mirror-github-bspeice/cuburn
1
0
Fork 0
mirror of https://github.com/stevenrobertson/cuburn.git synced 2025-11-03 18:00:55 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

Correct dither fail.

Browse Source
This commit is contained in:
Steven Robertson
2011-12-21 11:59:40 -05:00
parent adb79901f1
commit 09725ba794
3 changed files with 16 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
cuburn/code/interp.py
View File

@ -221,14 +221,13 @@ class GenomePacker(HunkOCode):
__global__ __global__
void interp_{{tname}}( void interp_{{tname}}(
{{tname}}* out, mwc_st *rctxes, {{tname}}* out,
const float *times, const float *knots, const float *times, const float *knots,
float tstart, float tstep, int maxid float tstart, float tstep, int maxid
) { ) {
int id = gtid(); int id = gtid();
if (id >= maxid) return; if (id >= maxid) return;
out = &out[id]; out = &out[id];
mwc_st rctx = rctxes[id];
float time = tstart + id * tstep; float time = tstart + id * tstep;
float *outf = reinterpret_cast<float*>(out); float *outf = reinterpret_cast<float*>(out);

27
cuburn/code/iter.py
View File

@ -70,23 +70,15 @@ def precalc_camera(pcam):
float cenx = {{pre_cam.center.x}}, ceny = {{pre_cam.center.y}}; float cenx = {{pre_cam.center.x}}, ceny = {{pre_cam.center.y}};
float scale = {{pre_cam.scale}} * acc_size.width; float scale = {{pre_cam.scale}} * acc_size.width;
float ditherwidth = {{pre_cam.dither_width}} * 0.33f;
float u0 = mwc_next_01(rctx);
float r = ditherwidth * sqrtf(-2.0f * log2f(u0) / M_LOG2E);
float u1 = 2.0f * M_PI * mwc_next_01(rctx);
float ditherx = r * cos(u1);
float dithery = r * sin(u1);
{{pre_cam._set('xx')}} = scale * rotcos; {{pre_cam._set('xx')}} = scale * rotcos;
{{pre_cam._set('xy')}} = scale * -rotsin; {{pre_cam._set('xy')}} = scale * -rotsin;
{{pre_cam._set('xo')}} = scale * (rotsin * ceny - rotcos * cenx) {{pre_cam._set('xo')}} = scale * (rotsin * ceny - rotcos * cenx)
+ 0.5f * acc_size.awidth + ditherx; + 0.5f * acc_size.awidth;
{{pre_cam._set('yx')}} = scale * rotsin; {{pre_cam._set('yx')}} = scale * rotsin;
{{pre_cam._set('yy')}} = scale * rotcos; {{pre_cam._set('yy')}} = scale * rotcos;
{{pre_cam._set('yo')}} = scale * -(rotsin * cenx + rotcos * ceny) {{pre_cam._set('yo')}} = scale * -(rotsin * cenx + rotcos * ceny)
+ 0.5f * acc_size.aheight + dithery; + 0.5f * acc_size.aheight;
""").substitute(locals())) """).substitute(locals()))
@ -203,13 +195,24 @@ void iter(
reinterpret_cast<const float*>(global_params)[i]; reinterpret_cast<const float*>(global_params)[i];
__shared__ int rb_idx; __shared__ int rb_idx;
if (threadIdx.x == 1 && threadIdx.y == 1) if (threadIdx.y == 1 && threadIdx.x == 1)
rb_idx = 32 * blockDim.y * (atomicAdd(&rb_head, 1) % rb_size); rb_idx = 32 * blockDim.y * (atomicAdd(&rb_head, 1) % rb_size);
__syncthreads(); __syncthreads();
int this_rb_idx = rb_idx + threadIdx.x + 32 * threadIdx.y; int this_rb_idx = rb_idx + threadIdx.x + 32 * threadIdx.y;
mwc_st rctx = msts[this_rb_idx]; mwc_st rctx = msts[this_rb_idx];
{{precalc_camera(pcp.camera)}}
if (threadIdx.y == 5 && threadIdx.x == 4) {
float ditherwidth = {{pcp.camera.dither_width}} * 0.5f;
float u0 = mwc_next_01(rctx);
float r = ditherwidth * sqrtf(-2.0f * log2f(u0) / M_LOG2E);
float u1 = 2.0f * M_PI * mwc_next_01(rctx);
{{pcp.camera.xo}} += r * cos(u1);
{{pcp.camera.yo}} += r * sin(u1);
}
{{if info.acc_mode == 'global'}} {{if info.acc_mode == 'global'}}
__shared__ float time_frac; __shared__ float time_frac;
time_frac = blockIdx.x / (float) gridDim.x; time_frac = blockIdx.x / (float) gridDim.x;
@ -332,8 +335,6 @@ void iter(
float cx, cy, cc; float cx, cy, cc;
{{precalc_camera(pcp.camera)}}
{{if 'final' in cp.xforms}} {{if 'final' in cp.xforms}}
{{apply_affine('fx', 'fy', 'cx', 'cy', pcp.camera)}} {{apply_affine('fx', 'fy', 'cx', 'cy', pcp.camera)}}
cc = fcolor; cc = fcolor;

2
cuburn/render.py
View File

@ -293,7 +293,7 @@ class Renderer(object):
block=(256,1,1), grid=(self.palette_height,1), block=(256,1,1), grid=(self.palette_height,1),
stream=write_stream) stream=write_stream)
packer_fun(d_infos, d_seeds, d_genome_times, d_genome_knots, packer_fun(d_infos, d_genome_times, d_genome_knots,
f32(ts), f32(td / ntemporal_samples), f32(ts), f32(td / ntemporal_samples),
i32(ntemporal_samples), block=(256,1,1), i32(ntemporal_samples), block=(256,1,1),
grid=(int(np.ceil(ntemporal_samples/256.)),1), grid=(int(np.ceil(ntemporal_samples/256.)),1),