mirror of
https://github.com/Cubitect/cubiomes.git
synced 2025-09-23 03:33:50 -04:00
Perfromace and memory optimization for VoronoiZoom 1.15+
This commit is contained in:
Cubitect
parent
6f2210094f
commit
c86145a640
266
layers.c
266
layers.c
@ -356,7 +356,7 @@ static float distsq(const float *a, const float *b, int n)
|
||||
|
||||
int getNetherBiome(const NetherNoise *nn, int x, int y, int z)
|
||||
{
|
||||
const float mixpoints[5][6] = {
|
||||
const float bpoints[5][6] = {
|
||||
{ nether_wastes, 0, 0, 0, 0, 0 },
|
||||
{ soul_sand_valley, 0, -0.5, 0, 0, 0 },
|
||||
{ crimson_forest, 0.4, 0, 0, 0, 0 },
|
||||
@ -376,7 +376,7 @@ int getNetherBiome(const NetherNoise *nn, int x, int y, int z)
|
||||
float dmin = FLT_MAX;
|
||||
for (i = 0; i < 5; i++)
|
||||
{
|
||||
float dsq = distsq(point, &mixpoints[i][1], 5);
|
||||
float dsq = distsq(point, &bpoints[i][1], 5);
|
||||
if (dsq < dmin)
|
||||
{
|
||||
dmin = dsq;
|
||||
@ -384,7 +384,7 @@ int getNetherBiome(const NetherNoise *nn, int x, int y, int z)
|
||||
}
|
||||
}
|
||||
|
||||
id = (int) mixpoints[id][0];
|
||||
id = (int) bpoints[id][0];
|
||||
return id;
|
||||
}
|
||||
|
||||
@ -1999,6 +1999,23 @@ int mapOceanMix(const Layer * l, int * out, int x, int z, int w, int h)
|
||||
return 0;
|
||||
}
|
||||
|
||||
static inline void getVoronoiCell(int64_t sha, int a, int b, int c,
|
||||
int *x, int *y, int *z)
|
||||
{
|
||||
int64_t s = sha;
|
||||
s = mcStepSeed(s, a);
|
||||
s = mcStepSeed(s, b);
|
||||
s = mcStepSeed(s, c);
|
||||
s = mcStepSeed(s, a);
|
||||
s = mcStepSeed(s, b);
|
||||
s = mcStepSeed(s, c);
|
||||
|
||||
*x = (((s >> 24) & 1023) - 512) * 36;
|
||||
s = mcStepSeed(s, sha);
|
||||
*y = (((s >> 24) & 1023) - 512) * 36;
|
||||
s = mcStepSeed(s, sha);
|
||||
*z = (((s >> 24) & 1023) - 512) * 36;
|
||||
}
|
||||
|
||||
int mapVoronoiZoom(const Layer * l, int * out, int x, int z, int w, int h)
|
||||
{
|
||||
@ -2016,120 +2033,147 @@ int mapVoronoiZoom(const Layer * l, int * out, int x, int z, int w, int h)
|
||||
int64_t sha = l->startSalt;
|
||||
int *buf = (int *) malloc(w*h*sizeof(*buf));
|
||||
|
||||
int x000, x001, x010, x011, x100, x101, x110, x111;
|
||||
int y000, y001, y010, y011, y100, y101, y110, y111;
|
||||
int z000, z001, z010, z011, z100, z101, z110, z111;
|
||||
int pi, pj, ii, jj, dx, dz, pjz, pix, i4, j4;
|
||||
int v00, v01, v10, v11, v;
|
||||
int prev_skip;
|
||||
int64_t r;
|
||||
uint64_t d, dmin;
|
||||
int i, j;
|
||||
typedef struct Touple { int x, y, z; } Touple;
|
||||
Touple *prbuf = (Touple *) malloc(2 * pW * pH * sizeof(Touple));
