Logo by mauxuam - Contribute your own Logo!
News: Follow us on Twitter
 
*
Welcome, Guest. Please login or register. November 20, 2017, 11:39:27 PM


Login with username, password and session length


The All New FractalForums is now in Public Beta Testing! Visit FractalForums.org and check it out!


Pages: [1]   Go Down
  Print  
Share this topic on DiggShare this topic on FacebookShare this topic on GoogleShare this topic on RedditShare this topic on StumbleUponShare this topic on Twitter
Author Topic: 2.11 Client Communication  (Read 599 times)
Description: Working vs. Ready
0 Members and 1 Guest are viewing this topic.
CCV
Navigator
*****
Posts: 73


« on: June 05, 2017, 12:02:16 PM »

I have to say I have noticed some networking improvements with 2.11, compared to 2.10.
Still, I'm noticing more of a problem with Clients switching from Working to Ready for no apparent reason.
I used to think it mostly occurred with Navigation changes and possibly Undo/Redo, but now it seems more prevalent with any changes.

I have stated previously that although a client is marked as Ready it is in fact Working - as far as Progress Bar and Task Manager are any indication. That seems to be the case still, but the work being done by Client (in this situation) isn't being communicated to Server.

From past experience I can say that sometimes Disconnect/Connect Client often gets it Working again. Mostly, it's just easier to work from the Server end.
In some cases it's enough to click Render (up to several times) to get everything Working again.
Failing that, Stop Server and Launch again and restart Render seems most efficacious.

Thank you for your attention.
« Last Edit: July 04, 2017, 04:53:15 AM by CCV » Logged
CCV
Navigator
*****
Posts: 73


« Reply #1 on: June 16, 2017, 11:44:00 AM »

I have to say, it is getting a little frustrating.
As said, clicking Render again as often as necessary often fixes the problem. It doesn't seem to (thinking 10x Render enough) be sufficient in this case:

# Mandelbulber settings file
# version 2.11
# only modified parameters
[main_parameters]
aux_light_colour_1 ff00 ff00 ff00;
aux_light_enabled_1 true;
aux_light_intensity_1 4.1;
aux_light_position_1 4 -4 4;
background_color_1 0000 0000 0000;
background_color_2 2a00 3e00 3600;
background_color_3 ff00 ed00 ab00;
boolean_operator_1 0;
boolean_operator_2 0;
boolean_operator_3 0;
boolean_operators true;
box_folding_limit 1.5;
camera -0.08287696032625239 -0.07941575649323826 -0.07847726559017255;
camera_distance_to_target 0.1198842972517015;
camera_rotation 129.9749721735614 -36.95097150828011 -1.00000000000004;
camera_top -0.4493796031152589 -0.3995140274796599 0.79902848143915;
DE_factor 0.1;
delta_DE_function 2;
detail_level 2;
flight_last_to_render 0;
formula_1 88;
formula_2 83;
formula_3 7;
formula_material_id_1 2;
formula_material_id_2 3;
image_height 272;
image_width 480;
julia_c 2.774110549353855 -2.656528596132875 1.904831177860441;
julia_mode_1 true;
keyframe_last_to_render 0;
main_light_alpha 24;
main_light_beta 36;
mat1_is_defined true;
mat1_surface_color_palette fd6029 698403 fff59c f5bd22 0b5e87 6a533f 2282a5 3b9fee d4ffd4 aba53c;
mat2_is_defined true;
mat2_shading 0.8;
mat2_surface_color_palette fd6029 698403 fff59c f5bd22 0b5e87 c68876 64103d 3ba4ef d4ffd4 aba21b;
mat2_transparency_of_surface 0.5;
mat3_is_defined true;
mat3_surface_color_palette fd6029 7b7402 fff59c f5bd22 633100 301b15 aa0000 f87c00 d4ffd4 f1e754;
mat4_fresnel_reflectance true;
mat4_is_defined true;
mat4_name glass;
mat4_reflectance 1;
mat4_specular 20;
mat4_specular_width 0.108;
mat4_transparency_interior_color 1900 3f00 0000;
mat4_transparency_of_interior 0.74131;
mat4_transparency_of_surface 1;
mat4_use_colors_from_palette false;
raytraced_reflections true;
spherical_folding true;
target -0.156295170297991 -0.1409663186432736 -0.1505434814690062;
view_distance_max 36.06895516380957;
[fractal_1]
IFS_rotation_enabled true;
[fractal_2]
IFS_rotation_enabled true;
[fractal_3]
IFS_rotation_enabled true;
[fractal_4]
IFS_rotation_enabled true;
Logged
CCV
Navigator
*****
Posts: 73


« Reply #2 on: June 25, 2017, 07:33:13 AM »

A couple of updates:
  • 1
Another example where repeatedly clicking Render fails to set Clients to Working. Stop and relaunch Server, then start Render again works fine.
I find it of interest because the same settings, except with the camera at a greater distance, haven't caused so much problem. And, the closer I move in the worse it seems to get.

