Generate basic example in plain C++

An attempt to produce a minimal LBM implementation to benchmark various
memory and vectorization schemes on the CPU.

2 files changed, 141 insertions, 0 deletions

diff --git a/standalone_cpp_codegen.py b/standalone_cpp_codegen.py
new file mode 100644
index 0000000..b38898a
--- /dev/null
+++ b/standalone_cpp_codegen.py
@@ -0,0 +1,32 @@
+import sympy
+from mako.template import Template
+from pathlib import Path
+
+from simulation         import Geometry
+from symbolic.generator import LBM
+
+import symbolic.D2Q9 as D2Q9
+
+lbm = LBM(D2Q9)
+
+moments = lbm.moments(optimize = False)
+collide = lbm.bgk(f_eq = lbm.equilibrium(), tau = 0.6)
+
+geometry = Geometry(512, 512)
+
+program_src = Template(filename = str(Path(__file__).parent/'template/standalone.mako')).render(
+    descriptor = lbm.descriptor,
+    geometry   = geometry,
+
+    steps = 100,
+
+    moments_subexpr    = moments[0],
+    moments_assignment = moments[1],
+    collide_subexpr    = collide[0],
+    collide_assignment = collide[1],
+
+    float_type = 'double',
+    ccode = sympy.ccode
+)
+
+print(program_src)
diff --git a/template/standalone.mako b/template/standalone.mako
new file mode 100644
index 0000000..6750de9
--- /dev/null
+++ b/template/standalone.mako
@@ -0,0 +1,109 @@
+#include <array>
+#include <cstdint>
+#include <memory>
+
+<%
+def pop_offset(i):
+    return i * geometry.volume
+%>
+
+void equilibrilize(${float_type}*    f_next,
+                   ${float_type}*    f_prev,
+                   const std::size_t gid)
+{
+    ${float_type}* preshifted_f_next = f_next + gid;
+    ${float_type}* preshifted_f_prev = f_prev + gid;
+
+% for i, w_i in enumerate(descriptor.w):
+    preshifted_f_next[${pop_offset(i)}] = ${w_i}.f;
+    preshifted_f_prev[${pop_offset(i)}] = ${w_i}.f;
+% endfor
+}
+
+<%
+def neighbor_offset(c_i):
+    return {
+        2: lambda:                                          c_i[1]*geometry.size_x + c_i[0],
+        3: lambda: c_i[2]*geometry.size_x*geometry.size_y + c_i[1]*geometry.size_x + c_i[0]
+    }.get(descriptor.d)()
+
+%>
+
+void collide_and_stream(      ${float_type}* f_next,
+                        const ${float_type}* f_prev,
+                        const int*           material,
+                        const std::size_t    gid)
+{
+    const int m = material[gid];
+
+    if ( m == 0 ) {
+        return;
+    }
+
+          ${float_type}* preshifted_f_next = f_next + gid;
+    const ${float_type}* preshifted_f_prev = f_prev + gid;
+
+% for i, c_i in enumerate(descriptor.c):
+    const ${float_type} f_curr_${i} = preshifted_f_prev[${pop_offset(i) + neighbor_offset(-c_i)}];
+% endfor
+
+% for i, expr in enumerate(moments_subexpr):
+    const ${float_type} ${expr[0]} = ${ccode(expr[1])};
+% endfor
+
+% for i, expr in enumerate(moments_assignment):
+    ${float_type} ${ccode(expr)}
+% endfor
+
+% for i, expr in enumerate(collide_subexpr):
+    const ${float_type} ${expr[0]} = ${ccode(expr[1])};
+% endfor
+
+% for i, expr in enumerate(collide_assignment):
+    const ${float_type} ${ccode(expr)}
+% endfor
+
+% for i in range(0,descriptor.q):
+    preshifted_f_next[${pop_offset(i)}] = f_next_${i};
+% endfor
+}
+
+int main()
+{
+    auto f_a = std::make_unique<${float_type}[]>(${geometry.volume*descriptor.q});
+    auto f_b = std::make_unique<${float_type}[]>(${geometry.volume*descriptor.q});
+    auto material = std::make_unique<int[]>(${geometry.volume});
+
+    ${float_type}* f_prev = f_a.get();
+    ${float_type}* f_next = f_b.get();
+
+    for (int iX = 0; iX < ${geometry.size_x}; ++iX) {
+        for (int iY = 0; iY < ${geometry.size_x}; ++iY) {
+            if (iX == 0 || iY == 0 || iX == ${geometry.size_x-1} || iY == ${geometry.size_y-1}) {
+                material[iY*${geometry.size_x} + iX] = 0;
+            } else {
+                material[iY*${geometry.size_x} + iX] = 1;
+            }
+        }
+    }
+
+    for (std::size_t iCell = 0; iCell < ${geometry.volume}; ++iCell) {
+        equilibrilize(f_prev, f_next, iCell);
+    }
+
+    for (std::size_t iStep = 0; iStep < ${steps}; ++iStep) {
+        if (iStep % 2 == 0) {
+            f_next = f_a.get();
+            f_prev = f_b.get();
+        } else {
+            f_next = f_b.get();
+            f_prev = f_a.get();
+        }
+
+        for (std::size_t iCell = 0; iCell < ${geometry.volume}; ++iCell) {
+            collide_and_stream(f_next, f_prev, material.get(), iCell);
+        }
+    }
+
+    return 0;
+}