diff --git a/src/M2ulPhyS2Boltzmann.cpp b/src/M2ulPhyS2Boltzmann.cpp
index a8aba27b6..8bdf6a5cc 100644
--- a/src/M2ulPhyS2Boltzmann.cpp
+++ b/src/M2ulPhyS2Boltzmann.cpp
@@ -90,12 +90,12 @@ void M2ulPhyS::fetch(TPS::Tps2Boltzmann &interface) {
   mfem::ParFiniteElementSpace *reaction_rates_fes(&(interface.NativeFes(TPS::Tps2Boltzmann::Index::ReactionRates)));
   externalReactionRates.reset(new mfem::ParGridFunction(reaction_rates_fes));
   interface.interpolateToNativeFES(*externalReactionRates, TPS::Tps2Boltzmann::Index::ReactionRates);
-  #if defined(_CUDA_) || defined(_HIP_)
-    const double * data(externalReactionRates->Read() );
-    int size(externalReactionRates->FESpace()->GetNDofs() );
-    assert(externalReactionRates->FESpace()->GetOrdering() == mfem::Ordering::byNODES);
-    gpu::deviceSetChemistryReactionData<<<1, 1>>>(data, size, chemistry_);
-  #else
-    chemistry_->setGridFunctionRates(*externalReactionRates);
-  #endif
+#if defined(_CUDA_) || defined(_HIP_)
+  const double *data(externalReactionRates->Read());
+  int size(externalReactionRates->FESpace()->GetNDofs());
+  assert(externalReactionRates->FESpace()->GetOrdering() == mfem::Ordering::byNODES);
+  gpu::deviceSetChemistryReactionData<<<1, 1>>>(data, size, chemistry_);
+#else
+  chemistry_->setGridFunctionRates(*externalReactionRates);
+#endif
 }
diff --git a/src/calorically_perfect.cpp b/src/calorically_perfect.cpp
index 539a52a4a..c182c14e6 100644
--- a/src/calorically_perfect.cpp
+++ b/src/calorically_perfect.cpp
@@ -115,10 +115,23 @@ CaloricallyPerfectThermoChem::CaloricallyPerfectThermoChem(mfem::ParMesh *pmesh,
 
   tpsP_->getInput("loMach/openSystem", domain_is_open_, false);
 
-  tpsP_->getInput("loMach/calperfect/linear-solver-rtol", rtol_, 1e-12);
-  tpsP_->getInput("loMach/calperfect/linear-solver-max-iter", max_iter_, 1000);
-  tpsP_->getInput("loMach/calperfect/linear-solver-verbosity", pl_solve_, 0);
+  // NOTE: this should default to FALSE (but would break all related test cases...)
   tpsP_->getInput("loMach/calperfect/numerical-integ", numerical_integ_, true);
+
+  tpsP_->getInput("loMach/calperfect/linear-solver-rtol", rtol_, default_rtol_);
+  tpsP_->getInput("loMach/calperfect/linear-solver-max-iter", max_iter_, default_max_iter_);
+  tpsP_->getInput("loMach/calperfect/linear-solver-verbosity", pl_solve_, 0);
+
+  // not deleting above block to maintain backwards-compatability
+  tpsP_->getInput("loMach/calperfect/hsolve-rtol", hsolve_rtol_, rtol_);
+  tpsP_->getInput("loMach/calperfect/hsolve-atol", hsolve_atol_, default_atol_);
+  tpsP_->getInput("loMach/calperfect/hsolve-max-iter", hsolve_max_iter_, max_iter_);
+  tpsP_->getInput("loMach/calperfect/hsolve-verbosity", hsolve_pl_, pl_solve_);
+
+  tpsP_->getInput("loMach/calperfect/msolve-rtol", mass_inverse_rtol_, rtol_);
+  tpsP_->getInput("loMach/calperfect/msolve-atol", mass_inverse_atol_, default_atol_);
+  tpsP_->getInput("loMach/calperfect/msolve-max-iter", mass_inverse_max_iter_, max_iter_);
+  tpsP_->getInput("loMach/calperfect/msolve-verbosity", mass_inverse_pl_, pl_solve_);
 }
 
 CaloricallyPerfectThermoChem::~CaloricallyPerfectThermoChem() {
@@ -400,6 +413,9 @@ void CaloricallyPerfectThermoChem::initializeSelf() {
 void CaloricallyPerfectThermoChem::initializeOperators() {
   dt_ = time_coeff_.dt;
 
+  // polynomial order for smoother of precons
+  smoother_poly_order_ = order_ + 1;
+
   Array<int> empty;
 
   // GLL integration rule (Numerical Integration)
@@ -438,9 +454,6 @@ void CaloricallyPerfectThermoChem::initializeOperators() {
     at_blfi->SetIntRule(&ir_nli);
   }
   At_form_->AddDomainIntegrator(at_blfi);
-  if (partial_assembly_) {
-    At_form_->SetAssemblyLevel(AssemblyLevel::PARTIAL);
-  }
   At_form_->Assemble();
   At_form_->FormSystemMatrix(empty, At_);
   if (rank0_) std::cout << "CaloricallyPerfectThermoChem At operator set" << endl;
@@ -466,9 +479,6 @@ void CaloricallyPerfectThermoChem::initializeOperators() {
     // msrho_blfi->SetIntRule(&ir_di);
   }
   MsRho_form_->AddDomainIntegrator(msrho_blfi);
-  if (partial_assembly_) {
-    MsRho_form_->SetAssemblyLevel(AssemblyLevel::PARTIAL);
-  }
   MsRho_form_->Assemble();
   MsRho_form_->FormSystemMatrix(empty, MsRho_);
   if (rank0_) std::cout << "CaloricallyPerfectThermoChem MsRho operator set" << endl;
@@ -483,9 +493,6 @@ void CaloricallyPerfectThermoChem::initializeOperators() {
   }
   Ht_form_->AddDomainIntegrator(hmt_blfi);
   Ht_form_->AddDomainIntegrator(hdt_blfi);
-  if (partial_assembly_) {
-    Ht_form_->SetAssemblyLevel(AssemblyLevel::PARTIAL);
-  }
   Ht_form_->Assemble();
   Ht_form_->FormSystemMatrix(temp_ess_tdof_, Ht_);
   if (rank0_) std::cout << "CaloricallyPerfectThermoChem Ht operator set" << endl;
@@ -496,33 +503,34 @@ void CaloricallyPerfectThermoChem::initializeOperators() {
     MsInvPC_ = new OperatorJacobiSmoother(diag_pa, empty);
   } else {
     MsInvPC_ = new HypreSmoother(*Ms_.As<HypreParMatrix>());
-    dynamic_cast<HypreSmoother *>(MsInvPC_)->SetType(HypreSmoother::Jacobi, 1);
+    dynamic_cast<HypreSmoother *>(MsInvPC_)->SetType(HypreSmoother::Jacobi, smoother_passes_);
+    dynamic_cast<HypreSmoother *>(MsInvPC_)->SetSOROptions(smoother_relax_weight_, smoother_relax_omega_);
+    dynamic_cast<HypreSmoother *>(MsInvPC_)->SetPolyOptions(smoother_poly_order_, smoother_poly_fraction_,
+                                                            smoother_eig_est_);
   }
   MsInv_ = new CGSolver(sfes_->GetComm());
   MsInv_->iterative_mode = false;
   MsInv_->SetOperator(*Ms_);
   MsInv_->SetPreconditioner(*MsInvPC_);
-  MsInv_->SetPrintLevel(pl_solve_);
-  MsInv_->SetRelTol(rtol_);
-  MsInv_->SetMaxIter(max_iter_);
+  MsInv_->SetPrintLevel(mass_inverse_pl_);
+  MsInv_->SetRelTol(mass_inverse_rtol_);
+  MsInv_->SetAbsTol(mass_inverse_atol_);
+  MsInv_->SetMaxIter(mass_inverse_max_iter_);
 
