diff --git a/src/cpp/main_3d/engine_cpu.cpp b/src/cpp/main_3d/engine_cpu.cpp
index 5a7d47c..1305e01 100644
--- a/src/cpp/main_3d/engine_cpu.cpp
+++ b/src/cpp/main_3d/engine_cpu.cpp
@@ -27,7 +27,9 @@
 #include <cstdlib>
 #include <vector>
 
-auto runSim(Simulation3D& sd) -> double {
+namespace pffdtd {
+
+auto run(Simulation3D& sd) -> double {
   // keep local ints, scalars
   int64_t Ns   = sd.Ns;
   int64_t Nr   = sd.Nr;
@@ -416,3 +418,5 @@ double process_bnl_pts_fd(
   }
   return omp_get_wtime() - tstart;
 }
+
+} // namespace pffdtd
diff --git a/src/cpp/main_3d/engine_cpu.hpp b/src/cpp/main_3d/engine_cpu.hpp
index b57271b..228cb91 100644
--- a/src/cpp/main_3d/engine_cpu.hpp
+++ b/src/cpp/main_3d/engine_cpu.hpp
@@ -19,7 +19,9 @@
 
 #include <cstdint>
 
-auto runSim(Simulation3D& sd) -> double;
+namespace pffdtd {
+
+auto run(Simulation3D& sd) -> double;
 double process_bnl_pts_fd(
     Real* u0b,
     Real const* u2b,
@@ -33,3 +35,5 @@ double process_bnl_pts_fd(
     MatQuad const* mat_quads,
     Real const* mat_beta
 );
+
+} // namespace pffdtd
diff --git a/src/cpp/main_3d/engine_cuda.hpp b/src/cpp/main_3d/engine_cuda.hpp
index 4aff488..e579c3e 100644
--- a/src/cpp/main_3d/engine_cuda.hpp
+++ b/src/cpp/main_3d/engine_cuda.hpp
@@ -66,7 +66,7 @@ __constant__ int8_t cuMb[MNm]; //to store Mb per mat (MNm has to be hash-defined
 uint64_t print_gpu_details(int i);
 void check_sorted( Simulation3D const& sd);
 void split_data( Simulation3D const& sd, struct gpuHostData *ghds, int ngpus);
-double runSim(Simulation3D &sd);
+double run(Simulation3D &sd);
 //CUDA kernels
 __global__ void KernelAirCart(Real * __restrict__ u0, const Real * __restrict__ u1, const uint8_t * __restrict__ bn_mask);
 __global__ void KernelAirFCC(Real * __restrict__ u0, const Real * __restrict__ u1, const uint8_t * __restrict__ bn_mask);
@@ -648,7 +648,7 @@ void split_data( Simulation3D const& sd, struct gpuHostData *ghds, int ngpus) {
 }
 
 //run the sim!
-double runSim(Simulation3D &sd)
+double run(Simulation3D &sd)
 {
    //if you want to test synchronous, env variable for that
    const char* s = getenv("CUDA_LAUNCH_BLOCKING");
diff --git a/src/cpp/main_3d/main.cpp b/src/cpp/main_3d/main.cpp
index 8324134..6c2ea61 100644
--- a/src/cpp/main_3d/main.cpp
+++ b/src/cpp/main_3d/main.cpp
@@ -33,10 +33,10 @@
 auto main(int /*argc*/, char** /*argv*/) -> int {
   auto const start = std::chrono::steady_clock::now();
 
-  auto sim = Simulation3D{};
+  auto sim = pffdtd::Simulation3D{};
   loadSimulation3D(sim);
   scaleInput(sim);
-  runSim(sim);
+  run(sim);
   rescaleOutput(sim);
   writeOutputs(sim);
   printLastSample(sim);
diff --git a/src/cpp/pffdtd/hdf.hpp b/src/cpp/pffdtd/hdf.hpp
index 17a81d1..5c11c1f 100644
--- a/src/cpp/pffdtd/hdf.hpp
+++ b/src/cpp/pffdtd/hdf.hpp
@@ -10,6 +10,14 @@
 
 namespace pffdtd {
 
+enum DataType : unsigned char {
+  FLOAT64,
+  FLOAT32,
+  INT64,
+  INT8,
+  BOOL,
+};
+
 template<typename T>
 inline constexpr auto isStdVector = false;
 
diff --git a/src/cpp/pffdtd/simulation_3d.cpp b/src/cpp/pffdtd/simulation_3d.cpp
index 41ae847..208bf07 100644
--- a/src/cpp/pffdtd/simulation_3d.cpp
+++ b/src/cpp/pffdtd/simulation_3d.cpp
@@ -15,17 +15,14 @@
 
