summaryrefslogtreecommitdiff
path: root/src
diff options
context:
space:
mode:
authorVito Caputo <vcaputo@pengaru.com>2019-10-14 14:58:37 -0700
committerVito Caputo <vcaputo@pengaru.com>2019-10-14 16:42:26 -0700
commit96a68810d21eeabf24985d9e07e897c6cb3aedd5 (patch)
treef59f39e39c260ff1d30619554e816c553fefc444 /src
parenta21f1febe29fbc2c7e00ff97913c04c247bfb95b (diff)
modules/flui2d: add 2D fluid dynamics simulation
This implements near verbatim the code found in the paper titled: Real-Time Fluid Dynamics for Games By Jos Stam It sometimes has the filename GDC03.PDF, or Stam_fluids_GDC03.pdf The density field is rendered using simple linear interpolation of the samples, in a grayscale palette. No gamma correction is being performed. There are three configurable defines of interest: VISCOSITY, DIFFUSION, and ROOT. This module is only threaded in the drawing stage, so basically the linear interpolation uses multiple cores. The simulation itself is not threaded, the implementation from the paper made no such considerations. It would be nice to reimplement this in a threaded fashion with a good generalized API, then move it into libs. Something where a unit square can be sampled for interpolated densities would be nice. Then extend it into 3 dimensions for volumetric effects...
Diffstat (limited to 'src')
-rw-r--r--src/Makefile.am2
-rw-r--r--src/modules/Makefile.am2
-rw-r--r--src/modules/flui2d/Makefile.am3
-rw-r--r--src/modules/flui2d/flui2d.c291
-rw-r--r--src/rototiller.c2
5 files changed, 298 insertions, 2 deletions
diff --git a/src/Makefile.am b/src/Makefile.am
index 7f0b63f..c1e726f 100644
--- a/src/Makefile.am
+++ b/src/Makefile.am
@@ -4,4 +4,4 @@ rototiller_SOURCES = fb.c fb.h fps.c fps.h rototiller.c rototiller.h sdl_fb.c se
if ENABLE_DRM
rototiller_SOURCES += drm_fb.c
endif
-rototiller_LDADD = modules/julia/libjulia.a modules/plasma/libplasma.a modules/ray/libray.a modules/roto/libroto.a modules/sparkler/libsparkler.a modules/stars/libstars.a modules/submit/libsubmit.a libs/grid/libgrid.a libs/ray/libray.a -lm
+rototiller_LDADD = modules/flui2d/libflui2d.a modules/julia/libjulia.a modules/plasma/libplasma.a modules/ray/libray.a modules/roto/libroto.a modules/sparkler/libsparkler.a modules/stars/libstars.a modules/submit/libsubmit.a libs/grid/libgrid.a libs/ray/libray.a -lm
diff --git a/src/modules/Makefile.am b/src/modules/Makefile.am
index 8706332..de59b4a 100644
--- a/src/modules/Makefile.am
+++ b/src/modules/Makefile.am
@@ -1 +1 @@
-SUBDIRS = julia plasma ray roto sparkler stars submit
+SUBDIRS = flui2d julia plasma ray roto sparkler stars submit
diff --git a/src/modules/flui2d/Makefile.am b/src/modules/flui2d/Makefile.am
new file mode 100644
index 0000000..239de9e
--- /dev/null
+++ b/src/modules/flui2d/Makefile.am
@@ -0,0 +1,3 @@
+noinst_LIBRARIES = libflui2d.a
+libflui2d_a_SOURCES = flui2d.c
+libflui2d_a_CPPFLAGS = -I@top_srcdir@/src
diff --git a/src/modules/flui2d/flui2d.c b/src/modules/flui2d/flui2d.c
new file mode 100644
index 0000000..2f0a472
--- /dev/null
+++ b/src/modules/flui2d/flui2d.c
@@ -0,0 +1,291 @@
+#include <stdint.h>
+#include <inttypes.h>
+#include <math.h>
+#include <stdlib.h>
+
+#include "fb.h"
+#include "rototiller.h"
+
+
+/* This code is almost entirely taken from the paper:
+ * Real-Time Fluid Dynamics for Games
+ * Joe Stam - Alias | Wavefront
+ *
+ * I take zero credit for it, I only wrote the rototiller integration.