-  if (partial_assembly_) {
-    Vector diag_pa(sfes_->GetTrueVSize());
-    Ht_form_->AssembleDiagonal(diag_pa);
-    HtInvPC_ = new OperatorJacobiSmoother(diag_pa, temp_ess_tdof_);
-  } else {
-    HtInvPC_ = new HypreSmoother(*Ht_.As<HypreParMatrix>());
-    dynamic_cast<HypreSmoother *>(HtInvPC_)->SetType(HypreSmoother::Jacobi, 1);
-  }
+  HtInvPC_ = new HypreSmoother(*Ht_.As<HypreParMatrix>());
+  dynamic_cast<HypreSmoother *>(HtInvPC_)->SetType(HypreSmoother::Jacobi, smoother_passes_);
+  dynamic_cast<HypreSmoother *>(HtInvPC_)->SetSOROptions(hsmoother_relax_weight_, hsmoother_relax_omega_);
+  dynamic_cast<HypreSmoother *>(HtInvPC_)->SetPolyOptions(smoother_poly_order_, smoother_poly_fraction_,
+                                                          smoother_eig_est_);
 
   HtInv_ = new CGSolver(sfes_->GetComm());
-
   HtInv_->iterative_mode = true;
   HtInv_->SetOperator(*Ht_);
   HtInv_->SetPreconditioner(*HtInvPC_);
-  HtInv_->SetPrintLevel(pl_solve_);
-  HtInv_->SetRelTol(rtol_);
-  HtInv_->SetMaxIter(max_iter_);
+  HtInv_->SetPrintLevel(hsolve_pl_);
+  HtInv_->SetRelTol(hsolve_rtol_);
+  HtInv_->SetAbsTol(hsolve_atol_);
+  HtInv_->SetMaxIter(hsolve_max_iter_);
   if (rank0_) std::cout << "Temperature operators set" << endl;
 
   // Qt .....................................
@@ -546,15 +554,19 @@ void CaloricallyPerfectThermoChem::initializeOperators() {
     MqInvPC_ = new OperatorJacobiSmoother(diag_pa, empty);
   } else {
     MqInvPC_ = new HypreSmoother(*Mq_.As<HypreParMatrix>());
-    dynamic_cast<HypreSmoother *>(MqInvPC_)->SetType(HypreSmoother::Jacobi, 1);
+    dynamic_cast<HypreSmoother *>(MqInvPC_)->SetType(HypreSmoother::Jacobi, smoother_passes_);
+    dynamic_cast<HypreSmoother *>(MqInvPC_)->SetSOROptions(smoother_relax_weight_, smoother_relax_omega_);
+    dynamic_cast<HypreSmoother *>(MqInvPC_)->SetPolyOptions(smoother_poly_order_, smoother_poly_fraction_,
+                                                            smoother_eig_est_);
   }
   MqInv_ = new CGSolver(sfes_->GetComm());
   MqInv_->iterative_mode = false;
   MqInv_->SetOperator(*Mq_);
   MqInv_->SetPreconditioner(*MqInvPC_);
-  MqInv_->SetPrintLevel(pl_solve_);
-  MqInv_->SetRelTol(rtol_);
-  MqInv_->SetMaxIter(max_iter_);
+  MqInv_->SetPrintLevel(mass_inverse_pl_);
+  MqInv_->SetRelTol(mass_inverse_rtol_);
+  MqInv_->SetAbsTol(mass_inverse_atol_);
+  MqInv_->SetMaxIter(mass_inverse_max_iter_);
 
   LQ_form_ = new ParBilinearForm(sfes_);
   auto *lqd_blfi = new DiffusionIntegrator(*thermal_diff_total_coeff_);
@@ -664,15 +676,7 @@ void CaloricallyPerfectThermoChem::step() {
   Ht_form_->Update();
   Ht_form_->Assemble();
   Ht_form_->FormSystemMatrix(temp_ess_tdof_, Ht_);
-
   HtInv_->SetOperator(*Ht_);
-  if (partial_assembly_) {
-    delete HtInvPC_;
-    Vector diag_pa(sfes_->GetTrueVSize());
-    Ht_form_->AssembleDiagonal(diag_pa);
-    HtInvPC_ = new OperatorJacobiSmoother(diag_pa, temp_ess_tdof_);
-    HtInv_->SetPreconditioner(*HtInvPC_);
-  }
 
   // Prepare for the solve
   for (auto &temp_dbc : temp_dbcs_) {
@@ -681,12 +685,7 @@ void CaloricallyPerfectThermoChem::step() {
   sfes_->GetRestrictionMatrix()->MultTranspose(resT_, resT_gf_);
 
   Vector Xt2, Bt2;
-  if (partial_assembly_) {
-    auto *HC = Ht_.As<ConstrainedOperator>();
-    EliminateRHS(*Ht_form_, *HC, temp_ess_tdof_, Tn_next_gf_, resT_gf_, Xt2, Bt2, 1);
-  } else {
-    Ht_form_->FormLinearSystem(temp_ess_tdof_, Tn_next_gf_, resT_gf_, Ht_, Xt2, Bt2, 1);
-  }
+  Ht_form_->FormLinearSystem(temp_ess_tdof_, Tn_next_gf_, resT_gf_, Ht_, Xt2, Bt2, 1);
 
   // solve helmholtz eq for temp
   HtInv_->Mult(Bt2, Xt2);
diff --git a/src/calorically_perfect.hpp b/src/calorically_perfect.hpp
index 570124e02..118cd3dd5 100644
--- a/src/calorically_perfect.hpp
+++ b/src/calorically_perfect.hpp
@@ -76,15 +76,39 @@ class CaloricallyPerfectThermoChem : public ThermoChemModelBase {
   // Flags
   bool rank0_;                      /**< true if this is rank 0 */
   bool partial_assembly_ = false;   /**< Enable/disable partial assembly of forms. */
-  bool numerical_integ_ = true;     /**< Enable/disable numerical integration rules of forms. */
+  bool numerical_integ_ = false;    /**< Enable/disable numerical integration rules of forms. */
   bool constant_viscosity_ = false; /**< Enable/disable constant viscosity */
   bool constant_density_ = false;   /**< Enable/disable constant density */
   bool domain_is_open_ = false;     /**< true if domain is open */
 
   // Linear-solver-related options
-  int pl_solve_ = 0;    /**< Verbosity level passed to mfem solvers */
-  int max_iter_;        /**< Maximum number of linear solver iterations */
-  double rtol_ = 1e-12; /**< Linear solver relative tolerance */
+  int smoother_poly_order_;
+  double smoother_poly_fraction_ = 0.75;
+  int smoother_eig_est_ = 10;
+  int smoother_passes_ = 1;
+  double smoother_relax_weight_ = 0.4;
+  double smoother_relax_omega_ = 1.0;
+  double hsmoother_relax_weight_ = 0.8;
+  double hsmoother_relax_omega_ = 0.1;
+
+  // solver tolerance options
+  int pl_solve_; /**< Verbosity level passed to mfem solvers */
+  int max_iter_; /**< Maximum number of linear solver iterations */
+  double rtol_;  /**< Linear solver relative tolerance */
+
+  int default_max_iter_ = 1000;
+  double default_rtol_ = 1.0e-10;
+  double default_atol_ = 1.0e-12;
+
+  int mass_inverse_pl_ = 0;
+  int mass_inverse_max_iter_;
+  double mass_inverse_rtol_;
+  double mass_inverse_atol_;
+
+  int hsolve_pl_ = 0;
+  int hsolve_max_iter_;
+  double hsolve_rtol_;
+  double hsolve_atol_;
 
   // Boundary condition info
   Array<int> temp_ess_attr_; /**< List of patches with Dirichlet BC on temperature */
diff --git a/src/chemistry.cpp b/src/chemistry.cpp
index fed7512d3..5960b463e 100644
--- a/src/chemistry.cpp
+++ b/src/chemistry.cpp
@@ -109,7 +109,7 @@ MFEM_HOST_DEVICE Chemistry::~Chemistry() {
   }
 }
 
