Skip to content

Commit

Permalink
Add screenshots
Browse files Browse the repository at this point in the history
  • Loading branch information
@AdamYuan
AdamYuan committed Apr 13, 2019
1 parent 116fc93 commit 1c767cf
Show file tree
Hide file tree
Showing 5 changed files with 247 additions and 2 deletions.
2 changes: 2 additions & 0 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -35,4 +35,6 @@ make
## Screenshots
![](https://raw.githubusercontent.com/AdamYuan/Adypt/master/screenshots/1.png)
![](https://raw.githubusercontent.com/AdamYuan/Adypt/master/screenshots/2.png)
![](https://raw.githubusercontent.com/AdamYuan/Adypt/master/screenshots/3.png)
![](https://raw.githubusercontent.com/AdamYuan/Adypt/master/screenshots/4.png)

Binary file added screenshots/3.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added screenshots/4.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
4 changes: 4 additions & 0 deletions shaders/pathtracer.glsl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -129,6 +129,7 @@ vec3 Render(vec4 origin, vec3 dir)

if(tri_idx == -1)
{
//if(!BVHIntersection(origin, normalize(vec3(0.6, 1, 0.2))))
ret += color * vec3(uSunR, uSunG, uSunB);
break;
}
Expand All @@ -137,6 +138,9 @@ vec3 Render(vec4 origin, vec3 dir)
origin.xyz = position;
ret += color * emissive;

if(mtl.m_illum < 6 && dot(dir, normal) > 0)
normal = -normal;

switch(mtl.m_illum)
{
case 2: //glossy reflection
Expand Down
243 changes: 241 additions & 2 deletions shaders/traversal.glsl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,10 +9,10 @@ layout(std430, binding = 0) readonly buffer uuBVHNodes { Node uBVHNodes[]; };
layout(std430, binding = 1) readonly buffer uuTriIndices { int uTriIndices[]; };
layout(std430, binding = 2) readonly buffer uuTriMatrices { Woop uTriMatrices[]; };

uvec2 stack[TRAVERSAL_STACK_SIZE];

void BVHIntersection(in const vec4 origin_tmin, vec3 dir, inout int o_hit_tri_idx, inout vec2 o_hit_uv)
{
uvec2 stack[TRAVERSAL_STACK_SIZE];

const float ooeps = exp2(-64.0f);
dir.x = abs(dir.x) > ooeps ? dir.x : (dir.x >= 0 ? ooeps : -ooeps);
dir.y = abs(dir.y) > ooeps ? dir.y : (dir.y >= 0 ? ooeps : -ooeps);
Expand Down Expand Up @@ -253,3 +253,242 @@ void BVHIntersection(in const vec4 origin_tmin, vec3 dir, inout int o_hit_tri_id
if(o_hit_tri_idx != -1)
o_hit_tri_idx = uTriIndices[o_hit_tri_idx];
}

bool BVHIntersection(in const vec4 origin_tmin, vec3 dir)
{
const float ooeps = exp2(-64.0f);
dir.x = abs(dir.x) > ooeps ? dir.x : (dir.x >= 0 ? ooeps : -ooeps);
dir.y = abs(dir.y) > ooeps ? dir.y : (dir.y >= 0 ? ooeps : -ooeps);
dir.z = abs(dir.z) > ooeps ? dir.z : (dir.z >= 0 ? ooeps : -ooeps);
dir = normalize(dir);
vec3 idir = 1.0f / dir;
uint octinv = 7u - ((dir.x < 0 ? 1 : 0) | (dir.y < 0 ? 2 : 0) | (dir.z < 0 ? 4 : 0));
uint octinv4 = octinv * 0x01010101u;

vec3 origin = origin_tmin.xyz;

float hit_tmin = origin_tmin.w;
float hit_t = 1e9f;

//stack
int stack_ptr = 0;

//traversal states
uvec2 tri_group, node_group = uvec2(0u, 0x80000000u);

//aabb intersection variables
float txmin[4], tymin[4], tzmin[4], txmax[4], tymax[4], tzmax[4], ctmin, ctmax;

//triangle intersection variables
vec4 tv00, tv11, tv22;
float tdx, tdy, tidz, tox, toy, toz;
float tt, tu, tv;

while(true)
{
if(node_group.y > 0x00ffffffu)
{
// G represents a node group
// n <- GetClosestNode(G, r)
uint imask = node_group.y;
uint child_bit_index = findMSB(node_group.y);
uint child_node_base_index = node_group.x;

// Clear corresponding bit in hits field
// G <- G / n
node_group.y &= ~(1u << child_bit_index);

if(node_group.y > 0x00ffffffu)
stack[stack_ptr ++] = node_group;

// Intersect with n
// G, Gt <- IntersectChildren(n, r)
uint slot_index = (child_bit_index - 24) ^ octinv;
uint relative_index = bitCount(imask & ~(0xffffffffu << slot_index));
uint child_node_index = child_node_base_index + relative_index;