|
||||
|
||||
for (j = 0; j < pH; j++)
|
||||
for (pj = 0; pj < pH-1; pj++)
|
||||
{
|
||||
for (i = 0; i < pW; i++)
|
||||
v00 = out[(pj+0)*pW];
|
||||
v10 = out[(pj+1)*pW];
|
||||
pjz = pZ + pj;
|
||||
j4 = ((pjz) << 2) - z;
|
||||
prev_skip = 1;
|
||||
|
||||
for (pi = 0; pi < pW-1; pi++)
|
||||
{
|
||||
int64_t s = sha;
|
||||
s = mcStepSeed(s, pX+i);
|
||||
s = mcStepSeed(s, -1);
|
||||
s = mcStepSeed(s, pZ+j);
|
||||
s = mcStepSeed(s, pX+i);
|
||||
s = mcStepSeed(s, -1);
|
||||
s = mcStepSeed(s, pZ+j);
|
||||
PREFETCH( buf + ((pjz << 2) + 0) * w + pi, 1, 1 );
|
||||
PREFETCH( buf + ((pjz << 2) + 1) * w + pi, 1, 1 );
|
||||
PREFETCH( buf + ((pjz << 2) + 2) * w + pi, 1, 1 );
|
||||
PREFETCH( buf + ((pjz << 2) + 3) * w + pi, 1, 1 );
|
||||
|
||||
int rx = (((s >> 24) & 1023) - 512) * 36;
|
||||
s = mcStepSeed(s, sha);
|
||||
int ry = (((s >> 24) & 1023) - 512) * 36;
|
||||
s = mcStepSeed(s, sha);
|
||||
int rz = (((s >> 24) & 1023) - 512) * 36;
|
||||
|
||||
Touple *pt = prbuf + 2*(j*pW+i);
|
||||
pt->x = rx;
|
||||
pt->y = ry;
|
||||
pt->z = rz;
|
||||
|
||||
s = sha;
|
||||
s = mcStepSeed(s, pX+i);
|
||||
s = mcStepSeed(s, 0);
|
||||
s = mcStepSeed(s, pZ+j);
|
||||
s = mcStepSeed(s, pX+i);
|
||||
s = mcStepSeed(s, 0);
|
||||
s = mcStepSeed(s, pZ+j);
|
||||
|
||||
rx = (((s >> 24) & 1023) - 512) * 36;
|
||||
s = mcStepSeed(s, sha);
|
||||
ry = (((s >> 24) & 1023) - 512) * 36;
|
||||
s = mcStepSeed(s, sha);
|
||||
rz = (((s >> 24) & 1023) - 512) * 36;
|
||||
|
||||
pt++;
|
||||
pt->x = rx;
|
||||
pt->y = ry;
|
||||
pt->z = rz;
|
||||
}
|
||||
}
|
||||
|
||||
for (j = 0; j < h; j++)
|
||||
{
|
||||
for (i = 0; i < w; i++)
|
||||
{
|
||||
int xi = x+i;
|
||||
int zj = z+j;
|
||||
int pi = (xi >> 2) - pX;
|
||||
int pj = (zj >> 2) - pZ;
|
||||
|
||||
int v00 = out[(pj+0)*pW + (pi+0)];
|
||||
int v01 = out[(pj+0)*pW + (pi+1)];
|
||||
int v10 = out[(pj+1)*pW + (pi+0)];
|
||||
int v11 = out[(pj+1)*pW + (pi+1)];
|
||||
int v = v00;
|
||||
v01 = out[(pj+0)*pW + (pi+1)];
|
||||
v11 = out[(pj+1)*pW + (pi+1)];
|
||||
pix = pX + pi;
|
||||
i4 = ((pix) << 2) - x;
|
||||
|
||||
if (v00 == v01 && v00 == v10 && v00 == v11)
|
||||
{
|
||||
buf[j*w + i] = v00;
|
||||
for (jj = 0; jj < 4; jj++)
|
||||
{
|
||||
j = j4 + jj;
|
||||
if (j < 0 || j >= h) continue;
|
||||
for (ii = 0; ii < 4; ii++)
|
||||
{
|
||||
i = i4 + ii;
|
||||
if (i < 0 || i >= w) continue;
|
||||
buf[j*w + i] = v00;
|
||||
}
|
||||
}
|
||||
prev_skip = 1;
|
||||
continue;
|
||||
}
|
||||
|
||||
int dx = (xi & 3) * 10240;
|
||||
int dz = (zj & 3) * 10240;