-MFEM_HOST_DEVICE void Chemistry::setRates(const double * data, int size) {
+MFEM_HOST_DEVICE void Chemistry::setRates(const double *data, int size) {
   for (int r = 0; r < numReactions_; r++) {
     if (reactions_[r]->reactionModel == GRIDFUNCTION_RXN) {
       GridFunctionReaction *rx = (GridFunctionReaction *)(reactions_[r]);
diff --git a/src/chemistry.hpp b/src/chemistry.hpp
index 906d249af..79af5a885 100644
--- a/src/chemistry.hpp
+++ b/src/chemistry.hpp
@@ -97,7 +97,7 @@ class Chemistry {
 
   // Set the grid function rates for GRIDFUNCTION_RXN reaction types
   void setGridFunctionRates(mfem::GridFunction &f);
-  MFEM_HOST_DEVICE void setRates(const double * data, int size);
+  MFEM_HOST_DEVICE void setRates(const double *data, int size);
 
   // return Vector of reaction rate coefficients, with the size of numReaction_.
   // WARNING(marc) I have removed "virtual" qualifier here assuming these functions will not
diff --git a/src/gpu_constructor.cpp b/src/gpu_constructor.cpp
index bf58b88a4..62c5d6f1b 100644
--- a/src/gpu_constructor.cpp
+++ b/src/gpu_constructor.cpp
@@ -124,11 +124,11 @@ __global__ void freeDeviceRadiation(Radiation *radiation) {
 //---------------------------------------------------
 // And finally device setters
 //---------------------------------------------------
-__global__ void deviceSetGridFunctionReactionData(const double * data, int size, GridFunctionReaction * reaction) {
+__global__ void deviceSetGridFunctionReactionData(const double *data, int size, GridFunctionReaction *reaction) {
   reaction->setData(data, size);
 }
 
-__global__ void deviceSetChemistryReactionData(const double * data, int size, Chemistry * chem) {
+__global__ void deviceSetChemistryReactionData(const double *data, int size, Chemistry *chem) {
   chem->setRates(data, size);
 }
 
diff --git a/src/gpu_constructor.hpp b/src/gpu_constructor.hpp
index 64e98d34d..18d516dd2 100644
--- a/src/gpu_constructor.hpp
+++ b/src/gpu_constructor.hpp
@@ -161,8 +161,8 @@ __global__ void freeDeviceChemistry(Chemistry *chem);
 __global__ void freeDeviceRadiation(Radiation *radiation);
 
 //! Set the data to a GridFunctionReaction
-__global__ void deviceSetGridFunctionReactionData(const double * data, int size, GridFunctionReaction * reaction);
-__global__ void deviceSetChemistryReactionData(const double * data, int size, Chemistry * chem);
+__global__ void deviceSetGridFunctionReactionData(const double *data, int size, GridFunctionReaction *reaction);
+__global__ void deviceSetChemistryReactionData(const double *data, int size, Chemistry *chem);
 
 #endif  // cuda or hip
 }  // namespace gpu
diff --git a/src/reaction.cpp b/src/reaction.cpp
index 18efb6072..54a02e478 100644
--- a/src/reaction.cpp
+++ b/src/reaction.cpp
@@ -84,12 +84,12 @@ MFEM_HOST_DEVICE GridFunctionReaction::GridFunctionReaction(int comp)
 
 MFEM_HOST_DEVICE GridFunctionReaction::~GridFunctionReaction() {}
 
-MFEM_HOST_DEVICE void GridFunctionReaction::setData(const double * data, int size) {
+MFEM_HOST_DEVICE void GridFunctionReaction::setData(const double *data, int size) {
   data_ = data + comp_ * size_;
   size_ = size;
 }
 
-void GridFunctionReaction::setGridFunction(const mfem::GridFunction & f) {
+void GridFunctionReaction::setGridFunction(const mfem::GridFunction &f) {
   size_ = f.FESpace()->GetNDofs();
   assert(comp_ < f.FESpace()->GetVDim());
   assert(f.FESpace()->GetOrdering() == mfem::Ordering::byNODES);
diff --git a/src/reaction.hpp b/src/reaction.hpp
index c563e0dd8..078c26faf 100644
--- a/src/reaction.hpp
+++ b/src/reaction.hpp
@@ -129,9 +129,9 @@ class GridFunctionReaction : public Reaction {
 
   MFEM_HOST_DEVICE virtual ~GridFunctionReaction();
 
-  void setGridFunction(const mfem::GridFunction & f);
+  void setGridFunction(const mfem::GridFunction &f);
 
-  MFEM_HOST_DEVICE void setData(const double * data, int size);
+  MFEM_HOST_DEVICE void setData(const double *data, int size);
 
   MFEM_HOST_DEVICE virtual double computeRateCoefficient([[maybe_unused]] const double &T_h,
                                                          [[maybe_unused]] const double &T_e, const int &dofindex,
diff --git a/src/tomboulides.cpp b/src/tomboulides.cpp
index d83265b6d..0932a1889 100644
--- a/src/tomboulides.cpp
+++ b/src/tomboulides.cpp
@@ -115,6 +115,19 @@ Tomboulides::Tomboulides(mfem::ParMesh *pmesh, int vorder, int porder, temporalS
     // locates quadrature points on the axis, which can lead to
     // singular mass matrices and evaluations of 1/r = 1/0.
     if (axisym_) assert(!numerical_integ_);
+
+    // exposing solver tolerance options to user
+    tps->getInput("loMach/tomboulides/psolve-atol", pressure_solve_atol_, default_atol_);
+    tps->getInput("loMach/tomboulides/hsolve-atol", hsolve_atol_, default_atol_);
+    tps->getInput("loMach/tomboulides/msolve-atol", mass_inverse_atol_, default_atol_);
+
+    tps->getInput("loMach/tomboulides/psolve-rtol", pressure_solve_rtol_, default_rtol_);
+    tps->getInput("loMach/tomboulides/hsolve-rtol", hsolve_rtol_, default_rtol_);
+    tps->getInput("loMach/tomboulides/msolve-rtol", mass_inverse_rtol_, default_rtol_);
+
+    tps->getInput("loMach/tomboulides/psolve-maxIters", pressure_solve_max_iter_, default_max_iter_);
+    tps->getInput("loMach/tomboulides/hsolve-maxIters", hsolve_max_iter_, default_max_iter_);
+    tps->getInput("loMach/tomboulides/msolve-maxIters", mass_inverse_max_iter_, default_max_iter_);
   }
 }
 
@@ -562,6 +575,16 @@ void Tomboulides::initializeSelf() {
 void Tomboulides::initializeOperators() {
   assert(thermo_interface_ != NULL);
 
+  // polynomial order for smoother of precons
+  smoother_poly_order_ = vorder_ + 1;
+
+  // AMG strength threshold
+  if (dim_ == 2) {
+    pressure_strength_thres_ = 0.25;
+  } else {
+    pressure_strength_thres_ = 0.6;
+  }
+
   // Create all the Coefficient objects we need
   rho_coeff_ = new GridFunctionCoefficient(thermo_interface_->density);
 
@@ -657,16 +680,13 @@ void Tomboulides::initializeOperators() {
     L_iorho_blfi->SetIntRule(&ir_ni_p);
   }
   L_iorho_form_->AddDomainIntegrator(L_iorho_blfi);
-  if (partial_assembly_) {
-    L_iorho_form_->SetAssemblyLevel(AssemblyLevel::PARTIAL);
-  }
   L_iorho_form_->Assemble();
   L_iorho_form_->FormSystemMatrix(pres_ess_tdof_, L_iorho_op_);
 
-  // Variable coefficient Laplacian inverse
+  // Variable coefficient Laplacian inverse (gets destoryed and rebuilt every step)
   L_iorho_lor_ = new ParLORDiscretization(*L_iorho_form_, pres_ess_tdof_);
   L_iorho_inv_pc_ = new HypreBoomerAMG(L_iorho_lor_->GetAssembledMatrix());
-  L_iorho_inv_pc_->SetPrintLevel(0);
+  L_iorho_inv_pc_->SetPrintLevel(pressure_solve_pl_);
   L_iorho_inv_ortho_pc_ = new OrthoSolver(pfes_->GetComm());
   L_iorho_inv_ortho_pc_->SetSolver(*L_iorho_inv_pc_);
 