//read node data
vec4 head = uBVHNodes[child_node_index].m_head;
uint head_w = floatBitsToUint(head.w);
uvec4 base_meta = uBVHNodes[child_node_index].m_base_meta;
uvec4 lox_loy = uBVHNodes[child_node_index].m_lox_loy;
uvec4 loz_hix = uBVHNodes[child_node_index].m_loz_hix;
uvec4 hiy_hiz = uBVHNodes[child_node_index].m_hiy_hiz;

float adjusted_idir_x = uintBitsToFloat(((head_w ) & 0xffu) << 23u) * idir.x;
float adjusted_idir_y = uintBitsToFloat(((head_w >> 8u) & 0xffu) << 23u) * idir.y;
float adjusted_idir_z = uintBitsToFloat(((head_w >> 16u) & 0xffu) << 23u) * idir.z;
vec3 adjusted_origin = (head.xyz - origin) * idir;

node_group.x = base_meta.x;
tri_group.x = base_meta.y;
tri_group.y = 0u;

uint hitmask = 0u;
{
uint meta4 = base_meta.z;
uint is_inner4 = (meta4 & (meta4 << 1u)) & 0x10101010u;
uint bit_index4 = (meta4 ^ (octinv4 & (is_inner4>>4u)*0xffu)) & 0x1f1f1f1fu;
uint child_bits4 = (meta4 >> 5u) & 0x07070707u;

uint swizzled_lox = (idir.x < 0) ? loz_hix.z : lox_loy.x;
uint swizzled_hix = (idir.x < 0) ? lox_loy.x : loz_hix.z;

uint swizzled_loy = (idir.y < 0) ? hiy_hiz.x : lox_loy.z;
uint swizzled_hiy = (idir.y < 0) ? lox_loy.z : hiy_hiz.x;

uint swizzled_loz = (idir.z < 0) ? hiy_hiz.z : loz_hix.x;
uint swizzled_hiz = (idir.z < 0) ? loz_hix.x : hiy_hiz.z;

txmin[0] = float((swizzled_lox ) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmin[1] = float((swizzled_lox >> 8) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmin[2] = float((swizzled_lox >> 16) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmin[3] = float((swizzled_lox >> 24) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
tymin[0] = float((swizzled_loy ) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymin[1] = float((swizzled_loy >> 8) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymin[2] = float((swizzled_loy >> 16) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymin[3] = float((swizzled_loy >> 24) & 0xffu) * adjusted_idir_y + adjusted_origin.y;

tzmin[0] = float((swizzled_loz ) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmin[1] = float((swizzled_loz >> 8) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmin[2] = float((swizzled_loz >> 16) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmin[3] = float((swizzled_loz >> 24) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
txmax[0] = float((swizzled_hix ) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmax[1] = float((swizzled_hix >> 8) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmax[2] = float((swizzled_hix >> 16) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmax[3] = float((swizzled_hix >> 24) & 0xffu) * adjusted_idir_x + adjusted_origin.x;

tymax[0] = float((swizzled_hiy ) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymax[1] = float((swizzled_hiy >> 8) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymax[2] = float((swizzled_hiy >> 16) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymax[3] = float((swizzled_hiy >> 24) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tzmax[0] = float((swizzled_hiz ) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmax[1] = float((swizzled_hiz >> 8) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmax[2] = float((swizzled_hiz >> 16) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmax[3] = float((swizzled_hiz >> 24) & 0xffu) * adjusted_idir_z + adjusted_origin.z;

ctmin = max(max(txmin[0], tymin[0]), max(tzmin[0], hit_tmin));
ctmax = min(min(txmax[0], tymax[0]), min(tzmax[0], hit_t));
if(ctmin <= ctmax) hitmask |= ((child_bits4 ) & 0xffu) << ((bit_index4 ) & 0xffu);

ctmin = max(max(txmin[1], tymin[1]), max(tzmin[1], hit_tmin));
ctmax = min(min(txmax[1], tymax[1]), min(tzmax[1], hit_t));
if(ctmin <= ctmax) hitmask |= ((child_bits4 >> 8u) & 0xffu) << ((bit_index4 >> 8u) & 0xffu);

ctmin = max(max(txmin[2], tymin[2]), max(tzmin[2], hit_tmin));
ctmax = min(min(txmax[2], tymax[2]), min(tzmax[2], hit_t));
if(ctmin <= ctmax) hitmask |= ((child_bits4 >> 16u) & 0xffu) << ((bit_index4 >> 16u) & 0xffu);

ctmin = max(max(txmin[3], tymin[3]), max(tzmin[3], hit_tmin));
ctmax = min(min(txmax[3], tymax[3]), min(tzmax[3], hit_t));
if(ctmin <= ctmax) hitmask |= ((child_bits4 >> 24u) & 0xffu) << ((bit_index4 >> 24u) & 0xffu);
}

