diff --git a/pennylane_lightning/core/src/algorithms/tests/mpi/Test_AdjointJacobianMPI.cpp b/pennylane_lightning/core/src/algorithms/tests/mpi/Test_AdjointJacobianMPI.cpp
index 51888c1b37..117d1dc980 100644
--- a/pennylane_lightning/core/src/algorithms/tests/mpi/Test_AdjointJacobianMPI.cpp
+++ b/pennylane_lightning/core/src/algorithms/tests/mpi/Test_AdjointJacobianMPI.cpp
@@ -58,6 +58,7 @@ template <typename TypeList> void testAdjointJacobian() {
         using ComplexT = typename StateVectorT::ComplexT;
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         const std::vector<PrecisionT> param{-M_PI / 7, M_PI / 5, 2 * M_PI / 3};
 
@@ -81,6 +82,7 @@ template <typename TypeList> void testAdjointJacobian() {
 
             int nDevices = 0; // Number of GPU devices per node
             cudaGetDeviceCount(&nDevices);
+            CHECK(nDevices >= 2);
             int deviceId = mpi_manager.getRank() % nDevices;
             cudaSetDevice(deviceId);
             DevTag<int> dt_local(deviceId, 0);
@@ -131,6 +133,7 @@ template <typename TypeList> void testAdjointJacobian() {
 
             int nDevices = 0; // Number of GPU devices per node
             cudaGetDeviceCount(&nDevices);
+            CHECK(nDevices >= 2);
             int deviceId = mpi_manager.getRank() % nDevices;
             cudaSetDevice(deviceId);
             DevTag<int> dt_local(deviceId, 0);
@@ -181,6 +184,7 @@ template <typename TypeList> void testAdjointJacobian() {
 
             int nDevices = 0; // Number of GPU devices per node
             cudaGetDeviceCount(&nDevices);
+            CHECK(nDevices >= 2);
             int deviceId = mpi_manager.getRank() % nDevices;
             cudaSetDevice(deviceId);
             DevTag<int> dt_local(deviceId, 0);
@@ -293,6 +297,7 @@ template <typename TypeList> void testAdjointJacobian() {
 
             int nDevices = 0; // Number of GPU devices per node
             cudaGetDeviceCount(&nDevices);
+            CHECK(nDevices >= 2);
             int deviceId = mpi_manager.getRank() % nDevices;
             cudaSetDevice(deviceId);
             DevTag<int> dt_local(deviceId, 0);
@@ -344,6 +349,7 @@ template <typename TypeList> void testAdjointJacobian() {
 
             int nDevices = 0; // Number of GPU devices per node
             cudaGetDeviceCount(&nDevices);
+            CHECK(nDevices >= 2);
             int deviceId = mpi_manager.getRank() % nDevices;
             cudaSetDevice(deviceId);
             DevTag<int> dt_local(deviceId, 0);
diff --git a/pennylane_lightning/core/src/measurements/tests/mpi/Test_MeasurementsBaseMPI.cpp b/pennylane_lightning/core/src/measurements/tests/mpi/Test_MeasurementsBaseMPI.cpp
index 54a5d299ed..733b8a5743 100644
--- a/pennylane_lightning/core/src/measurements/tests/mpi/Test_MeasurementsBaseMPI.cpp
+++ b/pennylane_lightning/core/src/measurements/tests/mpi/Test_MeasurementsBaseMPI.cpp
@@ -68,6 +68,7 @@ template <typename TypeList> void testProbabilities() {
         size_t num_qubits = 3;
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -77,6 +78,7 @@ template <typename TypeList> void testProbabilities() {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -127,6 +129,7 @@ template <typename TypeList> void testNamedObsExpval() {
         size_t num_qubits = 3;
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -136,6 +139,7 @@ template <typename TypeList> void testNamedObsExpval() {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -197,6 +201,7 @@ template <typename TypeList> void testHermitianObsExpval() {
         size_t num_qubits = 3;
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -206,6 +211,7 @@ template <typename TypeList> void testHermitianObsExpval() {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -294,6 +300,7 @@ template <typename TypeList> void testNamedObsVar() {
         size_t num_qubits = 3;
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -303,6 +310,7 @@ template <typename TypeList> void testNamedObsVar() {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -365,6 +373,7 @@ template <typename TypeList> void testHermitianObsVar() {
         size_t num_qubits = 3;
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -374,6 +383,7 @@ template <typename TypeList> void testHermitianObsVar() {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -470,6 +480,7 @@ template <typename TypeList> void testSamples() {
         size_t num_qubits = 3;
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -479,6 +490,7 @@ template <typename TypeList> void testSamples() {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
diff --git a/pennylane_lightning/core/src/observables/tests/mpi/Test_ObservablesMPI.cpp b/pennylane_lightning/core/src/observables/tests/mpi/Test_ObservablesMPI.cpp
index dee6e37532..20f8466e09 100644
--- a/pennylane_lightning/core/src/observables/tests/mpi/Test_ObservablesMPI.cpp
+++ b/pennylane_lightning/core/src/observables/tests/mpi/Test_ObservablesMPI.cpp
@@ -422,20 +422,6 @@ template <typename TypeList> void testHamiltonianBase() {
 
