From 6a8cce0a3bc7dd5a2a7cd22f5aa7377afb6761d5 Mon Sep 17 00:00:00 2001
From: Shuli Shu <08cnbj@gmail.com>
Date: Wed, 23 Aug 2023 19:11:08 -0700
Subject: [PATCH] add docstring

---
 .../src/algorithms/ObservablesGPUMPI.hpp              |  5 ++---
 .../src/simulator/StateVectorCudaMPI.hpp              |  1 -
 pennylane_lightning_gpu/src/util/CSRMatrix.hpp        | 11 +++++++----
 3 files changed, 9 insertions(+), 8 deletions(-)

diff --git a/pennylane_lightning_gpu/src/algorithms/ObservablesGPUMPI.hpp b/pennylane_lightning_gpu/src/algorithms/ObservablesGPUMPI.hpp
index d99431ba..fdaefc15 100644
--- a/pennylane_lightning_gpu/src/algorithms/ObservablesGPUMPI.hpp
+++ b/pennylane_lightning_gpu/src/algorithms/ObservablesGPUMPI.hpp
@@ -364,9 +364,8 @@ class HamiltonianGPUMPI final : public ObservableGPUMPI<T> {
         for (size_t term_idx = 0; term_idx < coeffs_.size(); term_idx++) {
             DevTag<int> dt_local(sv.getDataBuffer().getDevTag());
             dt_local.refresh();
-            StateVectorCudaMPI<PrecisionT> tmp(
-                dt_local, sv.getNumGlobalQubits(), sv.getNumLocalQubits(),
-                sv.getData());
+            StateVectorCudaMPI<T> tmp(dt_local, sv.getNumGlobalQubits(),
+                                      sv.getNumLocalQubits(), sv.getData());
             obs_[term_idx]->applyInPlace(tmp);
             scaleAndAddC_CUDA(std::complex<T>{coeffs_[term_idx], 0.0},
                               tmp.getData(), buffer.getData(), tmp.getLength(),
diff --git a/pennylane_lightning_gpu/src/simulator/StateVectorCudaMPI.hpp b/pennylane_lightning_gpu/src/simulator/StateVectorCudaMPI.hpp
index 4de7efe1..d8efa764 100644
--- a/pennylane_lightning_gpu/src/simulator/StateVectorCudaMPI.hpp
+++ b/pennylane_lightning_gpu/src/simulator/StateVectorCudaMPI.hpp
@@ -959,7 +959,6 @@ class StateVectorCudaMPI
         const index_type *csrOffsets_ptr, const index_type csrOffsets_size,
         const index_type *columns_ptr,
         const std::complex<Precision> *values_ptr, const index_type numNNZ) {
-
         if (mpi_manager_.getRank() == 0) {
             PL_ABORT_IF_NOT(static_cast<size_t>(csrOffsets_size - 1) ==
                                 (size_t{1} << this->getTotalNumQubits()),
diff --git a/pennylane_lightning_gpu/src/util/CSRMatrix.hpp b/pennylane_lightning_gpu/src/util/CSRMatrix.hpp
index 59856d46..78a60ddf 100644
--- a/pennylane_lightning_gpu/src/util/CSRMatrix.hpp
+++ b/pennylane_lightning_gpu/src/util/CSRMatrix.hpp
@@ -61,9 +61,10 @@ template <class Precision, class index_type> class CSRMatrix {
  * @brief Convert a global CSR (Compressed Sparse Row) format matrix into
  * local blocks. This operation should be conducted on the rank 0.
  *
+ * @tparam Precision Floating-point precision type.
  * @tparam index_type Integer type used as indices of the sparse matrix.
- * @param num_row_blocks Number of local blocks per global row.
- * @param num_col_blocks Number of local blocks per global column.
+ * @param mpi_manager MPIManager object.
+ * @param num_rows Number of rows of the CSR matrix.
  * @param csrOffsets_ptr Pointer to the array of row offsets of the sparse
  * matrix. Array of size csrOffsets_size.
  * @param columns_ptr Pointer to the array of column indices of the sparse
@@ -147,8 +148,11 @@ auto splitCSRMatrix(MPIManager &mpi_manager, const size_t &num_rows,
 /**
  * @brief Scatter a CSR (Compressed Sparse Row) format matrix.
  *
+ * @tparam Precision Floating-point precision type.
  * @tparam index_type Integer type used as indices of the sparse matrix.
- * @param matrix CSR (Compressed Sparse Row) format matrix.
+ * @param mpi_manager MPIManager object.
+ * @param matrix CSR (Compressed Sparse Row) format matrix vector.
+ * @param local_num_rows Number of rows of local CSR matrix.
  * @param root Root rank of the scatter operation.
  */
 template <class Precision, class index_type>
@@ -156,7 +160,6 @@ auto scatterCSRMatrix(MPIManager &mpi_manager,
                       std::vector<CSRMatrix<Precision, index_type>> &matrix,
                       size_t local_num_rows, size_t root)
     -> CSRMatrix<Precision, index_type> {
-    // Bcast num_rows and num_cols
     size_t num_col_blocks = mpi_manager.getSize();
 
     std::vector<size_t> nnzs;