gen_dgp_profile throws "ds_illum.exe stopped working" error

radiance-daysim
1

Hi all,

I'm running a number of annual glare analysis using gen_dgp_profile and it
randomly gives me "ds_illum.exe stopped working " error. Interestingly
enough when I close the error. The study keeps running and returns some
[wrong] results.

The error looks to be view dependent as it works fine for some views and
not for the other ones. Anybody has encountered similar issue and/or knows
the solution?

Thanks in advance,
Mostapha

PS: I'm using Daysim 4.0 and Here is the view file:

rview Perspective -vth -vp 2.004 5.478 1.496 -vd -0.526326 -2.530093 0.0
-vu 0.0 0.0 1.0 -vh 180.000 -vv 180.000 -vs 0.000 -vl 0.000 -x 250 -y 250

2

Hi Mostapha,

sorry I just saw by chance your posting... do you still have this problem?

Have you tried with angular fish eye view? I'm also not sure if the keyword "Perspective" in the view file might cause a problem.
Try it with this and see if this solves your problem:

rview -vta -vp 2.004 5.478 1.496 -vd -0.526326 -2.530093 0.0 -vu 0.0 0.0 1.0 -vh 180.000 -vv 180.000

Although -vth is implemented in evalglare, I'm still a little suspicious if a hemispherical view works in all kind of situations correctly (especially if a glare source is close to the border). The main thing is, that "outside" the "circle", the view is not defined for a hemispherical view. If a pixel lies partly inside and outside the circle, problems might occur and then division by zero is causing overflows in variables. I tried to prevent this and for the testing scenes it worked.
But I still recommend also in the description to use an angular fish eye view, where this cannot happen. So as long as you are simulating, please use always a angular fish eye projection.

But, I also don�t know if this is really the problem here, but it might be.

sorry for the late answer.

Best and good luck

Jan

···

On 09/08/2014 02:53 AM, Mostapha Sadeghipour wrote:

Hi all,

I'm running a number of annual glare analysis using gen_dgp_profile and it randomly gives me "ds_illum.exe stopped working " error. Interestingly enough when I close the error. The study keeps running and returns some [wrong] results.

The error looks to be view dependent as it works fine for some views and not for the other ones. Anybody has encountered similar issue and/or knows the solution?

Thanks in advance,
Mostapha

PS: I'm using Daysim 4.0 and Here is the view file:

rview Perspective -vth -vp 2.004 5.478 1.496 -vd -0.526326 -2.530093 0.0 -vu 0.0 0.0 1.0 -vh 180.000 -vv 180.000 -vs 0.000 -vl 0.000 -x 250 -y 250

_______________________________________________
Radiance-daysim mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-daysim

3

Hi Jan,

Thanks for the reply. I couldn't find this email and reply back to the
group to update everyone.

The issue turned out to be exporting an empty sensor file. I assumed that
Daysim doesn't need sensor information and will only look for .vf file if I
want to only run an annual glare analysis but then I checked the source
code and I saw that gen_dgp_profile looks for sensor file in two different
functions.

I'm not still clear how gen_dgp_profile uses sensor points for annual glare
analysis but adding the sensor files solved the issue. I really appreciate
it if you can shed some light about the relationship between sensor files
and annual glare analysis.

Best,
Mostapha

···

On Wed, Oct 29, 2014 at 11:27 AM, Jan Wienold <[email protected]> wrote:

Hi Mostapha,

sorry I just saw by chance your posting... do you still have this problem?

Have you tried with angular fish eye view? I'm also not sure if the
keyword "Perspective" in the view file might cause a problem.
Try it with this and see if this solves your problem:

rview -vta -vp 2.004 5.478 1.496 -vd -0.526326 -2.530093 0.0 -vu 0.0 0.0
1.0 -vh 180.000 -vv 180.000

Although -vth is implemented in evalglare, I'm still a little suspicious
if a hemispherical view works in all kind of situations correctly
(especially if a glare source is close to the border). The main thing is,
that "outside" the "circle", the view is not defined for a hemispherical
view. If a pixel lies partly inside and outside the circle, problems might
occur and then division by zero is causing overflows in variables. I tried
to prevent this and for the testing scenes it worked.
But I still recommend also in the description to use an angular fish eye
view, where this cannot happen. So as long as you are simulating, please
use always a angular fish eye projection.

But, I also don´t know if this is really the problem here, but it might be.

sorry for the late answer.

Best and good luck

Jan

On 09/08/2014 02:53 AM, Mostapha Sadeghipour wrote:

Hi all,

I'm running a number of annual glare analysis using gen_dgp_profile and
it randomly gives me "ds_illum.exe stopped working " error. Interestingly
enough when I close the error. The study keeps running and returns some
[wrong] results.

The error looks to be view dependent as it works fine for some views and
not for the other ones. Anybody has encountered similar issue and/or knows
the solution?

Thanks in advance,
Mostapha

PS: I'm using Daysim 4.0 and Here is the view file:

rview Perspective -vth -vp 2.004 5.478 1.496 -vd -0.526326 -2.530093
0.0 -vu 0.0 0.0 1.0 -vh 180.000 -vv 180.000 -vs 0.000 -vl 0.000 -x 250 -y
250

_______________________________________________
Radiance-daysim mailing [email protected]://www.radiance-online.org/mailman/listinfo/radiance-daysim

_______________________________________________
Radiance-daysim mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-daysim