|
||||
|
||||
uint64_t dmin = (uint64_t)-1;
|
||||
int bx, bz;
|
||||
for (bz = 0; bz <= 1; bz++)
|
||||
if (prev_skip)
|
||||
{
|
||||
for (bx = 0; bx <= 1; bx++)
|
||||
getVoronoiCell(sha, pix, -1, pjz+0, &x000, &y000, &z000);
|
||||
getVoronoiCell(sha, pix, 0, pjz+0, &x001, &y001, &z001);
|
||||
getVoronoiCell(sha, pix, -1, pjz+1, &x100, &y100, &z100);
|
||||
getVoronoiCell(sha, pix, 0, pjz+1, &x101, &y101, &z101);
|
||||
prev_skip = 0;
|
||||
}
|
||||
getVoronoiCell(sha, pix+1, -1, pjz+0, &x010, &y010, &z010);
|
||||
getVoronoiCell(sha, pix+1, 0, pjz+0, &x011, &y011, &z011);
|
||||
getVoronoiCell(sha, pix+1, -1, pjz+1, &x110, &y110, &z110);
|
||||
getVoronoiCell(sha, pix+1, 0, pjz+1, &x111, &y111, &z111);
|
||||
|
||||
|
||||
for (jj = 0; jj < 4; jj++)
|
||||
{
|
||||
j = j4 + jj;
|
||||
if (j < 0 || j >= h) continue;
|
||||
for (ii = 0; ii < 4; ii++)
|
||||
{
|
||||
int ax = pi + bx;
|
||||
int az = pj + bz;
|
||||
int64_t r;
|
||||
uint64_t d;
|
||||
i = i4 + ii;
|
||||
if (i < 0 || i >= w) continue;
|
||||
|
||||
Touple *pt = prbuf + 2*( (az)*pW + (ax) );
|
||||
const int A = 40*1024;
|
||||
const int B = 20*1024;
|
||||
dx = ii * 10*1024;
|
||||
dz = jj * 10*1024;
|
||||
dmin = (uint64_t)-1;
|
||||
|
||||
v = v00;
|
||||
d = 0;
|
||||
r = pt->x - 40*1024*bx + dx; d += r*r;
|
||||
r = pt->y + 20*1024; d += r*r;
|
||||
r = pt->z - 40*1024*bz + dz; d += r*r;
|
||||
|
||||
if (d < dmin)
|
||||
{
|
||||
dmin = d;
|
||||
v = out[az*pW + ax];
|
||||
}
|
||||
|
||||
pt++;
|
||||
r = x000 - 0 + dx; d += r*r;
|
||||
r = y000 + B; d += r*r;
|
||||
r = z000 - 0 + dz; d += r*r;
|
||||
if (d < dmin) { dmin = d; }
|
||||
d = 0;
|
||||
r = pt->x - 40*1024*bx + dx; d += r*r;
|
||||
r = pt->y - 20*1024; d += r*r;
|
||||
r = pt->z - 40*1024*bz + dz; d += r*r;
|
||||
r = x001 - 0 + dx; d += r*r;
|
||||
r = y001 - B; d += r*r;
|
||||
r = z001 - 0 + dz; d += r*r;
|
||||
if (d < dmin) { dmin = d; }
|
||||
|
||||
if (d < dmin)
|
||||
{
|
||||
dmin = d;
|
||||
v = out[az*pW + ax];
|
||||
}
|
||||
d = 0;
|
||||
r = x010 - A + dx; d += r*r;
|
||||
r = y010 + B; d += r*r;
|
||||
r = z010 - 0 + dz; d += r*r;
|
||||
if (d < dmin) { dmin = d; v = v01; }
|
||||
d = 0;
|
||||
r = x011 - A + dx; d += r*r;
|
||||
r = y011 - B; d += r*r;
|
||||
r = z011 - 0 + dz; d += r*r;
|
||||
if (d < dmin) { dmin = d; v = v01; }
|
||||
|
||||
d = 0;
|
||||
r = x100 - 0 + dx; d += r*r;
|
||||
r = y100 + B; d += r*r;
|
||||
r = z100 - A + dz; d += r*r;
|
||||
if (d < dmin) { dmin = d; v = v10; }
|
||||
d = 0;
|
||||
r = x101 - 0 + dx; d += r*r;
|
||||