@@ -745,9 +765,6 @@ void Tomboulides::initializeOperators() {
     msr_blfi->SetIntRule(&ir_ni_p);
   }
   Ms_rho_form_->AddDomainIntegrator(msr_blfi);
-  if (partial_assembly_) {
-    Ms_rho_form_->SetAssemblyLevel(AssemblyLevel::PARTIAL);
-  }
   Ms_rho_form_->Assemble();
   Ms_rho_form_->FormSystemMatrix(empty, Ms_rho_op_);
 
@@ -781,9 +798,6 @@ void Tomboulides::initializeOperators() {
     mvr_blfi->SetIntRule(&ir_ni_v);
   }
   Mv_rho_form_->AddDomainIntegrator(mvr_blfi);
-  if (partial_assembly_) {
-    Mv_rho_form_->SetAssemblyLevel(AssemblyLevel::PARTIAL);
-  }
   Mv_rho_form_->Assemble();
   Mv_rho_form_->FormSystemMatrix(empty, Mv_rho_op_);
 
@@ -794,29 +808,32 @@ void Tomboulides::initializeOperators() {
     Mv_inv_pc_ = new OperatorJacobiSmoother(diag_pa, empty);
   } else {
     Mv_inv_pc_ = new HypreSmoother(*Mv_op_.As<HypreParMatrix>());
-    dynamic_cast<HypreSmoother *>(Mv_inv_pc_)->SetType(HypreSmoother::Jacobi, 1);
+    dynamic_cast<HypreSmoother *>(Mv_inv_pc_)->SetType(smoother_type_, smoother_passes_);
+    dynamic_cast<HypreSmoother *>(Mv_inv_pc_)->SetSOROptions(smoother_relax_weight_, smoother_relax_omega_);
+    dynamic_cast<HypreSmoother *>(Mv_inv_pc_)
+        ->SetPolyOptions(smoother_poly_order_, smoother_poly_fraction_, smoother_eig_est_);
   }
   Mv_inv_ = new CGSolver(vfes_->GetComm());
   Mv_inv_->iterative_mode = false;
   Mv_inv_->SetOperator(*Mv_op_);
   Mv_inv_->SetPreconditioner(*Mv_inv_pc_);
   Mv_inv_->SetPrintLevel(mass_inverse_pl_);
+  Mv_inv_->SetAbsTol(mass_inverse_atol_);
   Mv_inv_->SetRelTol(mass_inverse_rtol_);
   Mv_inv_->SetMaxIter(mass_inverse_max_iter_);
 
-  if (partial_assembly_) {
-    Vector diag_pa(vfes_->GetTrueVSize());
-    Mv_rho_form_->AssembleDiagonal(diag_pa);
-    Mv_rho_inv_pc_ = new OperatorJacobiSmoother(diag_pa, empty);
-  } else {
-    Mv_rho_inv_pc_ = new HypreSmoother(*Mv_rho_op_.As<HypreParMatrix>());
-    dynamic_cast<HypreSmoother *>(Mv_rho_inv_pc_)->SetType(HypreSmoother::Jacobi, 1);
-  }
+  Mv_rho_inv_pc_ = new HypreSmoother(*Mv_rho_op_.As<HypreParMatrix>());
+  dynamic_cast<HypreSmoother *>(Mv_rho_inv_pc_)->SetType(smoother_type_, smoother_passes_);
+  dynamic_cast<HypreSmoother *>(Mv_rho_inv_pc_)->SetSOROptions(smoother_relax_weight_, smoother_relax_omega_);
+  dynamic_cast<HypreSmoother *>(Mv_rho_inv_pc_)
+      ->SetPolyOptions(smoother_poly_order_, smoother_poly_fraction_, smoother_eig_est_);
+
   Mv_rho_inv_ = new CGSolver(vfes_->GetComm());
   Mv_rho_inv_->iterative_mode = false;
   Mv_rho_inv_->SetOperator(*Mv_rho_op_);
   Mv_rho_inv_->SetPreconditioner(*Mv_rho_inv_pc_);
   Mv_rho_inv_->SetPrintLevel(mass_inverse_pl_);
+  Mv_rho_inv_->SetAbsTol(mass_inverse_atol_);
   Mv_rho_inv_->SetRelTol(mass_inverse_rtol_);
   Mv_rho_inv_->SetMaxIter(mass_inverse_max_iter_);
 
@@ -878,29 +895,22 @@ void Tomboulides::initializeOperators() {
     auto *hfv_blfi = new VectorMassIntegrator(*visc_forcing_coeff_);
     Hv_form_->AddDomainIntegrator(hfv_blfi);
   }
-  if (partial_assembly_) {
-    // Partial assembly is not supported for variable coefficient
-    // VectorMassIntegrator (as of mfem 4.5.2 at least)
-    assert(false);
-    Hv_form_->SetAssemblyLevel(AssemblyLevel::PARTIAL);
-  }
   Hv_form_->Assemble();
   Hv_form_->FormSystemMatrix(vel_ess_tdof_, Hv_op_);
 
   // Helmholtz solver
-  if (partial_assembly_) {
-    Vector diag_pa(vfes_->GetTrueVSize());
-    Hv_form_->AssembleDiagonal(diag_pa);
-    Hv_inv_pc_ = new OperatorJacobiSmoother(diag_pa, vel_ess_tdof_);
-  } else {
-    Hv_inv_pc_ = new HypreSmoother(*Hv_op_.As<HypreParMatrix>());
-    dynamic_cast<HypreSmoother *>(Hv_inv_pc_)->SetType(HypreSmoother::Jacobi, 1);
-  }
+  Hv_inv_pc_ = new HypreSmoother(*Hv_op_.As<HypreParMatrix>());
+  dynamic_cast<HypreSmoother *>(Hv_inv_pc_)->SetType(smoother_type_, smoother_passes_);
+  dynamic_cast<HypreSmoother *>(Hv_inv_pc_)->SetSOROptions(hsmoother_relax_weight_, hsmoother_relax_omega_);
+  dynamic_cast<HypreSmoother *>(Hv_inv_pc_)
+      ->SetPolyOptions(smoother_poly_order_, smoother_poly_fraction_, smoother_eig_est_);
+
   Hv_inv_ = new CGSolver(vfes_->GetComm());
   Hv_inv_->iterative_mode = true;
   Hv_inv_->SetOperator(*Hv_op_);
   Hv_inv_->SetPreconditioner(*Hv_inv_pc_);
   Hv_inv_->SetPrintLevel(hsolve_pl_);
+  Hv_inv_->SetAbsTol(hsolve_atol_);
   Hv_inv_->SetRelTol(hsolve_rtol_);
   Hv_inv_->SetMaxIter(hsolve_max_iter_);
 
@@ -986,13 +996,17 @@ void Tomboulides::initializeOperators() {
     Hs_form_->FormSystemMatrix(swirl_ess_tdof_, Hs_op_);
 
     Hs_inv_pc_ = new HypreSmoother(*Hs_op_.As<HypreParMatrix>());
-    dynamic_cast<HypreSmoother *>(Hs_inv_pc_)->SetType(HypreSmoother::Jacobi, 1);
+    dynamic_cast<HypreSmoother *>(Hs_inv_pc_)->SetType(smoother_type_, smoother_passes_);
+    dynamic_cast<HypreSmoother *>(Hs_inv_pc_)->SetSOROptions(hsmoother_relax_weight_, hsmoother_relax_omega_);
+    dynamic_cast<HypreSmoother *>(Hs_inv_pc_)
+        ->SetPolyOptions(smoother_poly_order_, smoother_poly_fraction_, smoother_eig_est_);
 