                 REQUIRE(ham1->getWires() == std::vector<size_t>{0, 5, 9});
             }
-
-            /*
-            DYNAMIC_SECTION("applyInPlace must fail - "
-                            << StateVectorMPIToName<StateVectorT>::name) {
-                auto ham =
-                    HamiltonianT::create({PrecisionT{1.0}, h, h}, {zz, x1, x2});
-                auto st_data = createZeroState<ComplexT>(2);
-
-                StateVectorT state_vector(st_data.data(), st_data.size());
-
-                REQUIRE_THROWS_AS(ham->applyInPlace(state_vector),
-                                  LightningException);
-            }
-            */
         }
         testHamiltonianBase<typename TypeList::Next>();
     }
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/StateVectorCudaMPI.hpp b/pennylane_lightning/core/src/simulators/lightning_gpu/StateVectorCudaMPI.hpp
index 8503efe570..32981cd48b 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/StateVectorCudaMPI.hpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/StateVectorCudaMPI.hpp
@@ -1867,6 +1867,7 @@ class StateVectorCudaMPI
         // LCOV_EXCL_STOP
 
         cuDoubleComplex expect_;
+
         // compute expectation
         PL_CUSTATEVEC_IS_SUCCESS(custatevecComputeExpectation(
             /* custatevecHandle_t */ handle_.get(),
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/algorithms/tests/mpi/Test_AdjointJacobianGPUMPI.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/algorithms/tests/mpi/Test_AdjointJacobianGPUMPI.cpp
index b9c4eba70c..26a13bd203 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/algorithms/tests/mpi/Test_AdjointJacobianGPUMPI.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/algorithms/tests/mpi/Test_AdjointJacobianGPUMPI.cpp
@@ -47,6 +47,7 @@ TEST_CASE("AdjointJacobianGPUMPI::adjointJacobianMPI Op=RX, Obs=[Z,Z]",
     using StateVectorT = StateVectorCudaMPI<double>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     AdjointJacobianMPI<StateVectorT> adj;
     std::vector<double> param{-M_PI / 7, M_PI / 5, 2 * M_PI / 3};
@@ -66,6 +67,7 @@ TEST_CASE("AdjointJacobianGPUMPI::adjointJacobianMPI Op=RX, Obs=[Z,Z]",
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -104,6 +106,7 @@ TEST_CASE("AdjointJacobianGPUMPI::adjointJacobianMPI Op=[QubitStateVector, "
     using StateVectorT = StateVectorCudaMPI<double>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     AdjointJacobianMPI<StateVectorT> adj;
     std::vector<double> param{-M_PI / 7, M_PI / 5, 2 * M_PI / 3};
@@ -124,6 +127,7 @@ TEST_CASE("AdjointJacobianGPUMPI::adjointJacobianMPI Op=[QubitStateVector, "
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -167,6 +171,7 @@ TEST_CASE(
     std::vector<double> jacobian_serial(num_obs * tp.size(), 0);
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -177,6 +182,7 @@ TEST_CASE(
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -234,6 +240,7 @@ TEST_CASE(
     std::vector<double> jacobian_serial(num_obs * tp.size(), 0);
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -244,6 +251,7 @@ TEST_CASE(
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -297,6 +305,7 @@ TEST_CASE("AdjointJacobianGPUMPI::adjointJacobian Op=[RX,RX,RX], Obs=[ZZZ]",
     std::vector<double> jacobian_serial(num_obs * tp.size(), 0);
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -307,6 +316,7 @@ TEST_CASE("AdjointJacobianGPUMPI::adjointJacobian Op=[RX,RX,RX], Obs=[ZZZ]",
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -358,6 +368,7 @@ TEST_CASE("AdjointJacobianGPUMPI::adjointJacobian Op=Mixed, Obs=[XXX]",
     std::vector<double> jacobian_serial(num_obs * tp.size(), 0);
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -368,6 +379,7 @@ TEST_CASE("AdjointJacobianGPUMPI::adjointJacobian Op=Mixed, Obs=[XXX]",
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -436,6 +448,7 @@ TEST_CASE("AdjointJacobianGPU::AdjointJacobianGPUMPI Op=[RX,RX,RX], "