r = y101 - B; d += r*r;
|
||||
r = z101 - A + dz; d += r*r;
|
||||
if (d < dmin) { dmin = d; v = v10; }
|
||||
|
||||
d = 0;
|
||||
r = x110 - A + dx; d += r*r;
|
||||
r = y110 + B; d += r*r;
|
||||
r = z110 - A + dz; d += r*r;
|
||||
if (d < dmin) { dmin = d; v = v11; }
|
||||
d = 0;
|
||||
r = x111 - A + dx; d += r*r;
|
||||
r = y111 - B; d += r*r;
|
||||
r = z111 - A + dz; d += r*r;
|
||||
if (d < dmin) { dmin = d; v = v11; }
|
||||
|
||||
buf[j*w + i] = v;
|
||||
}
|
||||
}
|
||||
|
||||
buf[j*w + i] = v;
|
||||
x000 = x010;
|
||||
y000 = y010;
|
||||
z000 = z010;
|
||||
x100 = x110;
|
||||
y100 = y110;
|
||||
z100 = z110;
|
||||
x001 = x011;
|
||||
y001 = y011;
|
||||
z001 = z011;
|
||||
x101 = x111;
|
||||
y101 = y111;
|
||||
z101 = z111;
|
||||
v00 = v01;
|
||||
v10 = v11;
|
||||
}
|
||||
}
|
||||
|
||||
free(prbuf);
|
||||
memcpy(out, buf, w*h*sizeof(*buf));
|
||||
free(buf);
|
||||
return 0;
|
||||
@ -2279,6 +2323,7 @@ void voronoiAccess3D(int64_t sha, int x, int y, int z, int *x4, int *y4, int *z4
|
||||
int dx = (x & 3) * 10240;
|
||||
int dy = (y & 3) * 10240;
|
||||
int dz = (z & 3) * 10240;
|
||||
int ax = 0, ay = 0, az = 0;
|
||||
uint64_t dmin = (uint64_t)-1;
|
||||
int i;
|
||||
|
||||
@ -2287,23 +2332,12 @@ void voronoiAccess3D(int64_t sha, int x, int y, int z, int *x4, int *y4, int *z4
|
||||
int bx = (i & 4) != 0;
|
||||
int by = (i & 2) != 0;
|
||||
int bz = (i & 1) != 0;
|
||||
int ax = pX + bx;
|
||||
int ay = pY + by;
|
||||
int az = pZ + bz;
|
||||
int cx = pX + bx;
|
||||
int cy = pY + by;
|
||||
int cz = pZ + bz;
|
||||
int rx, ry, rz;
|
||||
|
||||
int64_t s = sha;
|
||||
s = mcStepSeed(s, ax);
|
||||
s = mcStepSeed(s, ay);
|
||||
s = mcStepSeed(s, az);
|
||||
s = mcStepSeed(s, ax);
|
||||
s = mcStepSeed(s, ay);
|
||||
s = mcStepSeed(s, az);
|
||||
|
||||
int rx = (((s >> 24) & 1023) - 512) * 36;
|
||||
s = mcStepSeed(s, sha);
|
||||
int ry = (((s >> 24) & 1023) - 512) * 36;
|
||||
s = mcStepSeed(s, sha);
|
||||
int rz = (((s >> 24) & 1023) - 512) * 36;
|
||||
getVoronoiCell(sha, cx, cy, cz, &rx, &ry, &rz);
|
||||
|
||||
rx += dx - 40*1024*bx;
|
||||
ry += dy - 40*1024*by;
|
||||
@ -2313,11 +2347,15 @@ void voronoiAccess3D(int64_t sha, int x, int y, int z, int *x4, int *y4, int *z4
|
||||
if (d < dmin)
|
||||
{
|
||||
dmin = d;
|
||||
*x4 = ax;
|
||||
*y4 = ay;
|
||||
*z4 = az;
|
||||
ax = cx;
|
||||
ay = cy;
|
||||
az = cz;
|
||||
}
|
||||
}
|
||||
|
||||
if (x4) *x4 = ax;
|
||||
if (y4) *y4 = ay;
|
||||
if (z4) *z4 = az;
|
||||
}
|
||||
|
||||
|
||||
|
||||
21
layers.h
21
layers.h