     Hs_inv_ = new CGSolver(pfes_->GetComm());
     Hs_inv_->iterative_mode = true;
     Hs_inv_->SetOperator(*Hs_op_);
     Hs_inv_->SetPreconditioner(*Hs_inv_pc_);
     Hs_inv_->SetPrintLevel(hsolve_pl_);
+    Hs_inv_->SetAbsTol(hsolve_atol_);
     Hs_inv_->SetRelTol(hsolve_rtol_);
     Hs_inv_->SetMaxIter(hsolve_max_iter_);
 
@@ -1134,7 +1148,17 @@ void Tomboulides::step() {
 
   L_iorho_lor_ = new ParLORDiscretization(*L_iorho_form_, pres_ess_tdof_);
   L_iorho_inv_pc_ = new HypreBoomerAMG(L_iorho_lor_->GetAssembledMatrix());
-  L_iorho_inv_pc_->SetPrintLevel(0);
+  L_iorho_inv_pc_->SetPrintLevel(pressure_solve_pl_);
+  L_iorho_inv_pc_->SetCoarsening(amg_coarsening_);
+  L_iorho_inv_pc_->SetAggressiveCoarsening(amg_aggresive_);
+  L_iorho_inv_pc_->SetInterpolation(amg_interpolation_);
+  L_iorho_inv_pc_->SetStrengthThresh(pressure_strength_thres_);
+  L_iorho_inv_pc_->SetRelaxType(amg_relax_);
+  L_iorho_inv_pc_->SetMaxLevels(amg_max_levels_);
+  L_iorho_inv_pc_->SetMaxIter(amg_max_iters_);
+  L_iorho_inv_pc_->SetTol(0);
+  // NOTE: other cycle types (i.e. not V-cycle) do not scale well in parallel, so option removed for now
+  // L_iorho_inv_pc_->SetCycleType(1);
   L_iorho_inv_ortho_pc_ = new OrthoSolver(pfes_->GetComm());
   L_iorho_inv_ortho_pc_->SetSolver(*L_iorho_inv_pc_);
 
@@ -1168,10 +1192,7 @@ void Tomboulides::step() {
   Hv_form_->FormSystemMatrix(vel_ess_tdof_, Hv_op_);
 
   Hv_inv_->SetOperator(*Hv_op_);
-  if (partial_assembly_) {
-    // TODO(trevilo): Support partial assembly
-    assert(false);
-  }
+
   sw_helm_.Stop();
 
   //------------------------------------------------------------------------
@@ -1375,15 +1396,7 @@ void Tomboulides::step() {
   pfes_->GetRestrictionMatrix()->MultTranspose(resp_vec_, *resp_gf_);
 
   Vector X1, B1;
-  if (partial_assembly_) {
-    // TODO(trevilo): Support partial assembly here
-    assert(false);
-    // auto *SpC = Sp.As<ConstrainedOperator>();
-    // EliminateRHS(*Sp_form, *SpC, pres_ess_tdof, pn_gf, resp_gf, X1, B1, 1);
-  } else {
-    L_iorho_form_->FormLinearSystem(pres_ess_tdof_, *p_gf_, *resp_gf_, L_iorho_op_, X1, B1, 1);
-  }
-
+  L_iorho_form_->FormLinearSystem(pres_ess_tdof_, *p_gf_, *resp_gf_, L_iorho_op_, X1, B1, 1);
   L_iorho_inv_->Mult(B1, X1);
   if (!L_iorho_inv_->GetConverged()) {
     if (rank0_) std::cout << "ERROR: Poisson solve did not converge." << std::endl;
@@ -1430,15 +1443,7 @@ void Tomboulides::step() {
   vfes_->GetRestrictionMatrix()->MultTranspose(resu_vec_, *resu_gf_);
 
   Vector X2, B2;
-  if (partial_assembly_) {
-    // TODO(trevilo): Add partial assembly support
-    assert(false);
-    // auto *HC = H.As<ConstrainedOperator>();
-    // EliminateRHS(*Hv_form_, *HC, vel_ess_tdof, un_next_gf, resu_gf, X2, B2, 1);
-  } else {
-    Hv_form_->FormLinearSystem(vel_ess_tdof_, *u_next_gf_, *resu_gf_, Hv_op_, X2, B2, 1);
-  }
-
+  Hv_form_->FormLinearSystem(vel_ess_tdof_, *u_next_gf_, *resu_gf_, Hv_op_, X2, B2, 1);
   Hv_inv_->Mult(B2, X2);
   if (!Hv_inv_->GetConverged()) {
     if (rank0_) std::cout << "ERROR: Helmholtz solve did not converge." << std::endl;
diff --git a/src/tomboulides.hpp b/src/tomboulides.hpp
index 1dd9cc047..8bc5d8584 100644
--- a/src/tomboulides.hpp
+++ b/src/tomboulides.hpp
@@ -80,22 +80,88 @@ class Tomboulides final : public FlowBase {
   bool axisym_;
 
   // Options
-  // TODO(trevilo): hardcoded for testing.  Need to set based on input file.
-  bool numerical_integ_ = true;
+  bool numerical_integ_ = false;
   bool partial_assembly_ = false;
 
+  // linear-solver options
+  int smoother_poly_order_;
+  double smoother_poly_fraction_ = 0.75;
+  int smoother_eig_est_ = 10;
+  int smoother_passes_ = 1;
+  double smoother_relax_weight_ =
+      0.4;  // NOTE: setting relax wt to 0 should determine wt dynamically but this does not seem to work
+  double smoother_relax_omega_ = 1.0;
+  double hsmoother_relax_weight_ = 0.8;
+  double hsmoother_relax_omega_ = 0.1;
+
+  //  Options: Jacobi, l1Jacobi, l1GS, l1GStr, lumpedJacobi,
+  //           GS, OPFS, Chebyshev, Taubin, FIR
+  mfem::HypreSmoother::Type smoother_type_ = HypreSmoother::Jacobi;
+
+  double pressure_strength_thres_ = 0.6;
+  int amg_aggresive_ = 4;
+  int amg_max_levels_ = 10;
+  int amg_max_iters_ =
+      1;  // should be 1 for precon, setting to zero is ~30-40% faster per step but effectively turns off solve
+  int amg_relax_ = 18;  // only 0 or 18 now
+
+  /*
+  0 : CLJP-coarsening (a parallel coarsening algorithm using independent sets.
+  1 : classical Ruge-Stueben coarsening on each processor, no boundary treatment (not recommended!)
+  3 : classical Ruge-Stueben coarsening on each processor, followed by a third pass, which adds coarse points on the
+  boundaries 6 : Falgout coarsening (uses 1 first, followed by CLJP using the interior coarse points generated by 1 as
+  its first independent set) 7 : CLJP-coarsening (using a fixed random vector, for debugging purposes only) 8 :
+  PMIS-coarsening (a parallel coarsening algorithm using independent sets, generating lower complexities than CLJP,
+  might also lead to slower convergence) 9 : PMIS-coarsening (using a fixed random vector, for debugging purposes only)
+  10 : HMIS-coarsening (uses one pass Ruge-Stueben on each processor independently, followed by PMIS using the interior
+  C-points generated as its first independent set) 11 : one-pass Ruge-Stueben coarsening on each processor, no boundary
+  treatment (not recommended!) 21 : CGC coarsening by M. Griebel, B. Metsch and A. Schweitzer 22 : CGC-E coarsening by
+  M. Griebel, B. Metsch and A.Schweitzer
+  */
+  int amg_coarsening_ = 8;
+
+  /*
+  0 : classical modified interpolation
+  1 : LS interpolation (for use with GSMG)
+  2 : classical modified interpolation for hyperbolic PDEs
+  3 : direct interpolation (with separation of weights) (also for GPU use)
+  4 : multipass interpolation
+  5 : multipass interpolation (with separation of weights)
+  6 : extended+i interpolation (also for GPU use) ***
+  7 : extended+i (if no common C neighbor) interpolation
+  8 : standard interpolation
+  9 : standard interpolation (with separation of weights)
+  10 : classical block interpolation (for use with nodal systems version only)
+  11 : classical block interpolation (for use with nodal systems version only) with diagonalized diagonal blocks
+  12 : FF interpolation
+  13 : FF1 interpolation
+  14 : extended interpolation (also for GPU use)
+  15 : interpolation with adaptive weights (GPU use only)
+  16 : extended interpolation in matrix-matrix form
+  17 : extended+i interpolation in matrix-matrix form
+  18 : extended+e interpolation in matrix-matrix form
+  */
+  int amg_interpolation_ = 14;
+
+  // solver tolerance options
+  int default_max_iter_ = 1000;
+  double default_rtol_ = 1e-10;
+  double default_atol_ = 1e-12;
+
   int pressure_solve_pl_ = 0;
-  int pressure_solve_max_iter_ = 100;
-  double pressure_solve_rtol_ = 1e-8;
-  double pressure_solve_atol_ = 1e-18;
+  int pressure_solve_max_iter_;
+  double pressure_solve_rtol_;
+  double pressure_solve_atol_;
 