     std::vector<double> jacobian_serial(num_obs * tp.size(), 0);
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -446,6 +459,7 @@ TEST_CASE("AdjointJacobianGPU::AdjointJacobianGPUMPI Op=[RX,RX,RX], "
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -504,6 +518,7 @@ TEST_CASE("AdjointJacobianGPU::AdjointJacobianGPU Test HermitianObs",
     std::vector<double> jacobian2_serial(num_obs * tp.size(), 0);
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -514,6 +529,7 @@ TEST_CASE("AdjointJacobianGPU::AdjointJacobianGPU Test HermitianObs",
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_Generators.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_Generators.cpp
index 6dbd1f70f6..511e181d1b 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_Generators.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_Generators.cpp
@@ -64,6 +64,7 @@ using namespace Pennylane::LightningGPU::MPI;
         using cp_t = std::complex<TestType>;                                   \
         using PrecisionT = TestType;                                           \
         MPIManager mpi_manager(MPI_COMM_WORLD);                                \
+        CHECK(mpi_manager.getSize() == 2);                                     \
         size_t mpi_buffersize = 1;                                             \
         size_t nGlobalIndexBits =                                              \
             std::bit_width(static_cast<size_t>(mpi_manager.getSize())) - 1;    \
@@ -82,6 +83,7 @@ using namespace Pennylane::LightningGPU::MPI;
         mpi_manager.Barrier();                                                 \
         int nDevices = 0;                                                      \
         cudaGetDeviceCount(&nDevices);                                         \
+        CHECK(nDevices >= 2);                                                  \
         int deviceId = mpi_manager.getRank() % nDevices;                       \
         cudaSetDevice(deviceId);                                               \
         DevTag<int> dt_local(deviceId, 0);                                     \
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_NonParam.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_NonParam.cpp
index b8985e6642..fd8882769a 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_NonParam.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_NonParam.cpp
@@ -90,6 +90,7 @@ TEMPLATE_TEST_CASE("StateVectorCudaMPI::SetStateVector",
     using PrecisionT = TestType;
     using cp_t = std::complex<PrecisionT>;
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -132,6 +133,7 @@ TEMPLATE_TEST_CASE("StateVectorCudaMPI::SetStateVector",
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -165,6 +167,7 @@ TEMPLATE_TEST_CASE("StateVectorCudaMPI::SetIthStates",
     using PrecisionT = TestType;
     using cp_t = std::complex<PrecisionT>;
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -193,6 +196,7 @@ TEMPLATE_TEST_CASE("StateVectorCudaMPI::SetIthStates",
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -219,6 +223,7 @@ TEMPLATE_TEST_CASE("StateVectorCudaMPI::SetIthStates",
         using cp_t = std::complex<TestType>;                                   \
         using PrecisionT = TestType;                                           \
         MPIManager mpi_manager(MPI_COMM_WORLD);                                \
+        CHECK(mpi_manager.getSize() == 2);                                     \
         size_t mpi_buffersize = 1;                                             \
         size_t nGlobalIndexBits =                                              \
             std::bit_width(static_cast<size_t>(mpi_manager.getSize())) - 1;    \
@@ -237,6 +242,7 @@ TEMPLATE_TEST_CASE("StateVectorCudaMPI::SetIthStates",
