The OPAL Discussion Forum

[Can Davut <can.davut AT manchester.ac.uk>]

Hi Andy,

 

Thanks for tracking this down!

 

I can test the patch on my workstation first before moving to our clusters in my institute. Please send over the patch and any instructions, and I’ll give it a try right away.

 

Cheers,

Can

Kimden: Adelmann Andreas <andreas.adelmann AT psi.ch>
Tarih: 02 May 2025 21:52
Kime: Can Davut <can.davut AT manchester.ac.uk>
Bilgi: opal AT lists.psi.ch
Konu: Re: [Opal] CSR Calculation Crash in Symmetric C-Chicane Simulation (OPAL-t)

 

 Hi Can I think I found and fixed the problem. 

Are you able to build OPAL at your institute? If that is the case I can send you simply a patch

and you can test. 

 

Cheers Andy 

On 2 May 2025, at 09:49, Can Davut <can.davut AT manchester.ac.uk> wrote:

 

Hi Andy,

 

I use the previous version (2022.1) as it was the latest when I started using OPAL.

 

Best Wishes,
Can

 

---

From: opal-request AT lists.psi.ch <opal-request AT lists.psi.ch> on behalf of Adelmann Andreas <andreas.adelmann AT psi.ch>
Sent: Friday, May 2, 2025 3:45:16 pm
To: Can Davut <can.davut AT manchester.ac.uk>
Cc: opal AT lists.psi.ch <Opal AT lists.psi.ch>
Subject: Re: [Opal] CSR Calculation Crash in Symmetric C-Chicane Simulation (OPAL-t)

Hi Can thanks for the additional information. Will have a look at the problem now.

Last think, I assume you use the latest release of OPAL, correct?

 

A 

On 2 May 2025, at 05:58, Can Davut <can.davut AT manchester.ac.uk> wrote:

 

Hi Andy,

 

I initially thought the error was caused by the time step I used for the chicane—since the integration depends on it—so I reduced it to 30 μm, but the error persisted. I also tweaked the filter and wake-function parameters, but without success.

 

Because I use the same field map for all four dipoles and can save the CSR data for the first dipole and the drift without issue, I’ve ruled out the field map as the source of the problem.

 

It seems that increasing the time step before the chicane causes the beam to arrive in under 30 s. If it would help, I can share an HDF5 file so you can use the restart command to diagnose what’s going wrong in the chicane.

 

I’d greatly appreciate any insights you might have.

 

Best regards,

Can

 

Kimden: Adelmann Andreas <andreas.adelmann AT psi.ch>
Tarih: 01 May 2025 17:57
Kime: Can Davut <can.davut AT manchester.ac.uk>
Konu: Re: [Opal] CSR Calculation Crash in Symmetric C-Chicane Simulation (OPAL-t)

 

Hi Can, yes I got the reproducer and had a quick look. I can not see an immediate 

error on your side. Can you make the problem smaller i.e. just the chicane? This would help in debugging

the problem. Maybe it is just an unfortunate combination of parameters.

 

We need to solve a nonlinear set of equations and the Newton solver is not converging

within the 100 iterations. Maybe your case needs more than 100 but I do not yet know. 

 

Thanks Andy  

 

On 1 May 2025, at 11:18, Can Davut <can.davut AT manchester.ac.uk> wrote:

 

Hi Andy,

 

I sent the reproducer in private yesterday. I hope you received it. Just in case I forward it.

 

Thanks in advance for your help.

 

Best Regards,

Can

 

Kimden: Can Davut
Tarih: 30 April 2025 13:35
Kime: Adelmann Andreas <andreas.adelmann AT psi.ch>
Konu: YNT: [Opal] CSR Calculation Crash in Symmetric C-Chicane Simulation (OPAL-t)

 

Hi Andy,

 

Please find attached reproducer file including the field maps and input script.

 

Best Regards,

Can

 

Kimden: Adelmann Andreas <andreas.adelmann AT psi.ch>
Tarih: 30 April 2025 13:29
Kime: Can Davut <can.davut AT manchester.ac.uk>
Bilgi: opal AT lists.psi.ch
Konu: Re: [Opal] CSR Calculation Crash in Symmetric C-Chicane Simulation (OPAL-t)

 

Hi Can can you send us a reproducer?

 

Thanks Andy 

 

 

On 30 Apr 2025, at 04:13, Can Davut <can.davut AT manchester.ac.uk> wrote:

 

Dear OPAL developers and users,

 

I am experiencing issues with the CSR calculations in my OPAL-t simulations.

 

Following the examples provided in the manual, I set up a symmetric C-shaped chicane using RBEND dipoles and the default field map ("1DPROFILE1-DEFAULT.T7") for 25 cm magnets. The bunch compresses by the desired factor of 2, and the tracking completes successfully when CSR effects are not monitored.

 

However, when I attempt to monitor CSR effects, I define a filter as:

CSR_FILTER: FILTER, TYPE = "Savitzky-Golay", NPOINTS = 20, NLEFT = 4, NRIGHT = 4, POLYORDER = 4;

 

and then define the wake function as:

FS_CSR_WAKE: WAKE, TYPE = "1D-CSR-IGF", NBIN = 200, FILTERS = "CSR_FILTER";

 

This wake function is assigned to all dipoles and subsequent drifts. When the simulation runs with this setup, it crashes at the entrance of the second dipole. I also attempted to enable CSRDUMP = TRUE to monitor the raw CSR wake data but encountered the same issue.

The relevant error message is:

<image001.png>

Could you please help me understand what might be causing this issue, so that I can investigate the CSR impact on the bunch more effectively?

 

Thank you very much in advance for your support.

 

Best Regards,

Can

 

Dr. Can Davut | Research Associate

 

Accelerator Physics Group | Department of Physics & Astronomy | The University of Manchester

 

Office : 7.25 Schuster Building / A24B Cockcroft Institute