The OPAL Discussion Forum

[John Parrack <jparrack AT g.ucla.edu>]

Thank you very much. I'll try it now!

On Mon, Aug 28, 2023 at 8:52 PM Adelmann Andreas <andreas.adelmann AT psi.ch> wrote:

Hi John the version can be found at: https://amas.web.psi.ch/people/aadelmann/pub/opal

Thanks for testing and feedback. This will then be available in 2023.1, to be released soon.

Cheers A
------
Dr. sc. math. Andreas (Andy) Adelmann
Paul Scherrer Institut OHSA/D17 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Zoom ID: 470-582-4086 Password: AdA
Zoom Link: https://ethz.zoom.us/j/4705824086?pwd=dFcvT1pMMGY0bHg0dTNncUNZZTJkZz09

-------------------------------------------------------
Friday: ETH HPK G 28   +41 44 633 3076
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 28 Aug 2023, at 18:01, John Parrack <jparrack AT g.ucla.edu> wrote:

Hey thank you, good to hear. If you could send a version that would be great. 

On Mon, Aug 28, 2023 at 1:11 AM Adelmann Andreas <andreas.adelmann AT psi.ch> wrote:

Hi John, sorry for the delay, vacation and conferences interfered. Christof Kraus has a patch available 

for our problem. Do you build OPAL by yourself, hence I could send you the patch? Otherwise I will build a version for you.

Cheers A
------
Dr. sc. math. Andreas (Andy) Adelmann
Paul Scherrer Institut OHSA/D17 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Zoom ID: 470-582-4086 Password: AdA
Zoom Link: https://ethz.zoom.us/j/4705824086?pwd=dFcvT1pMMGY0bHg0dTNncUNZZTJkZz09

-------------------------------------------------------
Friday: ETH HPK G 28   +41 44 633 3076
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 31 Jul 2023, at 16:49, John Parrack <jparrack AT g.ucla.edu> wrote:

I ran simulations using stripes at twice, 4 times, and 8 times thickness. Here are the resulting distributions just after the mask, plotted. I can send files if desired. There seems to be the same bleed through effect, in the same region for all stripe thickness variants.

<image.png>

Also, no worries, Joel, I'm also confused! Let me know if any other ideas occur. 

On Mon, Jul 31, 2023 at 4:24 AM Joel Andersson <joean407 AT gmail.com> wrote:

Hello John,

My apologies! I had misunderstood the function of  the OUTFN command, for some reason I had understood it as the actual mask according to what OPAL had interpreted. However, it seems the output in the OUTFN file is the actual particles that are lost on the aperture, which then doesn't really help you very much, sorry about that! 

Cheers

//Joel

On Sat, Jul 29, 2023 at 5:43 PM John Parrack <jparrack AT g.ucla.edu> wrote:

Hi Andy, okay, thank you. I'll run that and send the results soon.

Hi Joel, It does almost seem like the particle simulation is just using the wrong mask but I've checked the input file and mask element every way I know how and it seems correct. Is there a way I could somehow check the mask as OPAL sees it?

On Wed, Jul 26, 2023 at 9:04 AM Joel Andersson <joean407 AT gmail.com> wrote:

Hello,

Out of curiousity I uploaded the files you provided John to our cluster where we have opal simulations running, and I can confirm that I see the same results as you do.

However, when I check the "col_out.h5" file I can see that the pattern observed in the particle distribution is already observed there. Wouldn't that indicate an issue with the OPAL function for creating the mask (so that the mask OPAL uses isn't actually the mask as specified from the image), rather than bleed through from the actual particle simulation? As far as I understand the mail conversation so far this isn't something that was previously observed, my apologies if this is the case.

Best regards,

Joel

On Wed, Jul 26, 2023 at 1:52 PM Adelmann Andreas <andreas.adelmann AT psi.ch> wrote:

Hi John (please call me Andy)

It would be very helpful to see if it is the thickness of the stripes. Please run the simulation

with just varying the thickness of the stripes.  

Thanks  and 

Cheers Andy 
------
Dr. sc. math. Andreas (Andy) Adelmann
Paul Scherrer Institut OHSA/D17 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Zoom ID: 470-582-4086 Password: AdA
Zoom Link: https://ethz.zoom.us/j/4705824086?pwd=dFcvT1pMMGY0bHg0dTNncUNZZTJkZz09

-------------------------------------------------------
Friday: ETH HPK G 28   +41 44 633 3076
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 25 Jul 2023, at 22:59, John Parrack <jparrack AT g.ucla.edu> wrote:

Hi Adelmann, yes, I've run identical versions of the mask without vertical stripes (this mask and other variations) Whenever the stripes are not present, I have no issue. 