         mpi_manager.Barrier();                                                 \
         int nDevices = 0;                                                      \
         cudaGetDeviceCount(&nDevices);                                         \
+        CHECK(nDevices >= 2);                                                  \
         int deviceId = mpi_manager.getRank() % nDevices;                       \
         cudaSetDevice(deviceId);                                               \
         DevTag<int> dt_local(deviceId, 0);                                     \
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_Param.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_Param.cpp
index d30115d4a9..4f00345d8f 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_Param.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/gates/tests/mpi/Test_StateVectorCudaMPI_Param.cpp
@@ -63,6 +63,7 @@ using namespace Pennylane::LightningGPU::MPI;
         using cp_t = std::complex<TestType>;                                   \
         using PrecisionT = TestType;                                           \
         MPIManager mpi_manager(MPI_COMM_WORLD);                                \
+        CHECK(mpi_manager.getSize() == 2);                                     \
         size_t mpi_buffersize = 1;                                             \
         size_t nGlobalIndexBits =                                              \
             std::bit_width(static_cast<size_t>(mpi_manager.getSize())) - 1;    \
@@ -81,6 +82,7 @@ using namespace Pennylane::LightningGPU::MPI;
         mpi_manager.Barrier();                                                 \
         int nDevices = 0;                                                      \
         cudaGetDeviceCount(&nDevices);                                         \
+        CHECK(nDevices >= 2);                                                  \
         int deviceId = mpi_manager.getRank() % nDevices;                       \
         cudaSetDevice(deviceId);                                               \
         DevTag<int> dt_local(deviceId, 0);                                     \
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Expval.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Expval.cpp
index 1a1a62c924..582eec1e2d 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Expval.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Expval.cpp
@@ -53,6 +53,7 @@ TEMPLATE_TEST_CASE("[Identity]", "[StateVectorCudaMPI_Expval]", float, double) {
     auto ONE = TestType(1);
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -62,6 +63,7 @@ TEMPLATE_TEST_CASE("[Identity]", "[StateVectorCudaMPI_Expval]", float, double) {
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -91,6 +93,7 @@ TEMPLATE_TEST_CASE("[PauliX]", "[StateVectorCudaMPI_Expval]", float, double) {
         auto ONE = TestType(1);
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -100,6 +103,7 @@ TEMPLATE_TEST_CASE("[PauliX]", "[StateVectorCudaMPI_Expval]", float, double) {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -162,6 +166,7 @@ TEMPLATE_TEST_CASE("[PauliY]", "[StateVectorCudaMPI_Expval]", float, double) {
         auto PI = TestType(M_PI);
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -171,6 +176,7 @@ TEMPLATE_TEST_CASE("[PauliY]", "[StateVectorCudaMPI_Expval]", float, double) {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -228,6 +234,7 @@ TEMPLATE_TEST_CASE("[PauliZ]", "[StateVectorCudaMPI_Expval]", float, double) {
         size_t num_qubits = 3;
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -237,6 +244,7 @@ TEMPLATE_TEST_CASE("[PauliZ]", "[StateVectorCudaMPI_Expval]", float, double) {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -266,6 +274,7 @@ TEMPLATE_TEST_CASE("[Hadamard]", "[StateVectorCudaMPI_Expval]", float, double) {
         auto INVSQRT2 = TestType(0.707106781186547524401);
 
