strides, dst_strides,
+                             // output
+                             simplified_shape, simplified_src_strides,
+                             simplified_dst_strides, src_offset, dst_offset);
 
     if (nd < 2) {
         if (nd == 1) {

dpctl_ext/tensor/libtensor/source/copy_as_contig.cpp

@@ -225,11 +225,11 @@ std::pair<sycl::event, sycl::event>
     int nd = src_nd;
 
     // nd, simplified_* and *_offset are modified by reference
-    dpctl::tensor::py_internal::simplify_iteration_space(
-        nd, src_shape_vec.data(), src_strides_vec, dst.get_strides_vector(),
-        // output
-        simplified_shape, simplified_src_strides, simplified_dst_strides,
-        src_offset, dst_offset);
+    simplify_iteration_space(nd, src_shape_vec.data(), src_strides_vec,
+                             dst.get_strides_vector(),
+                             // output
+                             simplified_shape, simplified_src_strides,
+                             simplified_dst_strides, src_offset, dst_offset);
 
     if (!((0 == src_offset) && (0 == dst_offset))) {
         throw std::runtime_error(
@@ -359,11 +359,11 @@ std::pair<sycl::event, sycl::event>
     int nd = src_nd;
 
     // nd, simplified_* and *_offset are modified by reference
-    dpctl::tensor::py_internal::simplify_iteration_space(
-        nd, src_shape_vec.data(), src_strides_vec, dst.get_strides_vector(),
-        // output
-        simplified_shape, simplified_src_strides, simplified_dst_strides,
-        src_offset, dst_offset);
+    simplify_iteration_space(nd, src_shape_vec.data(), src_strides_vec,
+                             dst.get_strides_vector(),
+                             // output
+                             simplified_shape, simplified_src_strides,
+                             simplified_dst_strides, src_offset, dst_offset);
 
     if (!((0 == src_offset) && (0 == dst_offset))) {
         throw std::runtime_error(
@@ -521,12 +521,11 @@ std::pair<sycl::event, sycl::event>
     int nd = static_cast<int>(batch_shape_vec.size());
 
     // nd, simplified_* and *_offset are modified by reference
-    dpctl::tensor::py_internal::simplify_iteration_space(
-        nd, batch_shape_vec.data(), src_batch_strides_vec,
-        dst_batch_strides_vec,
-        // output
-        simplified_shape, simplified_src_strides, simplified_dst_strides,
-        src_offset, dst_offset);
+    simplify_iteration_space(nd, batch_shape_vec.data(), src_batch_strides_vec,
+                             dst_batch_strides_vec,
+                             // output
+                             simplified_shape, simplified_src_strides,
+                             simplified_dst_strides, src_offset, dst_offset);
 
     if (!((0 == src_offset) && (0 == dst_offset))) {
         throw std::runtime_error(
@@ -714,12 +713,11 @@ std::pair<sycl::event, sycl::event>
     int nd = static_cast<int>(batch_shape_vec.size());
 
     // nd, simplified_* and *_offset are modified by reference
-    dpctl::tensor::py_internal::simplify_iteration_space(
-        nd, batch_shape_vec.data(), src_batch_strides_vec,
-        dst_batch_strides_vec,
-        // output
-        simplified_shape, simplified_src_strides, simplified_dst_strides,
-        src_offset, dst_offset);
+    simplify_iteration_space(nd, batch_shape_vec.data(), src_batch_strides_vec,
+                             dst_batch_strides_vec,
+                             // output
+                             simplified_shape, simplified_src_strides,
+                             simplified_dst_strides, src_offset, dst_offset);
 
     if (!((0 == src_offset) && (0 == dst_offset))) {
         throw std::runtime_error(

dpctl_ext/tensor/libtensor/source/copy_for_roll.cpp

@@ -197,11 +197,10 @@ std::pair<sycl::event, sycl::event>
     const py::ssize_t *shape = src_shape_ptr;
 
