Add NonParametric & Parametric Gates support to Lightning.Tensor

.github/CHANGELOG.md

@@ -2,6 +2,9 @@
 
 ### New features since last release
 
+* Add gate support to `cutensornet` backed `lightning.tensor` C++ layer.
+  [(#718)](https://github.com/PennyLaneAI/pennylane-lightning/pull/718)
+
 * Add `cutensornet` backed `MPS` C++ layer to `lightning.tensor`.
   [(#704)](https://github.com/PennyLaneAI/pennylane-lightning/pull/704)
 

pennylane_lightning/core/_version.py

@@ -16,4 +16,4 @@
    Version number (major.minor.patch[-label])
 """
 
-__version__ = "0.37.0-dev19"
+__version__ = "0.37.0-dev20"

pennylane_lightning/core/src/CMakeLists.txt

@@ -14,14 +14,10 @@ set(COMPONENT_SUBDIRS   algorithms
                         utils
 )
 
-if("${PL_BACKEND}" STREQUAL "lightning_tensor")
-    add_subdirectory(simulators)
-    add_subdirectory(utils)
-else()
-    foreach(COMP ${COMPONENT_SUBDIRS})
-        add_subdirectory(${COMP})
-    endforeach()
-endif()
+foreach(COMP ${COMPONENT_SUBDIRS})
+    add_subdirectory(${COMP})
+endforeach()
+
 
 if (BUILD_TESTS)
     # Include macros supporting tests.

pennylane_lightning/core/src/algorithms/CMakeLists.txt

@@ -15,7 +15,7 @@ target_link_libraries(lightning_algorithms INTERFACE    lightning_compile_option
                                                         lightning_observables
                                                         )
 
-if (BUILD_TESTS)
+if (BUILD_TESTS AND NOT DEFINED PL_TENSOR)
     enable_testing()
     add_subdirectory("tests")
-endif()
+endif()

pennylane_lightning/core/src/measurements/CMakeLists.txt

@@ -11,7 +11,7 @@ target_link_libraries(lightning_measurements INTERFACE  lightning_compile_option
                                                         lightning_utils
                                                         )
 
-if (BUILD_TESTS)
+if (BUILD_TESTS AND NOT DEFINED PL_TENSOR)
     enable_testing()
     add_subdirectory("tests")
-endif()
+endif()

pennylane_lightning/core/src/observables/CMakeLists.txt

@@ -11,7 +11,7 @@ target_link_libraries(lightning_observables INTERFACE   lightning_compile_option
                                                         )
 set_property(TARGET lightning_observables PROPERTY POSITION_INDEPENDENT_CODE ON)
 
-if (BUILD_TESTS)
+if (BUILD_TESTS AND NOT DEFINED PL_TENSOR)
     enable_testing()
     add_subdirectory("tests")
-endif()
+endif()

pennylane_lightning/core/src/simulators/base/CMakeLists.txt

@@ -11,7 +11,7 @@ target_link_libraries(lightning_base INTERFACE  lightning_compile_options
 
 set_property(TARGET lightning_base PROPERTY POSITION_INDEPENDENT_CODE ON)
 
-if (BUILD_TESTS)
+if (BUILD_TESTS AND NOT DEFINED PL_TENSOR)
     enable_testing()
     add_subdirectory("tests")
-endif()
+endif()

pennylane_lightning/core/src/simulators/lightning_tensor/tensor/CMakeLists.txt

@@ -11,7 +11,7 @@ if("${PL_TENSOR_BACKEND}" STREQUAL "cutensornet" AND NOT TNCUDA_UTILS_ADDED)
     add_subdirectory(tncuda)
     target_include_directories(${PL_BACKEND}_tensor INTERFACE ${CMAKE_CURRENT_SOURCE_DIR}/tncuda)
     target_link_libraries(${PL_BACKEND}_tensor INTERFACE tensor_cuda)
-    set(CUDA_UTILS_ADDED TRUE)
+    set(TNCUDA_UTILS_ADDED TRUE)
 endif()
 
 target_include_directories(${PL_BACKEND}_tensor INTERFACE ${CMAKE_CURRENT_SOURCE_DIR}/base)

pennylane_lightning/core/src/simulators/lightning_tensor/tensor/base/TensorBase.hpp