         MPIManager mpi_manager(MPI_COMM_WORLD);
+        CHECK(mpi_manager.getSize() == 2);
 
         size_t mpi_buffersize = 1;
 
@@ -275,6 +284,7 @@ TEMPLATE_TEST_CASE("[Hadamard]", "[StateVectorCudaMPI_Expval]", float, double) {
 
         int nDevices = 0;
         cudaGetDeviceCount(&nDevices);
+        CHECK(nDevices >= 2);
         int deviceId = mpi_manager.getRank() % nDevices;
         cudaSetDevice(deviceId);
         DevTag<int> dt_local(deviceId, 0);
@@ -299,6 +309,7 @@ TEMPLATE_TEST_CASE("Test expectation value of HamiltonianObs",
     using ComplexT = StateVectorT::ComplexT;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t num_qubits = 3;
     size_t mpi_buffersize = 1;
@@ -309,6 +320,7 @@ TEMPLATE_TEST_CASE("Test expectation value of HamiltonianObs",
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -345,6 +357,7 @@ TEMPLATE_TEST_CASE("Test expectation value of TensorProdObs",
     using ComplexT = StateVectorT::ComplexT;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t num_qubits = 3;
     size_t mpi_buffersize = 1;
@@ -355,6 +368,7 @@ TEMPLATE_TEST_CASE("Test expectation value of TensorProdObs",
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -384,11 +398,12 @@ TEMPLATE_TEST_CASE("Test expectation value of TensorProdObs",
 }
 
 TEMPLATE_TEST_CASE("StateVectorCudaMPI::Hamiltonian_expval_Sparse",
-                   "[StateVectorCudaMPI_Expval]", double) {
+                   "[StateVectorCudaMPI_Expval]", float, double) {
     using StateVectorT = StateVectorCudaMPI<TestType>;
     using ComplexT = StateVectorT::ComplexT;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t num_qubits = 3;
     size_t mpi_buffersize = 1;
@@ -399,6 +414,7 @@ TEMPLATE_TEST_CASE("StateVectorCudaMPI::Hamiltonian_expval_Sparse",
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Measure.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Measure.cpp
index ca19aae90e..5ca6b7684b 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Measure.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Measure.cpp
@@ -37,7 +37,7 @@ using Pennylane::Util::createNonTrivialState;
 }; // namespace
 /// @endcond
 
-TEMPLATE_TEST_CASE("Expected Values", "[MeasurementsMPI]", double) {
+TEMPLATE_TEST_CASE("Expected Values", "[MeasurementsMPI]", float, double) {
     using StateVectorT = StateVectorCudaMPI<TestType>;
     using PrecisionT = typename StateVectorT::PrecisionT;
     using ComplexT = typename StateVectorT::ComplexT;
@@ -49,6 +49,7 @@ TEMPLATE_TEST_CASE("Expected Values", "[MeasurementsMPI]", double) {
     size_t num_qubits = 3;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -58,6 +59,7 @@ TEMPLATE_TEST_CASE("Expected Values", "[MeasurementsMPI]", double) {
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -190,12 +192,15 @@ TEMPLATE_TEST_CASE("Expected Values", "[MeasurementsMPI]", double) {
     }
 }
 
-TEMPLATE_TEST_CASE("Pauliwords base on expval", "[MeasurementsMPI]", double) {
+TEMPLATE_TEST_CASE("Pauliwords base on expval", "[MeasurementsMPI]", float,
+                   double) {
     using PrecisionT = TestType;
     using cp_t = std::complex<PrecisionT>;
     using StateVectorT = StateVectorCudaMPI<TestType>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
+
     size_t numqubits = 4;
     size_t mpi_buffersize = 1;
 
@@ -225,6 +230,7 @@ TEMPLATE_TEST_CASE("Pauliwords base on expval", "[MeasurementsMPI]", double) {
 
     int nDevices = 0; // Number of GPU devices per node
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -294,7 +300,7 @@ TEMPLATE_TEST_CASE("Pauliwords base on expval", "[MeasurementsMPI]", double) {
     }
 }
 