+ * - Vito Caputo <vcaputo@pengaru.com> 10/13/2019
+ */
+
+#if 1
+ /* These knobs affect how the simulated fluid behaves */
+#define VISCOSITY .000000001f
+#define DIFFUSION .00001f
+#else
+#define VISCOSITY .00001f
+#define DIFFUSION .000001f
+#endif
+
+#define ROOT 128 // Change this to vary the density field resolution
+#define SIZE ((ROOT + 2) * (ROOT + 2))
+#define IX(i, j) ((i) + (ROOT + 2) * (j))
+#define SWAP(x0, x) {float *tmp = x0; x0 = x; x = tmp;}
+
+typedef struct flui2d_t {
+ float u[SIZE], v[SIZE], u_prev[SIZE], v_prev[SIZE];
+ float dens[SIZE], dens_prev[SIZE];
+ float visc, diff;
+} flui2d_t;
+
+static void set_bnd(int N, int b, float *x)
+{
+ for (int i = 1; i <= N; i++) {
+ x[IX(0, i)] = b == 1 ? -x[IX(1, i)] : x[IX(1, i)];
+ x[IX(N + 1, i)] = b == 1 ? -x[IX(N, i)] : x[IX(N, i)];
+ x[IX(i, 0)] = b == 2 ? -x[IX(i, 1)] : x[IX(i, 1)];
+ x[IX(i, N + 1)] = b == 2 ? -x[IX(i, N)] : x[IX(i, N)];
+ }
+
+ x[IX(0 , 0)] = 0.5 * (x[IX(1, 0)] + x[IX(0, 1)]);
+ x[IX(0 , N + 1)] = 0.5 * (x[IX(1, N + 1)] + x[IX(0, N)]);
+ x[IX(N + 1, 0)] = 0.5 * (x[IX(N, 0)] + x[IX(N + 1, 1)]);
+ x[IX(N + 1, N + 1)] = 0.5 * (x[IX(N, N + 1)] + x[IX(N + 1, N)]);
+}
+
+static void add_source(int N, float *x, float *s, float dt)
+{
+ int size = (N + 2) * (N + 2);
+
+ for (int i = 0; i < size; i++)
+ x[i] += dt * s[i];
+}
+
+static void diffuse(int N, int b, float *x, float *x0, float diff, float dt)
+{
+ float a = dt * diff * (float)N * (float)N;
+ int i, j, k;
+ float z = 1.f / (1.f + 4.f * a);
+
+ for (k = 0; k < 20; k++) {
+ for (i = 1; i <= N; i++) {
+ for (j = 1; j <= N; j++) {
+ x[IX(i, j)] = (x0[IX(i, j)] + a * (x[IX(i - 1, j)] + x[IX(i + 1, j)] + x[IX(i, j - 1)] + x[IX(i, j + 1)])) * z;
+ }
+ }
+ set_bnd(N, b, x);
+ }
+}
+
+static void advect(int N, int b, float *d, float *d0, float *u, float *v, float dt)
+{
+ float x, y, s0, t0, s1, t1, dt0;
+ int i, j, i0, j0, i1, j1;
+
+ dt0 = dt * (float)N;
+ for (i = 1 ; i <= N ; i++) {
+ for (j = 1 ; j <= N; j++) {
+ x = (float)i - dt0 * u[IX(i, j)];
+ y = (float)j - dt0 * v[IX(i, j)];
+
+ if (x < .5f)
+ x = .5f;
+ if (x > (float)N + .5f)
+ x = (float)N + .5f;
+
+ i0 = (int)x;
+ i1 = i0 + 1;
+
+ if (y < .5f)
+ y = .5f;
+ if (y > (float)N + .5f)
+ y = (float)N + .5f;
+
+ j0 = (int)y;
+ j1 = j0 + 1;
+
+ s1 = x - (float)i0;
+ s0 = 1.f - s1;
+ t1 = y - (float)j0;
+ t0 = 1.f - t1;
+
+ d[IX(i, j)] = s0 * (t0 * d0[IX(i0, j0)] + t1 * d0[IX(i0, j1)]) + s1 * (t0 * d0[IX(i1, j0)] + t1 * d0[IX(i1, j1)]);
+ }
+ }
+ set_bnd(N, b, d);
+}
+
+static void project(int N, float *u, float *v, float *p, float *div)
+{
+ float h = 1.f / (float)N;
+ int i, j, k;
+
+ for (i = 1; i <= N ; i++) {
+ for (j = 1; j <= N; j++) {