@@ -19,6 +19,7 @@
 
 #pragma once
 
+#include <functional>
 #include <vector>
 
 #include "Error.hpp"
@@ -28,7 +29,7 @@ namespace Pennylane::LightningTensor {
 /**
  * @brief CRTP-enabled base class for tensor.
  *
- * @tparam Precision Floating point precision.
+ * @tparam PrecisionT Floating point precision.
  * @tparam Derived Derived class to instantiate using CRTP.
  */
 template <class PrecisionT, class Derived> class TensorBase {
@@ -44,7 +45,6 @@ template <class PrecisionT, class Derived> class TensorBase {
         : rank_(rank), modes_(modes), extents_(extents) {
         PL_ABORT_IF_NOT(rank_ == extents_.size(),
                         "Please check if rank or extents are set correctly.");
-        length_ = 1;
         length_ = std::accumulate(extents.begin(), extents.end(),
                                   std::size_t{1}, std::multiplies<>());
     }
@@ -79,7 +79,7 @@ template <class PrecisionT, class Derived> class TensorBase {
     /**
      * @brief Return the number of elements of a tensor object.
      *
-     * @return std::vector<std::size_t> Number of elements of a tensor object.
+     * @return std::size_t Number of elements of a tensor object.
      */
     [[nodiscard]] std::size_t getLength() const { return length_; }
 };

pennylane_lightning/core/src/simulators/lightning_tensor/tensor/tncuda/TensorCuda.hpp

@@ -27,17 +27,19 @@
 #include "TensorBase.hpp"
 #include "cuda_helpers.hpp"
 
+/// @cond DEV
 namespace {
 namespace cuUtil = Pennylane::LightningGPU::Util;
 using namespace Pennylane::LightningGPU;
 } // namespace
+/// @endcond
 
 namespace Pennylane::LightningTensor::TNCuda {
 
 /**
  * @brief CRTP-enabled class for CUDA-capable Tensor.
  *
- * @tparam Precision Floating point precision.
+ * @tparam PrecisionT Floating point precision.
  */
 
 template <class PrecisionT>
@@ -61,14 +63,17 @@ class TensorCuda final : public TensorBase<PrecisionT, TensorCuda<PrecisionT>> {
     /**
      * @brief Explicitly copy data from GPU device to host memory.
      *
-     * @param sv Complex data pointer to receive data from device.
+     * @param host_tensor Complex data pointer to receive data from device.
+     * @param length Number of elements to copy.
+     * @param async If true, the copy is asynchronous. Only synchronous copy is
+     * supported now.
      */
-    inline void CopyGpuDataToHost(std::complex<PrecisionT> *host_sv,
+    inline void CopyGpuDataToHost(std::complex<PrecisionT> *host_tensor,
                                   std::size_t length,
                                   bool async = false) const {
         PL_ABORT_IF_NOT(BaseType::getLength() == length,
                         "Sizes do not match for Host and GPU data");
-        data_buffer_->CopyGpuDataToHost(host_sv, length, async);
+        data_buffer_->CopyGpuDataToHost(host_tensor, length, async);
     }
 
     DataBuffer<CFP_t> &getDataBuffer() { return *data_buffer_; }

pennylane_lightning/core/src/simulators/lightning_tensor/tncuda/CMakeLists.txt

@@ -37,9 +37,10 @@ option(PL_DISABLE_CUDA_SAFETY "Build without CUDA call safety checks" OFF)
 
 target_link_libraries(${PL_TENSOR} PUBLIC  lightning_compile_options
                                             lightning_external_libs
-                                            ${PL_BACKEND}_utils
+                                            ${PL_BACKEND}_gates
                                             ${PL_BACKEND}_tensor
                                             ${PL_BACKEND}_tensornetBase
+                                            ${PL_BACKEND}_utils
                                             )
 
 target_include_directories(${PL_TENSOR} PUBLIC ${CMAKE_CURRENT_SOURCE_DIR})
@@ -57,6 +58,7 @@ endif()
 # Include all nested sources directories
 ###############################################################################
 set(COMPONENT_SUBDIRS      base
+                           gates
                            utils
 )
 foreach(COMP ${COMPONENT_SUBDIRS})