     // nd, simplified_* and *_offset are modified by reference
-    dpctl::tensor::py_internal::simplify_iteration_space(
-        nd, shape, src_strides, dst_strides,
-        // output
-        simplified_shape, simplified_src_strides, simplified_dst_strides,
-        src_offset, dst_offset);
+    simplify_iteration_space(nd, shape, src_strides, dst_strides,
+                             // output
+                             simplified_shape, simplified_src_strides,
+                             simplified_dst_strides, src_offset, dst_offset);
 
     if (nd == 1 && simplified_src_strides[0] == 1 &&
         simplified_dst_strides[0] == 1) {

dpctl_ext/tensor/libtensor/source/device_support_queries.cpp

@@ -36,7 +36,7 @@
 
 #include "dpnp4pybind11.hpp"
 #include <pybind11/pybind11.h>
-#include <pybind11/stl.h>
+
 #include <sycl/sycl.hpp>
 
 namespace dpctl::tensor::py_internal

dpctl_ext/tensor/libtensor/source/full_ctor.cpp

@@ -32,7 +32,6 @@
 /// This file defines functions of dpctl.tensor._tensor_impl extensions
 //===--------------------------------------------------------------------===//
 
-#include <complex>
 #include <cstddef>
 #include <cstdint>
 #include <tuple>
@@ -42,11 +41,13 @@
 #include <sycl/sycl.hpp>
 
 #include "dpnp4pybind11.hpp"
-#include <pybind11/complex.h>
+#include <pybind11/complex.h> // py::cast<std::complex<T>>
 #include <pybind11/pybind11.h>
 
 #include "kernels/constructors.hpp"
+#include "utils/offset_utils.hpp"
 #include "utils/output_validation.hpp"
+#include "utils/sycl_alloc_utils.hpp"
 #include "utils/type_dispatch.hpp"
 #include "utils/type_utils.hpp"
 

dpctl_ext/tensor/libtensor/source/integer_advanced_indexing.cpp

@@ -34,7 +34,6 @@
 //===----------------------------------------------------------------------===//
 
 #include <algorithm>
-#include <complex>
 #include <cstddef>
 #include <cstdint>
 #include <iterator>
@@ -47,9 +46,7 @@
 #include <sycl/sycl.hpp>
 
 #include "dpnp4pybind11.hpp"
-#include <pybind11/complex.h>
 #include <pybind11/pybind11.h>
-#include <pybind11/stl.h>
 
 #include "kernels/integer_advanced_indexing.hpp"
 #include "utils/memory_overlap.hpp"

dpctl_ext/tensor/libtensor/source/tensor_ctors.cpp

@@ -72,9 +72,6 @@ static_assert(std::is_same_v<py::ssize_t, dpctl::tensor::ssize_t>);
 namespace
 {
 
-using dpctl::tensor::c_contiguous_strides;
-using dpctl::tensor::f_contiguous_strides;
-
 using dpctl::tensor::overlap::MemoryOverlap;
 using dpctl::tensor::overlap::SameLogicalTensors;
 

dpctl_ext/tensor/libtensor/source/zeros_ctor.cpp

@@ -32,7 +32,6 @@
 /// This file defines functions of dpctl.tensor._tensor_impl extensions
 //===--------------------------------------------------------------------===//
 
-#include <complex>
 #include <cstddef>
 #include <stdexcept>
 #include <utility>
@@ -41,7 +40,6 @@
 #include <sycl/sycl.hpp>
 
 #include "dpnp4pybind11.hpp"
-#include <pybind11/complex.h>
 #include <pybind11/pybind11.h>
 
 #include "utils/output_validation.hpp"

dpctl_ext/tensor/libtensor/source/zeros_ctor.hpp

@@ -39,7 +39,6 @@
 #include <sycl/sycl.hpp>
 
 #include "dpnp4pybind11.hpp"
-#include <pybind11/pybind11.h>
 
 namespace dpctl::tensor::py_internal
 {