# Mandelbulber settings file
# version 2.11
# only modified parameters
[main_parameters]
aux_light_colour_1 ff00 ff00 ff00;
aux_light_enabled_1 true;
aux_light_intensity_1 4.1;
aux_light_position_1 4 -4 4;
background_color_1 0000 0000 0000;
background_color_2 2a00 3e00 3600;
background_color_3 ff00 ed00 ab00;
boolean_operator_1 0;
boolean_operator_2 0;
boolean_operator_3 0;
boolean_operators true;
box_folding_limit 1.5;
camera -0.08733398805806962 -0.08292766540109509 -0.08139408355882762;
camera_distance_to_target 0.1198842972517017;
camera_rotation 137.9749721735613 -31.95097150828001 -8.000000000000011;
camera_top -0.2474384205152353 -0.4824570194695414 0.8402436863300388;
DE_factor 0.1;
delta_DE_function 2;
detail_level 2;
flight_last_to_render 0;
formula_1 88;
formula_2 83;
formula_3 7;
formula_material_id_1 2;
formula_material_id_2 3;
image_height 272;
image_width 480;
julia_c 2.774110549353855 -2.656528596132875 1.904831177860441;
julia_mode_1 true;
keyframe_last_to_render 0;
main_light_alpha 24;
mat1_is_defined true;
mat1_surface_color_palette fd6029 698403 fff59c f5bd22 0b5e87 6a533f 2282a5 3b9fee d4ffd4 aba53c;
mat2_is_defined true;
mat2_shading 0.8;
mat2_surface_color_palette fd6029 698403 fff59c f5bd22 0b5e87 c68876 64103d 3ba4ef d4ffd4 aba21b;
mat2_transparency_of_surface 0.5;
mat3_is_defined true;
mat3_surface_color_palette fd6029 7b7402 fff59c f5bd22 633100 301b15 aa0000 f87c00 d4ffd4 f1e754;
mat4_fresnel_reflectance true;
mat4_is_defined true;
mat4_name glass;
mat4_reflectance 1;
mat4_specular 20;
mat4_specular_width 0.108;
mat4_transparency_interior_color 1900 3f00 0000;
mat4_transparency_of_interior 0.74131;
mat4_transparency_of_surface 1;
mat4_use_colors_from_palette false;
raytraced_reflections true;
spherical_folding true;
target -0.1554322893465567 -0.1584920855007935 -0.1448360610197572;
view_distance_max 36.06895516380957;
[fractal_1]
IFS_rotation_enabled true;
[fractal_2]
IFS_rotation_enabled true;
[fractal_3]
IFS_rotation_enabled true;
[fractal_4]
IFS_rotation_enabled true;


  • 2
Only once have I seen Clients start working while Server already Rendering. I mean, normally they don't start until rendering is started again after connecting.
In this case they started working immediately upon connection. But, both clients produced the same 'error' as illustrated in the attachment.




* Mandel-com.jpg (235.62 KB, 672x511 - viewed 77 times.)
Logged
CCV
Navigator
*****
Posts: 73


« Reply #3 on: July 01, 2017, 09:35:34 AM »

This will require further testing, but:

In the last example, DE_factor is 0.1. Yet, when I use the "High quality" Optimisation button on similar situations, it's more like 0.5, 1.xxx or even as much as 3.xxx - dependent upon how much of what is in view is Menger sponge.
Having recently used Optimisation on one of my small scale 'preview' sessions, I find Clients much more readily showing as WORKING.

Will see how it goes in future and report back. Certainly intend further exploration of this arrangement.
Logged
CCV
Navigator
*****
Posts: 73


« Reply #4 on: July 03, 2017, 05:31:33 AM »

DE_factor does seem to play a role in this, tho it's a mystery to me as to why or how. Might it be connected with the Distance estimation function being set to Delta DE Logarithmic?
For example, DE_factor of 1.15625 compares very favourably against 0.65625 - both settings left over from previous optimisations, not optimised for current renderings.
For 0.65625 I switched it to 1, which was some improvement but about 'default' for client problems too.
After High quality Optimisation, instead of 1, the value this time is 0.703125 and, again, it does seem to help.

Edit: Particularly when moving camera (& target) in a few steps I see the problem reemerging. What did optimise as 0.703125, now does so at 0.47265625. So, I guess, that tells its own story.
« Last Edit: July 03, 2017, 07:13:31 AM by CCV » Logged
mclarekin
Fractal Senior
******
Posts: 1737



« Reply #5 on: July 03, 2017, 12:11:00 PM »

About Optimization function.

From the manual:

Quote
Since the estimation contains some error (sometimes quite large), there is a risk that the
step of moving the "photon" will be too large, and incorrectly it will flow into the surface of
the fractal. This may result in visible noise in the rendered image.
To prevent this, the "photon" can be moved by the estimated distance multiplied by a
number between 0 and 1 ("ray-marching step multiplier"). Steps are then smaller, so there
is less risk of "overshooting" the surface, but the rendering time increases due to more
steps being required.