   int mass_inverse_pl_ = 0;
-  int mass_inverse_max_iter_ = 200;
-  double mass_inverse_rtol_ = 1e-12;
+  int mass_inverse_max_iter_;
+  double mass_inverse_rtol_;
+  double mass_inverse_atol_;
 
   int hsolve_pl_ = 0;
-  int hsolve_max_iter_ = 200;
-  double hsolve_rtol_ = 1e-12;
+  int hsolve_max_iter_;
+  double hsolve_rtol_;
+  double hsolve_atol_;
 
   // To use "numerical integration", quadrature rule must persist
   mfem::IntegrationRules gll_rules;
diff --git a/src/tps.cpp b/src/tps.cpp
index b82b7caf3..7e0cd6c39 100644
--- a/src/tps.cpp
+++ b/src/tps.cpp
@@ -140,8 +140,7 @@ void Tps::parseCommandLineArgs(int argc, char *argv[]) {
   args.AddOption(&visualMode, "-visual", "--visualization", "", "--no-visualization",
                  "Launch post-process visualization.");
   gpu_aware_mpi_ = false;
-  args.AddOption(&gpu_aware_mpi_, "-ga", "--gpu-aware-mpi", "", "--no-gpu-aware-mpi",
-                 "Set GPU-aware MPI.");
+  args.AddOption(&gpu_aware_mpi_, "-ga", "--gpu-aware-mpi", "", "--no-gpu-aware-mpi", "Set GPU-aware MPI.");
   args.Parse();
 
   if (!args.Good()) {
diff --git a/src/tps2Boltzmann.cpp b/src/tps2Boltzmann.cpp
index 72f9b0678..689fded06 100644
--- a/src/tps2Boltzmann.cpp
+++ b/src/tps2Boltzmann.cpp
@@ -94,12 +94,10 @@ Tps2Boltzmann::Tps2Boltzmann(Tps *tps)
 
   save_to_paraview_dc = tps->getInput("boltzmannInterface/save_to_paraview", false);
 
-
   offsets.SetSize(NIndexes + 1);
   ncomps.SetSize(NIndexes + 1);
 }
 
-
 int Tps2Boltzmann::_countBTEReactions() {
   int total_reactions(0);
   int bte_reactions(0);
diff --git a/test/autoPeriodic.test b/test/autoPeriodic.test
index dc6a9d33e..de811ddf7 100755
--- a/test/autoPeriodic.test
+++ b/test/autoPeriodic.test
@@ -9,6 +9,7 @@ RESTART="ref_solns/sgsLoMach/restart_output.sol.h5"
 setup() {
     SOLN_FILE=restart_output.sol.h5
     REF_FILE=ref_solns/sgsLoMach/restart_output.sol.h5
+    OUT_FILE=output_solns/restart_output_autoPeriodic.sol.h5    
 }
 
 @test "[$TEST] check for input file $RUNFILE" {
@@ -26,5 +27,6 @@ setup() {
 @test "[$TEST] verify tps output with input -> $RUNFILE" {
     test -s $SOLN_FILE
     test -s $REF_FILE
-    h5diff -r --delta=1e-12 $SOLN_FILE $REF_FILE /velocity
+    h5diff -r --delta=1e-10 $SOLN_FILE $REF_FILE /velocity
+    # cp $SOLN_FILE $OUT_FILE
 }
diff --git a/test/aveLoMach.test b/test/aveLoMach.test
index 31af36add..008c43efc 100755
--- a/test/aveLoMach.test
+++ b/test/aveLoMach.test
@@ -9,6 +9,7 @@ RESTART="ref_solns/aveLoMach/restart_output.sol.h5"
 setup() {
     SOLN_FILE=restart_output.sol.h5
     REF_FILE=ref_solns/aveLoMach/restart_output.sol.h5
+    OUT_FILE=output_solns/restart_output_aveLoMach.sol.h5    
 }
 
 @test "[$TEST] check for input file $RUNFILE" {
@@ -26,5 +27,6 @@ setup() {
 @test "[$TEST] verify tps output with input -> $RUNFILE" {
     test -s $SOLN_FILE
     test -s $REF_FILE
-    h5diff -r --delta=1e-12 $SOLN_FILE $REF_FILE meanVel
+    h5diff -r --delta=1e-11 $SOLN_FILE $REF_FILE meanVel
+    # cp $SOLN_FILE $OUT_FILE
 }
diff --git a/test/heatEq.test b/test/heatEq.test
index 771aa9296..ac6108dd5 100755
--- a/test/heatEq.test
+++ b/test/heatEq.test
@@ -9,6 +9,7 @@ RESTART="ref_solns/heatedBox/restart_output.sol.h5"
 setup() {
     SOLN_FILE=restart_output.sol.h5
     REF_FILE=ref_solns/heatedBox/restart_output.sol.h5
+    OUT_FILE=output_solns/restart_output_heatedBox.sol.h5    
 }
 
 @test "[$TEST] check for input file $RUNFILE" {
@@ -25,5 +26,6 @@ setup() {
 @test "[$TEST] verify tps output with input -> $RUNFILE" {
     test -s $SOLN_FILE
     test -s $REF_FILE
-    h5diff -r --relative=1e-13 $SOLN_FILE $REF_FILE /temperature/temperature
+    h5diff -r --relative=1e-10 $SOLN_FILE $REF_FILE /temperature/temperature
+    # cp $SOLN_FILE $OUT_FILE
 }
diff --git a/test/inputs/input.autoPeriodic.ini b/test/inputs/input.autoPeriodic.ini
index 56fc43205..a462938b6 100644
--- a/test/inputs/input.autoPeriodic.ini
+++ b/test/inputs/input.autoPeriodic.ini
@@ -3,35 +3,34 @@ type = loMach
 
 [loMach]
 flow-solver = tomboulides
+thermo-solver = constant-property
 mesh = meshes/flatBox.msh
-uOrder = 1
-pOrder = 1
-nOrder = 1
+order = 1
 maxIters = 10
 outputFreq = 10
-fluid = dry_air
-refLength = 0.05
-equation_system = navier-stokes
 enablePressureForcing = False
-constantViscosity = 1.552e-5
-constantDensity = 1.1839
 enableGravity = False
+gravity = '0 0 0'
 openSystem = True
 ambientPressure = 101326.
 sgsModel = smagorinsky
 sgsModelConstant = 0.09
 
 [loMach/tomboulides]
-ic = tgv2d
+numerical-integ = false
+ic = tgv2d_uniform
+
+[loMach/constprop]
+mu = 1.552e-5
+rho = 1.1839
 
 [io]
 outdirBase = output
 
 [time]
-cfl = 0.5
+integrator = curlcurl
 dt_fixed = 1.0e-3
-maxSolverIteration = 1000
-solverRelTolerance = 1.0e-12
+bdfOrder = 3
 