@ -164,14 +164,14 @@ STRUCT(DoublePerlinNoise)
|
||||
|
||||
STRUCT(Layer)
|
||||
{
|
||||
int64_t layerSalt; // (depends only on layer salt)
|
||||
int64_t startSalt; // (world seed dependent) = worldGenSeed, used for RND beyond the first
|
||||
int64_t startSeed; // (world seed dependent) starting point for chunk seeds
|
||||
int64_t layerSalt; // processed salt or initialization mode
|
||||
int64_t startSalt; // (depends on world seed) used to step PRNG forward
|
||||
int64_t startSeed; // (depends on world seed) starting point for chunk seeds
|
||||
|
||||
void *noise; // seed dependent data for noise maps
|
||||
void *data; // generic data for custom layers
|
||||
|
||||
int scale; // map scale of this layer (map entry = scale x scale blocks)
|
||||
int scale; // scale of this layer (cell = scale x scale blocks)
|
||||
int edge; // maximum border required from parent layer
|
||||
|
||||
int (*getMap)(const Layer *, int *, int, int, int, int);
|
||||
@ -231,13 +231,13 @@ int getNetherBiome(const NetherNoise *nn, int x, int y, int z);
|
||||
/**
|
||||
* The seed pipeline:
|
||||
*
|
||||
* getLayerSalt(n) -> layerSalt (ls)
|
||||
* layerSalt (ls) & worldSeed (ws) -> startSalt (st) & startSeed (ss)
|
||||
* startSeed (ss) & coords (x,z) -> chunkSeed (cs)
|
||||
* getLayerSalt(n) -> layerSalt (ls)
|
||||
* layerSalt (ls), worldSeed (ws) -> startSalt (st), startSeed (ss)
|
||||
* startSeed (ss), coords (x,z) -> chunkSeed (cs)
|
||||
*
|
||||
* The chunkSeed alone is enough to generate the first RND integer with:
|
||||
* The chunkSeed alone is enough to generate the first PRNG integer with:
|
||||
* mcFirstInt(cs, mod)
|
||||
* subsequent RND integers are generated by stepping the chunkSeed forwards,
|
||||
* subsequent PRNG integers are generated by stepping the chunkSeed forwards,
|
||||
* salted with startSalt:
|
||||
* cs_next = mcStepSeed(cs, st)
|
||||
*/
|
||||
@ -415,6 +415,9 @@ int mapOceanMix (const Layer *, int *, int, int, int, int);
|
||||
int mapVoronoiZoom (const Layer *, int *, int, int, int, int);
|
||||
int mapVoronoiZoom114 (const Layer *, int *, int, int, int, int);
|
||||
|
||||
// With 1.15, biomes are only generated up to scale 1:4 OceanMix on the server-
|
||||
// side while voronoi is primarily treated as a client-side access pattern.
|
||||
// It is seeded by the first 8-bytes of the SHA-256 hash of the world seed.
|
||||
int64_t getVoroniSHA(int64_t worldSeed);
|
||||
void voronoiAccess3D(int64_t sha, int x, int y, int z, int *x4, int *y4, int *z4);
|
||||
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user