I could play around with the thickness of the vertical lines in this configuration, I've done other configurations with 20 bars instead of 26 and the thicker stripes between them caused no issue. 

I can definitely try to create this same mask with thicker gap lines if you think it will help diagnose the problem. Ultimately though, I ideally need to be able to simulate the mask with the thinner stripes as it models a physical mask element we're using.  

On Tue, Jul 25, 2023 at 12:55 Adelmann Andreas <andreas.adelmann AT psi.ch> wrote:

ok it looks like the simulation is setup nicely and from the output file I can not spot an obvious problem. 

I have two things you maybe can check easily: 

1. try to make the vertical line thicker 

2. Can you make two masks, one with the vertical stripes and one with the rest? 

Thanks & 

Cheers A
------
Dr. sc. math. Andreas (Andy) Adelmann
Paul Scherrer Institut OHSA/D17 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Zoom ID: 470-582-4086 Password: AdA
Zoom Link: https://ethz.zoom.us/j/4705824086?pwd=dFcvT1pMMGY0bHg0dTNncUNZZTJkZz09

-------------------------------------------------------
Friday: ETH HPK G 28   +41 44 633 3076
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 25 Jul 2023, at 17:29, John Parrack <jparrack AT g.ucla.edu> wrote:

On Mon, Jul 24, 2023 at 12:14 PM Adelmann Andreas <andreas.adelmann AT psi.ch> wrote:

Can you send me the slurm output from the EEX simulation?

Cheers A
------
Dr. sc. math. Andreas (Andy) Adelmann
Paul Scherrer Institut OHSA/D17 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Zoom ID: 470-582-4086 Password: AdA
Zoom Link: https://ethz.zoom.us/j/4705824086?pwd=dFcvT1pMMGY0bHg0dTNncUNZZTJkZz09

-------------------------------------------------------
Friday: ETH HPK G 28   +41 44 633 3076
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 24 Jul 2023, at 17:11, John Parrack <jparrack AT g.ucla.edu> wrote:

To reproduce, unzip MLCv2_Setup and convert input file to find requisite files in your directory. Run simulation on the EEX portion of the beamline using the MLCv2-Mask21f.pbm mask element. Extract the resulting h5 (or data around mask at step 16,17,18 or so) and view the distribution using the PlottingScript. 

If I follow this procedure using a mask without vertical stripes it seems to work normally, for the Mask21f (and 4o other variations on it) I observe the bleed through effect at the masking step and this leads to the witness beam being connected to the main distribution at the end of the beamline (an issue for us). 

Please let me know if you have any questions. Very possible I've made some mess along the way but I've been struggling to sort it out myself. 

 MLCv2_Setup.zip

On Mon, Jul 24, 2023 at 12:59 AM Adelmann Andreas <andreas.adelmann AT psi.ch> wrote:

Hi, this has nothing to do with AMR grids and yes in OPAL-t AMR is not supported.

Can you please send us a reproducer including the analysis script? Thanks Andy

------
Dr. sc. math. Andreas (Andy) Adelmann
Paul Scherrer Institut OHSA/D17 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Zoom ID: 470-582-4086 Password: AdA
Zoom Link: https://ethz.zoom.us/j/4705824086?pwd=dFcvT1pMMGY0bHg0dTNncUNZZTJkZz09

-------------------------------------------------------
Friday: ETH HPK G 28   +41 44 633 3076
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 21 Jul 2023, at 17:23, Joel Andersson <joean407 AT gmail.com> wrote:

Hello,

Just out of curiosity, have you tried with different grid sizes to see if there is any correlations between bleeding and grid points?

Best regards

Joel

Den fre 21 juli 2023 17:12John Parrack <jparrack AT g.ucla.edu> skrev:

Hello, I'm having an issue with some of the simulations I'm trying to run and having difficulty resolving it. I was hoping to reach out for help. 

The issue appears to be that when I use certain masks elements, I see bleed through in a particular region. An example is attached below. The masking element contains thin vertical stripes and I've attempted to test many variations on it but this seems to correlate with this bleed through effect in the lower left quadrant just after the masking step. If I use an identical mask without the vertical striping it looks as expected but the vertical striping is an important feature of the simulation.

Would love some direction or guidance towards eliminating this issue. I'm out of ideas. 

--

John Parrack

Particle Beam Physics Lab | Undergraduate Physics | UCLA

<PlottingScript.txt><MLCv2-Mask21f.png><PlottingScript.ipynb><MLCv2-Mask21f.pbm>

<slurm-3085.out>