Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jos Leys DeltaDE broken in OpenCL #627

Closed
mclarekin opened this issue Mar 8, 2019 · 8 comments
Closed

Jos Leys DeltaDE broken in OpenCL #627

mclarekin opened this issue Mar 8, 2019 · 8 comments
Labels

Comments

@mclarekin
Copy link
Collaborator

mclarekin commented Mar 8, 2019

maybe something i did, i will try and find it

# Mandelbulber settings file
# version 2.17
# only modified parameters
[main_parameters]
ambient_occlusion_enabled true;
background_3_colors_enable false;
background_color_1 0000 0000 0000;
camera -0.3497615304715637 1.935824043354695 3.349010474846709;
camera_distance_to_target 8.700000000000001;
camera_rotation -90 -90 -180;
camera_top -1 6.123233995736764e-17 -6.123233995736765e-17;
check_for_bailout_3 false;
DE_factor 0.3;
delta_DE_function 4;
delta_DE_method 1;
detail_level 16;
file_background C:\Users\graeme\mandelbulber\my textures\background.jpg;
file_envmap C:\Users\graeme\mandelbulber\my textures\envmap.jpg;
file_lightmap C:\Users\graeme\mandelbulber\my textures\lightmap.jpg;
flight_last_to_render 0;
formula_1 122;
formula_2 1084;
formula_3 1038;
formula_start_iteration_3 5;
formula_stop_iteration_2 1;
formula_stop_iteration_3 6;
fov 0.7;
fractal_enable_2 false;
fractal_enable_3 false;
glow_enabled false;
image_proportion 2;
keyframe_last_to_render 0;
mat1_fractal_coloring_algorithm 3;
mat1_fractal_coloring_aux_color_scale1 0;
mat1_fractal_coloring_orbit_trap_scale1 89.59999999999999;
mat1_fractal_coloring_rad_div_de_scale1 0;
mat1_is_defined true;
mat1_metallic false;
mat1_specular 1;
mat1_specular_plastic_enable false;
mat1_specular_width 1;
mat1_surface_color_palette fd6029 698403 fff59c 000000 550000 000000 550000 000000 550000 aba53c;
N 49;
raytraced_reflections true;
target -0.3497615304715631 1.935824043354695 -5.350989525153292;
[fractal_1]
transf_folding_value 1.99;
transf_function_enabled_false true;
transf_maxR2_1 0.4;
transf_offset_000 -0.2 -1.45 0.3;
transf_sphere_inversion_enabled_false true;
[fractal_2]
mandelbox_folding_fixed_radius 99;
mandelbox_folding_min_radius 1;
transf_constant_multiplier_111 0.8 0.5 0;
transf_function_enabled false;
transf_function_enabled_false true;
transf_maxR2_1 0.2;
transf_offset 0.11;
transf_offset_000 -0.3 1.5 0;
transf_sphere_inversion_enabled_false true;
[fractal_3]
info true;
mandelbox_folding_fixed_radius 99;
mandelbox_folding_min_radius 1;
transf_constant_multiplier_111 0.8 0.5 0;
transf_function_enabled false;
transf_maxR2_1 0.2;
transf_offset 0.11;
transf_offset_000 -0.3 1.5 0;
transf_sphere_inversion_enabled_false true;
@mclarekin mclarekin added the BUG label Mar 8, 2019
@rconcept
Copy link

rconcept commented Mar 8, 2019

I played a bit with different parameters (of your example) and it seems that the "noisy" pixels (only in OpenCL) are directly related to the InverR2 parameter.

@mclarekin
Copy link
Collaborator Author

I have just been looking at this. I find three problems, hopefully fixing the first two will also fix the third (only a color thing).

The first is the JosLeysys Kleinian does not render correctly with DeltaDE JosLeys with openCL, when the sphere inversion is enabled

No_openCL with sphInversion
FULL with sphInversion

@mclarekin
Copy link
Collaborator Author

mclarekin commented Apr 27, 2019

so turn off sphereInv and no problem. Hmmm, so looks like problem in the sphereInversion part of the openCL code (but maybe something else as well)
BACKGROUND info, I had made this change but it does not look like it is the cause.
// aux.pseudoKleinianDE *= iR; // TODO remove after testing
aux.DE *= fabs(iR);

FULL
No_openCL

@mclarekin
Copy link
Collaborator Author

So i will have a look at the code before recording the second problem

@mclarekin
Copy link
Collaborator Author

I cannot see any obvious difference between the c++ and openCL sphereinversion code

@buddhi1980
Copy link
Owner

buddhi1980 commented Apr 27, 2019

Now is better. DeltaDE for JosKleinian fractals was not implemented at all. I have no idea why I forgot about this.

@buddhi1980
Copy link
Owner

Test it and close issue if works with your different settings.

@mclarekin
Copy link
Collaborator Author

Thanks

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

3 participants