+ div[IX(i, j)] = -0.5 * h *(u[IX(i + 1, j)] - u[IX(i - 1, j)] + v[IX(i, j + 1)] - v[IX(i, j - 1)]);
+ p[IX(i, j)] = 0;
+ }
+ }
+
+ set_bnd(N, 0, div);
+ set_bnd(N, 0, p);
+
+ for (k = 0; k < 20; k++) {
+ for (i = 1; i <= N; i++) {
+ for (j = 1; j <= N; j++) {
+ p[IX(i, j)] = (div[IX(i, j)] + p[IX(i - 1, j)] + p[IX(i + 1, j)] + p[IX(i, j - 1)] + p[IX(i, j + 1)]) * .25f;
+ }
+ }
+ set_bnd(N, 0, p);
+ }
+
+ for (i = 1; i <= N; i++) {
+ for (j = 1; j <= N; j++) {
+ u[IX(i, j)] -= 0.5 * (p[IX(i + 1, j)] - p[IX(i - 1, j)]) / h;
+ v[IX(i, j)] -= 0.5 * (p[IX(i, j + 1)] - p[IX(i, j - 1)]) / h;
+ }
+ }
+
+ set_bnd(N, 1, u);
+ set_bnd(N, 2, v);
+}
+
+static void dens_step(int N, float *x, float *x0, float *u, float *v, float diff, float dt)
+{
+
+ /*
+ * The paper includes this, but it blows up the simulation.
+ * add_source(N, x, x0, dt);
+ * SWAP(x0, x);
+ */
+ diffuse(N, 0, x, x0, diff, dt);
+ SWAP(x0, x);
+ advect(N, 0, x, x0, u, v, dt);
+}
+
+static void vel_step(int N, float *u, float *v, float *u0, float *v0, float visc, float dt)
+{
+ add_source(N, u, u0, dt);
+ add_source(N, v, v0, dt);
+ SWAP(u0, u);
+ diffuse(N, 1, u, u0, visc, dt);
+ SWAP(v0, v);
+ diffuse(N, 2, v, v0, visc, dt);
+ project(N, u, v, u0, v0);
+ SWAP(u0, u);
+ SWAP(v0, v);
+ advect(N, 1, u, u0, u0, v0, dt);
+ advect(N, 2, v, v0, u0, v0, dt);
+ project(N, u, v, u0, v0);
+}
+
+
+typedef struct flui2d_context_t {
+ flui2d_t fluid;
+ float xf, yf;
+} flui2d_context_t;
+
+
+static void * flui2d_create_context(void)
+{
+ flui2d_context_t *ctxt;
+
+ ctxt = calloc(1, sizeof(flui2d_context_t));
+ if (!ctxt)
+ return NULL;
+
+ ctxt->fluid.visc = VISCOSITY;
+ ctxt->fluid.diff = DIFFUSION;
+
+ return ctxt;
+}
+
+
+static void flui2d_destroy_context(void *context)
+{
+ free(context);
+}
+
+
+static int flui2d_fragmenter(void *context, const fb_fragment_t *fragment, unsigned num, fb_fragment_t *res_fragment)
+{
+ return fb_fragment_tile_single(fragment, 64, num, res_fragment);
+}
+
+
+/* Prepare a frame for concurrent drawing of fragment using multiple fragments */
+static void flui2d_prepare_frame(void *context, unsigned n_cpus, fb_fragment_t *fragment, rototiller_fragmenter_t *res_fragmenter)
+{
+ flui2d_context_t *ctxt = context;
+ static float r;
+ int x = (cos(r += .01f) * .4f + .5f) * (float)ROOT; /* figure eight pattern for the added densities */
+ int y = (sin(r * 2.f) * .4f + .5f) * (float)ROOT;
+
+ *res_fragmenter = flui2d_fragmenter;
+
+ ctxt->fluid.dens_prev[IX(x, y)] = 1.f;
+
+ /* This orientation for the added velocities at the added densities isn't trying to
+ * emulate any sort of physical relationship to the movement - it's just creating a variety
+ * of turbulence. It'd be trivial to make it look like a rocket's jetstream or something.