-TEMPLATE_TEST_CASE("Variances", "[MeasurementsMPI]", double) {
+TEMPLATE_TEST_CASE("Variances", "[MeasurementsMPI]", float, double) {
     using StateVectorT = StateVectorCudaMPI<TestType>;
     using PrecisionT = typename StateVectorT::PrecisionT;
     using ComplexT = typename StateVectorT::ComplexT;
@@ -306,6 +312,7 @@ TEMPLATE_TEST_CASE("Variances", "[MeasurementsMPI]", double) {
     size_t num_qubits = 3;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -315,6 +322,7 @@ TEMPLATE_TEST_CASE("Variances", "[MeasurementsMPI]", double) {
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -404,7 +412,7 @@ TEMPLATE_TEST_CASE("Variances", "[MeasurementsMPI]", double) {
     }
 }
 
-TEMPLATE_TEST_CASE("Probabilities", "[MeasuresMPI]", double) {
+TEMPLATE_TEST_CASE("Probabilities", "[MeasuresMPI]", float, double) {
     using StateVectorT = StateVectorCudaMPI<TestType>;
     // Probabilities calculated with Pennylane default.qubit:
     std::vector<std::pair<std::vector<size_t>, std::vector<TestType>>> input = {
@@ -420,6 +428,7 @@ TEMPLATE_TEST_CASE("Probabilities", "[MeasuresMPI]", double) {
     size_t num_qubits = 3;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -429,6 +438,7 @@ TEMPLATE_TEST_CASE("Probabilities", "[MeasuresMPI]", double) {
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Var.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Var.cpp
index 141b987cb0..caea1e0d03 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Var.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/measurements/tests/mpi/Test_StateVectorCudaMPI_Var.cpp
@@ -41,6 +41,7 @@ TEMPLATE_TEST_CASE("Test variance of NamedObs", "[StateVectorCudaMPI_Var]",
     const std::size_t num_qubits = 2;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -50,6 +51,7 @@ TEMPLATE_TEST_CASE("Test variance of NamedObs", "[StateVectorCudaMPI_Var]",
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -108,12 +110,13 @@ TEMPLATE_TEST_CASE("Test variance of NamedObs", "[StateVectorCudaMPI_Var]",
 }
 
 TEMPLATE_TEST_CASE("Test variance of HermitianObs", "[StateVectorCudaMPI_Var]",
-                   double) {
+                   float, double) {
     const std::size_t num_qubits = 3;
     using StateVectorT = StateVectorCudaMPI<TestType>;
     using ComplexT = typename StateVectorT::ComplexT;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -123,6 +126,7 @@ TEMPLATE_TEST_CASE("Test variance of HermitianObs", "[StateVectorCudaMPI_Var]",
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -160,11 +164,12 @@ TEMPLATE_TEST_CASE("Test variance of HermitianObs", "[StateVectorCudaMPI_Var]",
 }
 
 TEMPLATE_TEST_CASE("Test variance of TensorProdObs", "[StateVectorCudaMPI_Var]",
-                   double) {
+                   float, double) {
     using StateVectorT = StateVectorCudaMPI<TestType>;
     const std::size_t num_qubits = 3;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -174,6 +179,7 @@ TEMPLATE_TEST_CASE("Test variance of TensorProdObs", "[StateVectorCudaMPI_Var]",
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
@@ -203,11 +209,12 @@ TEMPLATE_TEST_CASE("Test variance of TensorProdObs", "[StateVectorCudaMPI_Var]",
 }
 
 TEMPLATE_TEST_CASE("Test variance of HamiltonianObs",
-                   "[StateVectorCudaMPI_Var]", double) {
+                   "[StateVectorCudaMPI_Var]", float, double) {
     using StateVectorT = StateVectorCudaMPI<TestType>;
     const std::size_t num_qubits = 3;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     size_t mpi_buffersize = 1;
 
