diff --git a/src/particle_oriented_animated_quad_vbuf_pixel_shader.cpp b/src/particle_oriented_animated_quad_vbuf_pixel_shader.cpp
index f2d812f..cc3030e 100644
--- a/src/particle_oriented_animated_quad_vbuf_pixel_shader.cpp
+++ b/src/particle_oriented_animated_quad_vbuf_pixel_shader.cpp
@@ -49,7 +49,7 @@ const int vertex_shader_paths_length = (sizeof (vertex_shader_paths)) / (sizeof
 const char * fragment_shader_paths[] = {
   "clear.fs.bin",
   "particle_plane.fs.bin",
-  "particle_particle.fs.bin",
+  "particle_particle_colored.fs.bin",
   "texture_tile.fs.bin",
   "particle_physics.fs.bin",
   "vertex_buffer_copy.fs.bin",
@@ -965,7 +965,7 @@ void init_particles2(void * position_ptr,
     float vx = xorshift32f(state) * 2.0f - 1.0f;
     float vz = xorshift32f(state) * 2.0f - 1.0f;
 
-    float new_age = max_age * sinf(fi * rl * 2) * 0.5f + 0.5f;
+    float new_age = max_age * sinf(fi * rl * 2.0f) * 0.5f + 0.5f;
 
     vec3 new_position = vec3(sx, sy, sz);
     vec3 new_velocity = normalize(vec3(vx * 0.5f, 0.0f, vz * 0.5f));
@@ -1304,6 +1304,7 @@ int main()
   // colorbuffer
   colorbuffer_handle[0] = create_buffer(fd, colorbuffer_size, &colorbuffer_ptr[0]);
   colorbuffer_handle[1] = create_buffer(fd, colorbuffer_size, &colorbuffer_ptr[1]);
+
   zbuffer_handle = create_buffer(fd, colorbuffer_size, &zbuffer_ptr);
   //flush_handle = create_flush_buffer(fd);
   texturebuffer_handle = load_textures(fd, textures, textures_length);
@@ -1315,7 +1316,7 @@ int main()
   int colorbuffer_ix = 0;
   float theta = PI * 0.5;
 
-  const int floatbuffer_width = 512;
+  const int floatbuffer_width = 256;
   const int floatbuffer_height = 256;
   floatbuffer_state state = create_floatbuffers(fd, floatbuffer_width * floatbuffer_height);
 
@@ -1374,11 +1375,13 @@ int main()
         break;
 
 
-      struct timespec duration = {
-        .tv_sec = 0,
-        .tv_nsec = 18000000,
-      };
-      nanosleep(&duration, NULL);
+      if (0) {
+        struct timespec duration = {
+          .tv_sec = 0,
+          .tv_nsec = 17000000,
+        };
+        nanosleep(&duration, NULL);
+      }
     }
 
     primary_surface_address(rmmio, colorbuffer_ix);
diff --git a/src/particle_particle_animated_quad_vbuf.vs.asm b/src/particle_particle_animated_quad_vbuf.vs.asm
index 2c08251..26c6464 100644
--- a/src/particle_particle_animated_quad_vbuf.vs.asm
+++ b/src/particle_particle_animated_quad_vbuf.vs.asm
@@ -21,18 +21,18 @@ temp[2].xy   = VE_MAD  input[1].yx__  const[7].xx__  input[1].11__ ;
 temp[0].xyz  = VE_ADD  input[0].xyz_  input[0].000_ ;
 
 -- ppos = position.xxx * dx + ppos
-temp[0].xyz  = VE_MAD   temp[2].xxx_  const[4].xyz_  temp[0].xyz_ ;
+temp[0].xyz  = VE_MAD   temp[2].xxx_  const[4].xyz_   temp[0].xyz_ ;
 -- ppos = position.yyy * dy + ppos
-temp[0].xyz  = VE_MAD   temp[2].yyy_  const[5].xyz_  temp[0].xyz_ ;
+temp[0].xyz  = VE_MAD   temp[2].yyy_  const[5].xyz_   temp[0].xyz_ ;
 
--- ppos *= scale
-temp[0].xyzw = VE_MUL  temp[0].xyz1   const[6].www1 ;
+-- ppos = vec4(ppos.xyz * scale.xyz, age)
+temp[0].xyzw = VE_MAD  temp[0].xyz0   const[6].www0  input[0].000w ;
 
 -- ppos = transform_matrix * ppos
-temp[1].x    = VE_DOT  const[0].xyzw  temp[0].xyzw ;
-temp[1].y    = VE_DOT  const[1].xyzw  temp[0].xyzw ;
-temp[1].z    = VE_DOT  const[2].xyzw  temp[0].xyzw ;
-temp[1].w    = VE_DOT  const[3].xyzw  temp[0].xyzw ;
+temp[1].x    = VE_DOT  const[0].xyzw  temp[0].xyz1 ;
+temp[1].y    = VE_DOT  const[1].xyzw  temp[0].xyz1 ;
+temp[1].z    = VE_DOT  const[2].xyzw  temp[0].xyz1 ;
+temp[1].w    = VE_DOT  const[3].xyzw  temp[0].xyz1 ;
 