+ */
+ ctxt->fluid.u_prev[IX(x, y)] = cos(r * 3.f) * 10.f;
+ ctxt->fluid.v_prev[IX(x, y)] = sin(r * 3.f) * 10.f;
+
+ /* These are the core of the simulation, and can't currently be threaded using the paper's implementation, so they
+ * must occur serialized here in prepare_frame. It would be interesting to try refactor the API and tweak the
+ * implementation for threading, as it would really open up larger field sizes as well as map more naturally to
+ * a GLSL implementation for a fragment shader.
+ */
+ vel_step(ROOT, ctxt->fluid.u, ctxt->fluid.v, ctxt->fluid.u_prev, ctxt->fluid.v_prev, ctxt->fluid.visc, .1f);
+ dens_step(ROOT, ctxt->fluid.dens, ctxt->fluid.dens_prev, ctxt->fluid.u, ctxt->fluid.v, ctxt->fluid.diff, .1f);
+
+ ctxt->xf = 1.f / fragment->frame_width;
+ ctxt->yf = 1.f / fragment->frame_height;
+}
+
+
+/* Draw a the flui2d densities */
+static void flui2d_render_fragment(void *context, fb_fragment_t *fragment)
+{
+ flui2d_context_t *ctxt = context;
+
+ for (int y = fragment->y; y < fragment->y + fragment->height; y++) {
+ int y0, y1;
+ float Y;
+
+ Y = (float)y * ctxt->yf * (float)ROOT;
+ y0 = (int)Y;
+ y1 = y0 + 1;
+
+ for (int x = fragment->x; x < fragment->x + fragment->width; x++) {
+ float X, dens, dx0, dx1;
+ int x0, x1;
+ uint32_t pixel;
+
+ X = (float)x * ctxt->xf * (float)ROOT;
+ x0 = (int)X;
+ x1 = x0 + 1;
+
+ /* linear interpolation of density samples */
+ dx0 = ctxt->fluid.dens[(int)IX(x0, y0)] * (1.f - (X - x0));
+ dx0 += ctxt->fluid.dens[(int)IX(x1, y0)] * (X - x0);
+ dx1 = ctxt->fluid.dens[(int)IX(x0, y1)] * (1.f - (X - x0));
+ dx1 += ctxt->fluid.dens[(int)IX(x1, y1)] * (X - x0);
+ dens = dx0 * (1.f - (Y - y0)) + dx1 * (Y - y0);
+
+ pixel = ((float)dens * 256.f);
+ pixel = pixel << 16 | pixel << 8 | pixel;
+ fb_fragment_put_pixel_unchecked(fragment, x, y, pixel);
+ }
+ }
+}
+
+
+rototiller_module_t flui2d_module = {
+ .create_context = flui2d_create_context,
+ .destroy_context = flui2d_destroy_context,
+ .prepare_frame = flui2d_prepare_frame,
+ .render_fragment = flui2d_render_fragment,
+ .name = "flui2d",
+ .description = "Fluid dynamics simulation in 2D (threaded (poorly))",
+ .author = "Vito Caputo <vcaputo@pengaru.com>",
+ .license = "Unknown",
+};
diff --git a/src/rototiller.c b/src/rototiller.c
index 2a9a469..fe0c154 100644
--- a/src/rototiller.c
+++ b/src/rototiller.c
@@ -31,6 +31,7 @@ extern fb_ops_t drm_fb_ops;
extern fb_ops_t sdl_fb_ops;
fb_ops_t *fb_ops;
+extern rototiller_module_t flui2d_module;
extern rototiller_module_t julia_module;
extern rototiller_module_t plasma_module;
extern rototiller_module_t roto32_module;
@@ -51,6 +52,7 @@ static rototiller_module_t *modules[] = {
&julia_module,
&submit_module,
&submit_softly_module,
+ &flui2d_module,
};
© All Rights Reserved