{
uint meta4 = base_meta.w;
uint is_inner4 = (meta4 & (meta4 << 1u)) & 0x10101010u;
uint bit_index4 = (meta4 ^ (octinv4 & (is_inner4>>4u)*0xffu)) & 0x1f1f1f1fu;
uint child_bits4 = (meta4 >> 5u) & 0x07070707u;

uint swizzled_lox = (idir.x < 0) ? loz_hix.w : lox_loy.y;
uint swizzled_hix = (idir.x < 0) ? lox_loy.y : loz_hix.w;

uint swizzled_loy = (idir.y < 0) ? hiy_hiz.y : lox_loy.w;
uint swizzled_hiy = (idir.y < 0) ? lox_loy.w : hiy_hiz.y;

uint swizzled_loz = (idir.z < 0) ? hiy_hiz.w : loz_hix.y;
uint swizzled_hiz = (idir.z < 0) ? loz_hix.y : hiy_hiz.w;

txmin[0] = float((swizzled_lox ) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmin[1] = float((swizzled_lox >> 8) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmin[2] = float((swizzled_lox >> 16) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmin[3] = float((swizzled_lox >> 24) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
tymin[0] = float((swizzled_loy ) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymin[1] = float((swizzled_loy >> 8) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymin[2] = float((swizzled_loy >> 16) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymin[3] = float((swizzled_loy >> 24) & 0xffu) * adjusted_idir_y + adjusted_origin.y;

tzmin[0] = float((swizzled_loz ) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmin[1] = float((swizzled_loz >> 8) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmin[2] = float((swizzled_loz >> 16) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmin[3] = float((swizzled_loz >> 24) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
txmax[0] = float((swizzled_hix ) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmax[1] = float((swizzled_hix >> 8) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmax[2] = float((swizzled_hix >> 16) & 0xffu) * adjusted_idir_x + adjusted_origin.x;
txmax[3] = float((swizzled_hix >> 24) & 0xffu) * adjusted_idir_x + adjusted_origin.x;

tymax[0] = float((swizzled_hiy ) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymax[1] = float((swizzled_hiy >> 8) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymax[2] = float((swizzled_hiy >> 16) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tymax[3] = float((swizzled_hiy >> 24) & 0xffu) * adjusted_idir_y + adjusted_origin.y;
tzmax[0] = float((swizzled_hiz ) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmax[1] = float((swizzled_hiz >> 8) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmax[2] = float((swizzled_hiz >> 16) & 0xffu) * adjusted_idir_z + adjusted_origin.z;
tzmax[3] = float((swizzled_hiz >> 24) & 0xffu) * adjusted_idir_z + adjusted_origin.z;

ctmin = max(max(txmin[0], tymin[0]), max(tzmin[0], hit_tmin));
ctmax = min(min(txmax[0], tymax[0]), min(tzmax[0], hit_t));
if(ctmin <= ctmax) hitmask |= ((child_bits4 ) & 0xffu) << ((bit_index4 ) & 0xffu);

ctmin = max(max(txmin[1], tymin[1]), max(tzmin[1], hit_tmin));
ctmax = min(min(txmax[1], tymax[1]), min(tzmax[1], hit_t));
if(ctmin <= ctmax) hitmask |= ((child_bits4 >> 8u) & 0xffu) << ((bit_index4 >> 8u) & 0xffu);

ctmin = max(max(txmin[2], tymin[2]), max(tzmin[2], hit_tmin));
ctmax = min(min(txmax[2], tymax[2]), min(tzmax[2], hit_t));
if(ctmin <= ctmax) hitmask |= ((child_bits4 >> 16u) & 0xffu) << ((bit_index4 >> 16u) & 0xffu);

ctmin = max(max(txmin[3], tymin[3]), max(tzmin[3], hit_tmin));
ctmax = min(min(txmax[3], tymax[3]), min(tzmax[3], hit_t));
if(ctmin <= ctmax) hitmask |= ((child_bits4 >> 24u) & 0xffu) << ((bit_index4 >> 24u) & 0xffu);
}

node_group.y = (hitmask & 0xff000000u) | ((head_w >> 24u) & 0xffu);
tri_group.y = hitmask & 0x00ffffffu;
}
else
{
tri_group = node_group;
node_group = uvec2(0);
}

while(tri_group.y != 0)
{
uint tridx = findLSB(tri_group.y);
tri_group.y &= ~(1u << tridx);
tridx += tri_group.x;

tv00 = uTriMatrices[tridx].m0;
tv11 = uTriMatrices[tridx].m1;
tv22 = uTriMatrices[tridx].m2;

toz = tv00.w - dot(origin, tv00.xyz);
tidz = 1.0 / dot(dir, tv00.xyz);
tt = toz * tidz;

tox = tv11.w + dot(origin, tv11.xyz);
tdx = dot(dir, tv11.xyz);
tu = tox + tt*tdx;

toy = tv22.w + dot(origin, tv22.xyz);
tdy = dot(dir, tv22.xyz);
tv = toy + tt*tdy;

if(tt > hit_tmin && tt < hit_t)
if(tu >= 0.0 && tu <= 1.0)
if(tv >= 0.0 && tu + tv <= 1.0)
{
hit_t = tt;
return true;
}
}

if(node_group.y <= 0x00ffffffu)
{
if(stack_ptr == 0)
break;
node_group = stack[--stack_ptr];
}
}
return false;
}

0 comments on commit 1c767cf

Please sign in to comment.