 [initialConditions]
 rho = 1.1839
@@ -49,4 +48,4 @@ numOutlets = 0
 enablePeriodic = True
 periodicX = True
 periodicY = True
-periodicZ = True
\ No newline at end of file
+periodicZ = True
diff --git a/test/inputs/input.heatedBox.ini b/test/inputs/input.heatedBox.ini
index c3b68817a..c603db677 100644
--- a/test/inputs/input.heatedBox.ini
+++ b/test/inputs/input.heatedBox.ini
@@ -4,9 +4,6 @@ type = loMach
 [loMach]
 mesh = meshes/lequereBoxCrs3wall.msh
 order = 1
-uOrder = 1
-pOrder = 1
-nOrder = 1
 nFilter = 0
 filterWeight = 1.0
 maxIters = 5000
diff --git a/test/inputs/input.sgsLoMach.ini b/test/inputs/input.sgsLoMach.ini
index a884829c6..ad3e812db 100644
--- a/test/inputs/input.sgsLoMach.ini
+++ b/test/inputs/input.sgsLoMach.ini
@@ -3,35 +3,34 @@ type = loMach
 
 [loMach]
 flow-solver = tomboulides
+thermo-solver = constant-property
 mesh = meshes/flatBox.msh
-uOrder = 1
-pOrder = 1
-nOrder = 1
+order = 1
 maxIters = 10
 outputFreq = 10
-fluid = dry_air
-refLength = 0.05
-equation_system = navier-stokes
 enablePressureForcing = False
-constantViscosity = 1.552e-5
-constantDensity = 1.1839
 enableGravity = False
+gravity = '0 0 0'
 openSystem = True
 ambientPressure = 101326.
 sgsModel = smagorinsky
 sgsModelConstant = 0.09
 
 [loMach/tomboulides]
-ic = tgv2d
+numerical-integ = false
+ic = tgv2d_uniform
+
+[loMach/constprop]
+mu = 1.552e-5
+rho = 1.1839
 
 [io]
 outdirBase = output
 
 [time]
-cfl = 0.5
+integrator = curlcurl
 dt_fixed = 1.0e-3
-maxSolverIteration = 1000
-solverRelTolerance = 1.0e-12
+bdfOrder = 3
 
 [initialConditions]
 rho = 1.1839
diff --git a/test/interpInlet.test b/test/interpInlet.test
index e9fc0e30f..0af9a9283 100755
--- a/test/interpInlet.test
+++ b/test/interpInlet.test
@@ -9,6 +9,7 @@ RESTART=
 setup() {
     SOLN_FILE=restart_output.sol.h5
     REF_FILE=ref_solns/interpInlet/restart_output.sol.h5
+    OUT_FILE=output_solns/restart_output_interpInlet.sol.h5    
 }
 
 @test "[$TEST] check for input file $RUNFILE" {
@@ -33,5 +34,6 @@ setup() {
 @test "[$TEST] verify tps output with input -> $RUNFILE" {
     test -s $SOLN_FILE
     test -s $REF_FILE
-    h5diff -r --relative=1e-12 $SOLN_FILE $REF_FILE /temperature/temperature
+    h5diff -r --relative=1e-10 $SOLN_FILE $REF_FILE /temperature/temperature
+    # cp $SOLN_FILE $OUT_FILE
 }
diff --git a/test/lomach-flow.test b/test/lomach-flow.test
index a54cfaedd..eba58c5b9 100755
--- a/test/lomach-flow.test
+++ b/test/lomach-flow.test
@@ -16,12 +16,15 @@ setup() {
 
     RUNFILE_PIPE="inputs/lomach.pipe.ini"
     REF_SOLN_PIPE="ref_solns/pipe/restart_output-pipe-lam.sol.h5"
+    OUT_SOLN_PIPE="output_solns/restart_output-pipe-lam.sol.h5"    
 
     RUNFILE_TC="inputs/lomach.taylor.couette.ini"
     REF_SOLN_TC="ref_solns/taylor-couette/restart_output-tc.Re100.h5"
+    OUT_SOLN_TC="output_solns/restart_output-tc.Re100.h5"    
 
     RUNFILE_PIPE_ARANS="inputs/lomach.pipe.arans.ini"
     REF_SOLN_PIPE_ARANS="ref_solns/pipe/restart_output-pipe-arans.sol.h5"
+    OUT_SOLN_PIPE_ARANS="output_solns/restart_output-pipe-arans.sol.h5"    
 
     MPIRUN=`./sniff_mpirun.sh`
 }
@@ -109,8 +112,8 @@ setup() {
     [[ ${status} -eq 0 ]]
 
     # check solutions are the same!
-    h5diff -r --delta=1e-12 restart_output-lid.sol.h5 restart_output-lid.sol.2.h5
-
+    h5diff -r --delta=1e-11 restart_output-lid.sol.h5 restart_output-lid.sol.2.h5  
+    
     # delete intermediate files
     rm -f restart_output-lid*.h5
     rm $RUNFILE_LID_MOD
@@ -138,7 +141,7 @@ setup() {
     $MPIRUN -n 2 $EXE --runFile $RUNFILE_LID_MOD
 
     # check solutions are the same!
-    h5diff -r --delta=1e-12 restart_output-lid.sol.h5 restart_output-lid.sol.2.h5
+    h5diff -r --delta=1e-10 restart_output-lid.sol.h5 restart_output-lid.sol.2.h5
 
     # delete intermediate files
     rm -f restart_output-lid*.h5
@@ -149,22 +152,25 @@ setup() {
     run $EXE --runFile $RUNFILE_PIPE
     [[ ${status} -eq 0 ]]
     test -s restart_output-pipe-lam.sol.h5
-    h5diff -r --delta=1e-12 restart_output-pipe-lam.sol.h5 $REF_SOLN_PIPE /velocity
-    h5diff -r --delta=1e-12 restart_output-pipe-lam.sol.h5 $REF_SOLN_PIPE /swirl
+    h5diff -r --delta=1e-10 restart_output-pipe-lam.sol.h5 $REF_SOLN_PIPE /velocity
+    h5diff -r --delta=1e-10 restart_output-pipe-lam.sol.h5 $REF_SOLN_PIPE /swirl
+    # cp restart_output-pipe-lam.sol.h5 $OUT_SOLN_PIPE
 }
 
 @test "[$TEST] check that Taylor-Couette (axisymmetric, with swirl) runs and verify result" {
     run $EXE --runFile $RUNFILE_TC
     [[ ${status} -eq 0 ]]
     test -s restart_output-tc.sol.h5
-    h5diff -r --delta=1e-12 restart_output-tc.sol.h5 $REF_SOLN_TC /velocity
-    h5diff -r --delta=1e-12 restart_output-tc.sol.h5 $REF_SOLN_TC /swirl
+    h5diff -r --delta=1e-10 restart_output-tc.sol.h5 $REF_SOLN_TC /velocity
+    h5diff -r --delta=1e-10 restart_output-tc.sol.h5 $REF_SOLN_TC /swirl
+    # cp restart_output-tc.sol.h5 $OUT_SOLN_TC    
 }
 
 @test "[$TEST] check pipe (axisymmetric) flow with algebraic rans model" {
     run $EXE --runFile $RUNFILE_PIPE_ARANS
     [[ ${status} -eq 0 ]]
     test -s restart_output-pipe-arans.sol.h5
-    h5diff -r --delta=1e-12 restart_output-pipe-arans.sol.h5 $REF_SOLN_PIPE_ARANS /velocity
-    h5diff -r --delta=1e-12 restart_output-pipe-arans.sol.h5 $REF_SOLN_PIPE_ARANS /swirl
+    h5diff -r --delta=1e-9 restart_output-pipe-arans.sol.h5 $REF_SOLN_PIPE_ARANS /velocity
+    h5diff -r --delta=1e-9 restart_output-pipe-arans.sol.h5 $REF_SOLN_PIPE_ARANS /swirl
+    # cp restart_output-pipe-arans.sol.h5 $OUT_SOLN_PIPE_ARANS
 }
diff --git a/test/lomach-lequere.test b/test/lomach-lequere.test
index 4738f3ac9..546b91b13 100755
--- a/test/lomach-lequere.test
+++ b/test/lomach-lequere.test
@@ -10,6 +10,7 @@ setup() {
     RUNFILE="inputs/lomach.lequere.varmu.ini"
     REF_SOLN="ref_solns/lequere-varmu/reference-lequere-varmu.sol.h5"
     SOLN_FILE="restart_output-lequere-varmu.sol.h5"
+    OUT_SOLN="output_solns/reference-lequere-varmu.sol.h5"   
 }
 