 #include "simulation_3d.hpp"
 
-#include "pffdtd/hdf.hpp"
-
 #include <algorithm>
+#include <cassert>
 #include <cmath>
 #include <cstdio>
 #include <filesystem>
 
 namespace {
 
-#include <cassert>
-
 // linear indices to sub-indices in 3d, Nz continguous
 void ind2sub3d(
     int64_t idx,
@@ -62,6 +59,8 @@ void check_inside_grid(
 }
 } // namespace
 
+namespace pffdtd {
+
 // load the sim data from Python-written HDF5 files
 void loadSimulation3D(Simulation3D& sim) {
   // local values, to read in and attach to struct at end
@@ -970,3 +969,5 @@ void writeOutputs(Simulation3D& sim) {
   writer.write("u_out", std::span{u_out}, Nr, Nt);
   std::puts("wrote output dataset");
 }
+
+} // namespace pffdtd
diff --git a/src/cpp/pffdtd/simulation_3d.hpp b/src/cpp/pffdtd/simulation_3d.hpp
index 8afa86f..9386727 100644
--- a/src/cpp/pffdtd/simulation_3d.hpp
+++ b/src/cpp/pffdtd/simulation_3d.hpp
@@ -16,10 +16,9 @@
 #pragma once
 
 #include "pffdtd/config.hpp"
+#include "pffdtd/hdf.hpp"
 #include "pffdtd/utility.hpp"
 
-#include "hdf5.h"
-
 #include <cstdint>
 
 // maximum number of RLC branches in freq-dep (FD) boundaries (needed at
@@ -28,6 +27,8 @@
 // maximum number of materials allows (needed at compile-time for CUDA)
 #define MNm 64 // change as necssary
 
+namespace pffdtd {
+
 // main sim data, on host
 struct Simulation3D {
   int64_t* bn_ixyz;     // boundary node indices
@@ -93,3 +94,5 @@ void readH5Dataset(
     DataType t
 );
 void readH5Constant(hid_t file, char* dset_str, void* out, DataType t);
+
+} // namespace pffdtd
diff --git a/src/cpp/pffdtd/utility.hpp b/src/cpp/pffdtd/utility.hpp
index 22a373d..431f3d5 100644
--- a/src/cpp/pffdtd/utility.hpp
+++ b/src/cpp/pffdtd/utility.hpp
@@ -33,14 +33,6 @@
 #define SET_BIT_VAL(var, pos, val)                                               \
   ((var) = ((var) & ~(1ULL << (pos))) | ((val) << (pos)))
 
-enum DataType : unsigned char {
-  FLOAT64,
-  FLOAT32,
-  INT64,
-  INT8,
-  BOOL,
-};
-
 void allocate_zeros(void** arr, uint64_t Nbytes);
 int cmpfunc_int64_keys(void const* a, void const* b);
 void sort_keys(int64_t* val_arr, int64_t* key_arr, int64_t N);
diff --git a/src/cpp/pffdtd/video.cpp b/src/cpp/pffdtd/video.cpp
index fd9ebfa..a2ac387 100644
--- a/src/cpp/pffdtd/video.cpp
+++ b/src/cpp/pffdtd/video.cpp
@@ -1,5 +1,7 @@
 #include "video.hpp"
 
+#include "pffdtd/exception.hpp"
+
 #include <opencv2/imgproc.hpp>
 
 namespace pffdtd {
diff --git a/src/cpp/pffdtd/video.hpp b/src/cpp/pffdtd/video.hpp
index 33aa729..52f871a 100644
--- a/src/cpp/pffdtd/video.hpp
+++ b/src/cpp/pffdtd/video.hpp
@@ -1,11 +1,8 @@
 #pragma once
 
-#include "pffdtd/exception.hpp"
-
 #include <opencv2/videoio.hpp>
 
 #include <filesystem>
-#include <span>
 
 namespace pffdtd {