@@ -217,6 +224,7 @@ TEMPLATE_TEST_CASE("Test variance of HamiltonianObs",
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/observables/tests/mpi/Test_ObservablesGPUMPI.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/observables/tests/mpi/Test_ObservablesGPUMPI.cpp
index e757fb4b84..a39f9f5f7e 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/observables/tests/mpi/Test_ObservablesGPUMPI.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/observables/tests/mpi/Test_ObservablesGPUMPI.cpp
@@ -172,6 +172,8 @@ TEMPLATE_PRODUCT_TEST_CASE("HamiltonianMPI::ApplyInPlace", "[Observables]",
     using Hamiltonian = Hamiltonian<StateVectorCudaManaged<PrecisionT>>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
+
     size_t num_qubits = 8;
     size_t mpi_buffersize = 1;
     size_t nGlobalIndexBits =
@@ -191,6 +193,7 @@ TEMPLATE_PRODUCT_TEST_CASE("HamiltonianMPI::ApplyInPlace", "[Observables]",
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_CSRMatrix.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_CSRMatrix.cpp
index a7be042ce3..3cc508fb98 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_CSRMatrix.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_CSRMatrix.cpp
@@ -42,6 +42,7 @@ TEMPLATE_TEST_CASE("CRSMatrix::Split", "[CRSMatrix]", float, double) {
                                   int64_t>::type;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     int rank = mpi_manager.getRank();
     int size = mpi_manager.getSize();
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_LinearAlgebraMPI.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_LinearAlgebraMPI.cpp
index 4a3892b4fd..43abf11c62 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_LinearAlgebraMPI.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_LinearAlgebraMPI.cpp
@@ -42,6 +42,7 @@ TEMPLATE_TEST_CASE("Linear Algebra::SparseMV", "[Linear Algebra]", float,
     using CFP_t = StateVectorT::CFP_t;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     std::size_t num_qubits = 3;
 
@@ -79,6 +80,7 @@ TEMPLATE_TEST_CASE("Linear Algebra::SparseMV", "[Linear Algebra]", float,
 
     int nDevices = 0;
     cudaGetDeviceCount(&nDevices);
+    CHECK(nDevices >= 2);
     int deviceId = mpi_manager.getRank() % nDevices;
     cudaSetDevice(deviceId);
     DevTag<int> dt_local(deviceId, 0);
diff --git a/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_MPIManager.cpp b/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_MPIManager.cpp
index 549be55b85..da21c9028e 100644
--- a/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_MPIManager.cpp
+++ b/pennylane_lightning/core/src/simulators/lightning_gpu/utils/tests/mpi/Test_MPIManager.cpp
@@ -36,6 +36,7 @@ TEMPLATE_TEST_CASE("MPIManager::Scatter", "[MPIManager]", float, double) {
     using cp_t = std::complex<PrecisionT>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
 
     int rank = mpi_manager.getRank();
     int size = mpi_manager.getSize();
@@ -79,6 +80,8 @@ TEMPLATE_TEST_CASE("MPIManager::Allgather", "[MPIManager]", float, double) {
     using cp_t = std::complex<PrecisionT>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
+
     int rank = mpi_manager.getRank();
     int size = mpi_manager.getSize();
 
@@ -132,6 +135,8 @@ TEMPLATE_TEST_CASE("MPIManager::Reduce", "[MPIManager]", float, double) {
     using cp_t = std::complex<PrecisionT>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
+
     int rank = mpi_manager.getRank();
     int size = mpi_manager.getSize();
 
@@ -170,6 +175,8 @@ TEMPLATE_TEST_CASE("MPIManager::Allreduce", "[MPIManager]", float, double) {
     using cp_t = std::complex<PrecisionT>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
+
     int rank = mpi_manager.getRank();
     int size = mpi_manager.getSize();
 
@@ -216,6 +223,8 @@ TEMPLATE_TEST_CASE("MPIManager::Bcast", "[MPIManager]", float, double) {
     using cp_t = std::complex<PrecisionT>;
 
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
+
     int rank = mpi_manager.getRank();
 
     SECTION("Apply Bcast scalar") {
@@ -265,6 +274,7 @@ TEMPLATE_TEST_CASE("MPIManager::Sendrecv", "[MPIManager]", float, double) {
 
 TEST_CASE("MPIManager::split") {
     MPIManager mpi_manager(MPI_COMM_WORLD);
+    CHECK(mpi_manager.getSize() == 2);
     int rank = mpi_manager.getRank();
     int color = rank % 2;
     int key = rank;