 @test "[$TEST] Le Quere low Mach, thermally-driven cavity regression with variable viscosity" {
@@ -24,8 +25,9 @@ setup() {
     $EXE --runFile $RUNFILE
 
     # check
-    h5diff -r --relative=1e-12 $SOLN_FILE $REF_SOLN /temperature/temperature
-    h5diff -r --relative=1e-10 $SOLN_FILE $REF_SOLN /velocity/x-comp
-    h5diff -r --relative=1e-10 $SOLN_FILE $REF_SOLN /velocity/y-comp
+    h5diff -r --relative=1e-10 $SOLN_FILE $REF_SOLN /temperature/temperature
+    h5diff -r --delta=1e-10 $SOLN_FILE $REF_SOLN /velocity/x-comp
+    h5diff -r --delta=1e-10 $SOLN_FILE $REF_SOLN /velocity/y-comp
+    # cp $SOLN_FILE $OUT_SOLN
 }
 
diff --git a/test/ref_solns/aveLoMach/restart_output.sol.h5 b/test/ref_solns/aveLoMach/restart_output.sol.h5
index 48d9b67a7..fc8a71bcd 100644
--- a/test/ref_solns/aveLoMach/restart_output.sol.h5
+++ b/test/ref_solns/aveLoMach/restart_output.sol.h5
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c3d84bab6ef05f42cdad7b3e28a758b3a1f32b0eaaf55ec3cdf0a7b04a47fec0
+oid sha256:15bb2ec395f79dc7d28324a61467760e675f3501b5582b56ff7a294146cb337d
 size 382040
diff --git a/test/ref_solns/heatedBox/restart_output.sol.h5 b/test/ref_solns/heatedBox/restart_output.sol.h5
index dc428085f..cf108bb26 100644
--- a/test/ref_solns/heatedBox/restart_output.sol.h5
+++ b/test/ref_solns/heatedBox/restart_output.sol.h5
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a7e7088904222b6e9c29b70a410f82885f389e2c9fb672686df5ff418ea92696
-size 3504
+oid sha256:02d366574f35b75cc97fdfe330450832cdb98937bcbdf3fd39f4f79921a6bab9
+size 3568
diff --git a/test/ref_solns/interpInlet/restart_output.sol.h5 b/test/ref_solns/interpInlet/restart_output.sol.h5
index 5fc3e60c0..78a7f4a38 100644
--- a/test/ref_solns/interpInlet/restart_output.sol.h5
+++ b/test/ref_solns/interpInlet/restart_output.sol.h5
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8c782a4cdbb7228c7ffd19f262ba1f2440ed132ca055e2e76b54c3a3c2630c5d
-size 8664
+oid sha256:c0639671f8f1f9496eb313f87d4fa40bcada3713d9d6f4bbd67b4bbf81405a32
+size 8728
diff --git a/test/ref_solns/lequere-varmu/reference-lequere-varmu.sol.h5 b/test/ref_solns/lequere-varmu/reference-lequere-varmu.sol.h5
index 13e2013d8..b62cc8878 100644
--- a/test/ref_solns/lequere-varmu/reference-lequere-varmu.sol.h5
+++ b/test/ref_solns/lequere-varmu/reference-lequere-varmu.sol.h5
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:992d58bb5010f3576f13fd076f2b2ab67b4ead1536b627d7e05649c6acfccc87
+oid sha256:222e94021b4b7ed239f24c8b486a45b89a7bf7d7a15d621273f2b504bf14900a
 size 106416
diff --git a/test/ref_solns/pipe/restart_output-pipe-arans.sol.h5 b/test/ref_solns/pipe/restart_output-pipe-arans.sol.h5
index 5e71a3a80..dd9e8665d 100644
--- a/test/ref_solns/pipe/restart_output-pipe-arans.sol.h5
+++ b/test/ref_solns/pipe/restart_output-pipe-arans.sol.h5
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fc4128fc860c4f6f3a5419ed3117af719e8968cfee76b85bee28bb9019667e90
+oid sha256:39061c77f5cc54f664ce69f5f91ae3f5f5386a04616efa7223cce39793234b97
 size 56496
diff --git a/test/ref_solns/pipe/restart_output-pipe-lam.sol.h5 b/test/ref_solns/pipe/restart_output-pipe-lam.sol.h5
index 636445c61..953709889 100644
--- a/test/ref_solns/pipe/restart_output-pipe-lam.sol.h5
+++ b/test/ref_solns/pipe/restart_output-pipe-lam.sol.h5
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7fe39fdee999551ea2e82c2e492f0cafd8567e9c0bae0066a02063e05cac8bc4
-size 18416
+oid sha256:3af1089c90ac4118d913cf80d7e9e9ee260e2ad94e0a5c0310a2fde4ae63c574
+size 18480
diff --git a/test/ref_solns/sgsLoMach/restart_output.sol.h5 b/test/ref_solns/sgsLoMach/restart_output.sol.h5
index 333ab85e4..c56fe05a7 100644
--- a/test/ref_solns/sgsLoMach/restart_output.sol.h5
+++ b/test/ref_solns/sgsLoMach/restart_output.sol.h5
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a1d285d59acd77b7cc383d646e8b2ddf10b1da771dad96bf0a5f46afa08df0bc
+oid sha256:663299c7c61d64043b0e02a6da2e6c315bf0ff9d63736bbc7ba8ec27b64a4840
 size 78744
diff --git a/test/ref_solns/spongeBox/restart_output.sol.h5 b/test/ref_solns/spongeBox/restart_output.sol.h5
index 0251aed2e..75c4efff5 100644
--- a/test/ref_solns/spongeBox/restart_output.sol.h5
+++ b/test/ref_solns/spongeBox/restart_output.sol.h5
@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6369b60adb2786bcd2c477743345cc303d33e73a105db3d39fc114990a6f96dd
-size 8664
+oid sha256:197ed2f249869bc6f94a5649bb6d3f4aff73dc406fc94ca98681e48dbd01b6e3
+size 8728
diff --git a/test/sgsLoMach.test b/test/sgsLoMach.test
index d572d711d..8034a54b6 100755
--- a/test/sgsLoMach.test
+++ b/test/sgsLoMach.test
@@ -9,6 +9,7 @@ RESTART="ref_solns/sgsLoMach/restart_output.sol.h5"
 setup() {
     SOLN_FILE=restart_output.sol.h5
     REF_FILE=ref_solns/sgsLoMach/restart_output.sol.h5
+    OUT_FILE=output_solns/restart_output_sgsLoMach.sol.h5    
 }
 
 @test "[$TEST] check for input file $RUNFILE" {
@@ -26,5 +27,6 @@ setup() {
 @test "[$TEST] verify tps output with input -> $RUNFILE" {
     test -s $SOLN_FILE
     test -s $REF_FILE
-    h5diff -r --delta=1e-12 $SOLN_FILE $REF_FILE /velocity
+    h5diff -r --delta=1e-10 $SOLN_FILE $REF_FILE /velocity
+    # cp $SOLN_FILE $OUT_FILE
 }
diff --git a/test/sponge.test b/test/sponge.test
index c06d1757a..d69ff6edc 100755
--- a/test/sponge.test
+++ b/test/sponge.test
@@ -33,5 +33,5 @@ setup() {
 @test "[$TEST] verify tps output with input -> $RUNFILE" {
     test -s $SOLN_FILE
     test -s $REF_FILE
-    h5diff -r --relative=1e-13 $SOLN_FILE $REF_FILE /temperature/temperature
+    h5diff -r --relative=1e-10 $SOLN_FILE $REF_FILE /temperature/temperature
 }