The quality produced by the DE mode and function combinations is formula specific. The
setting of formula parameters can also greatly affect the quality produced by the DE. In
some cases the choice of fractal image is determined by what location and parameters
can produce good DE quality.

In the Statistics (enable in View menu) you can see Percentage of Wrong Distance
Estimations ("Bad DE"). As a general rule less than 0.01 is good, but it is case specific and
3.0 sometimes is OK and 0.0001 sometimes is not.




Optimization function is based  on the statistics  Percentage of Wrong Distance Estimations ("Bad DE").  When Optimization is selected the programs runs a loop.  The loop commences with a step multiplier of 1.0, renders the fractal formula and checks if the Bad DE is below a threshold. If not, the step multiplier is reduced and the loop run again until the  Bad DE is below this threshold.  Threshold may be  1.0, 0.1 0.01 for low, medium and high quality, I cannot remember.

I never use Optimization, as I always have the statistics showing and reduce the raymarching step multiplier until I get the best quality image, generally aiming for Bad DE < 0.01.

Using the statistic  "Percentage of Wrong Distance Estimations" is a guide that works well in a lot of cases, but not always.


Logged
CCV
Navigator
*****
Posts: 73


« Reply #6 on: July 04, 2017, 04:52:14 AM »

Optimization function is based  on the statistics  Percentage of Wrong Distance Estimations ("Bad DE").  When Optimization is selected the programs runs a loop.  The loop commences with a step multiplier of 1.0, renders the fractal formula and checks if the Bad DE is below a threshold. If not, the step multiplier is reduced and the loop run again until the  Bad DE is below this threshold.  Threshold may be  1.0, 0.1 0.01 for low, medium and high quality, I cannot remember.

I never use Optimization, as I always have the statistics showing and reduce the raymarching step multiplier until I get the best quality image, generally aiming for Bad DE < 0.01.

Using the statistic  "Percentage of Wrong Distance Estimations" is a guide that works well in a lot of cases, but not always.

From the manual:
Quote
The quality produced by the DE mode and function combinations is formula specific. The
setting of formula parameters can also greatly affect the quality produced by the DE. In
some cases the choice of fractal image is determined by what location and parameters
can produce good DE quality.

In the Statistics (enable in View menu) you can see Percentage of Wrong Distance
Estimations ("Bad DE"). As a general rule less than 0.01 is good, but it is case specific and
3.0 sometimes is OK and 0.0001 sometimes is not.
As I mentioned, the actual results of Optimisation vary wildly dependent upon (I'm supposing) how much of what I have in view is Menger sponge in my current arrangement - where it is combined with Collatz & IQ Bulb. Plus, with almost any formula, the amount of detail (potential noise) and distances involved will vary depending on location.
Theoretically, yes, you're right, High quality Optimisation "gives 0.01% of bad distance estimation", Medium 0.1 and Low 1 - according to popup "tooltip" information.
However, for example, my last image optimisation produced a DE factor of 0.390625 giving a Percentage of wrong distance estimation: 0.840421.
Btw., in this case (and others) I found running Optimisation a second time gave a better result - it was 0.7xxx first time round.

Thanks for you input tho. smiley
Basically, I'm barely coming to grips with steering the thing to get the view I think I want. Sceptical

Still doesn't address the issue of getting Clients to switch from READY to WORKING easily.
It makes sense to me that optimal DE will change with changes to position of camera and target. I tested the idea today, as soon as the problem arose again, (after clicking Render again a few times) by running Optimisation again and it did help. DE was 0.390625 with 'bad' DE 0.840421, now it's 0.5000076293945313/0.0205764 and everything is WORKING fine - for the time being..
(Hmm.. But then, the next camera movement gives me a "wrong distance" of 0 for a good while. Ended up being 0.00136565.)
Edit: Twice tested today, actually. After zooming in a few steps, ran into trouble. Optimisation now gives me 0.453125/0.00289679.

Anyway..
Seeing that this has become more of a problem for me in version 2.11, I'm wondering what has changed since 2.10 to cause it.
I can see there's a couple of NetRender fixes which do work, as apparently intended. But I'm also thinking (based on some previous discussion) the Default Distance estimation function for the Collatz formula may have been set to Delta DE Logarithmic. It is set to that, but it could be for IQ Bulb - idk.. In any case, for what it's worth, the DE algorithm is "different for each fractal".
« Last Edit: July 04, 2017, 05:20:39 AM by CCV » Logged
Pages: [1]   Go Down
  Print  
 
Jump to:  


Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

Valid XHTML 1.0! Valid CSS! Dilber MC Theme by HarzeM
Page created in 0.343 seconds with 29 queries. (Pretty URLs adds 0.014s, 2q)