-out[0].xyzw  = VE_MAD   temp[1].xyzw   temp[1].1111  temp[1].0000 ;
-out[1].xyzw  = VE_MAX  input[1].xy00  input[1].xy00 ;
+out[0].xyzw  = VE_ADD   temp[1].xyzw   temp[1].0000 ;
+out[1].xyzw  = VE_ADD  input[1].xy00   temp[0].00zw ; -- age
diff --git a/src/particle_particle_animated_quad_vbuf.vs.bin b/src/particle_particle_animated_quad_vbuf.vs.bin
index 450568f..efef34e 100644
Binary files a/src/particle_particle_animated_quad_vbuf.vs.bin and b/src/particle_particle_animated_quad_vbuf.vs.bin differ
diff --git a/src/particle_particle_colored.fs.asm b/src/particle_particle_colored.fs.asm
new file mode 100644
index 0000000..92323ae
--- /dev/null
+++ b/src/particle_particle_colored.fs.asm
@@ -0,0 +1,43 @@
+-- temp[0].xy: texture coordinate
+-- temp[0].a : age
+
+TEX TEX_SEM_WAIT TEX_SEM_ACQUIRE
+  temp[2].rgb = LD tex[0].rgba temp[0].rgaa ;
+
+-- i = vec3(0.25, 0.40625, 0.5625);
+src0.a = float(40)   , -- 0.25
+src1.a = float(45)   , -- 0.40625
+src2.rgb = float(49) : -- 0.5625
+  temp[1].rgb = MAD src0.a10 src1.1a0 src2.00r ;
+
+-- t = i + vec3(td)
+src0.a   = temp[0] ,
+src0.rgb = temp[1] ,
+src1.rgb = temp[0] :
+  temp[1].rgb = MAD src1.bbb src0.aaa src0.rgb ;
+
+-- j = fract(t)
+src0.rgb = temp[1] :
+  temp[1].rgb = FRC src0.rgb ;
+
+-- k = cos(j * 2π)
+src0.rgb = temp[1] :
+              COS src0.r ,
+  temp[1].r = SOP ;
+src0.rgb = temp[1] :
+              COS src0.g ,
+  temp[1].g = SOP ;
+src0.rgb = temp[1] :
+              COS src0.b ,
+  temp[1].b = SOP ;
+
+-- l = k * vec3(0.5, 0.5, 0.5) + vec3(0.5, 0.5, 0.5)
+src0.rgb = temp[1] ,
+src1.rgb = float(48) : -- 0.5
+  temp[1].rgb = MAD src0.rgb src1.rrr src1.rrr ;
+
+OUT TEX_SEM_WAIT
+src0.rgb = temp[2] ,
+src1.rgb = temp[1] :
+  out[0].a    = MAX src0.1 src0.1 ,
+  out[0].rgb  = MAD src0.rgb src1.rgb src0.000 ;
diff --git a/src/particle_particle_colored.fs.bin b/src/particle_particle_colored.fs.bin
new file mode 100644
index 0000000..3eb1524
Binary files /dev/null and b/src/particle_particle_colored.fs.bin differ
diff --git a/src/particle_physics.fs.asm b/src/particle_physics.fs.asm
index c570f3f..660098f 100644
--- a/src/particle_physics.fs.asm
+++ b/src/particle_physics.fs.asm
@@ -75,8 +75,9 @@ src0.rgb = temp[1] : -- velocity
 src0.a   = temp[2] :
   temp[2].a   = RSQ |src0.a| ;
 src0.a   = temp[2] ,
-src0.rgb = temp[1] : -- velocity
-  temp[2].rgb = MAD src0.r0b src0.a0a src0.000 ;
+src0.rgb = temp[1] , -- velocity
+src1.a   = temp[1] : -- delta
+  temp[2].rgb = MAD src0.r0b src0.a0a src1.0a0 ;
 
 -- age = age + max_age
 -- reset__position = reset__position * 20
@@ -88,10 +89,19 @@ src0.rgb = temp[2]  : -- reset__position
   temp[2].rgb = MAD src0.rgb src2.aaa src2.000 ;
 
 -- reset__velocity
-src0.a   = temp[1] , -- delta
-src1.a   = float(64) , -- 2.0
-src0.rgb = temp[1] : -- velocity
-  temp[3].rgb = MAD src0.rab src1.1a1 src1.000 ;
+-- (p.velocity.x * 0.5 + 0.5)
+-- velocity.xz = velocity.xz
+src0.rgb = temp[1]   , -- velocity.x
+src1.a   = float(48) : -- 0.5
+  temp[3].a   = MAD src0.r src1.a src1.a ,
+  temp[3].rgb = MAX src0.r0b src0.r0b ;
+
+-- reset__velocity
+-- velocity.y = (temp[3].a * delta + 2.0)
+src0.a   = temp[1]   , -- delta
+src1.a   = float(56) , -- 1.0
+src2.a   = temp[3]   : --
+  temp[3].g = MAD src2.0a0 src0.0a0 src1.0a0 ;
 
 OUT
 src0.a   = temp[4]  , -- update__age
diff --git a/src/particle_physics.fs.bin b/src/particle_physics.fs.bin
index 279fc2b..f2565a9 100644
Binary files a/src/particle_physics.fs.bin and b/src/particle_physics.fs.bin differ
diff --git a/src/vertex_buffer_copy.vs.asm b/src/vertex_buffer_copy.vs.asm
index eadd23e..75b82cf 100644
--- a/src/vertex_buffer_copy.vs.asm
+++ b/src/vertex_buffer_copy.vs.asm
@@ -1,2 +1,2 @@
 out[0].xyzw = VE_ADD  input[0].xyzw  input[0].0000 ;
-out[1].xyzw = VE_ADD  input[0].xy00  const[0].xy00 ;
+out[1].xyzw = VE_ADD  input[0].xy00  input[0].0000 ;
diff --git a/src/vertex_buffer_copy.vs.bin b/src/vertex_buffer_copy.vs.bin
index b140eb3..fc9a4c5 100644
Binary files a/src/vertex_buffer_copy.vs.bin and b/src/vertex_buffer_copy.vs.bin differ