-
Notifications
You must be signed in to change notification settings - Fork 121
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
[edited] GPU: keep_interpolated_fields #759
Comments
Thank you for reporting this bug. Are you able to try the latest version from GitHub to see if this was previously resolved? |
Thank you. After upgrading to the latest version (Version : 5.1-55-ga31451771-master), ID tracking of particle works normally, and the sorting process is much faster than before. However, after adding
I tried several times, but the same error still occurs. These files might be useful.
|
Indeed, Let me change the title of this issue to make it as a feature request |
Okay, thank you very much! |
When using GPU computation, the ID list in trackparticle seems to be reassigned based on the current position at each timestamp, while with CPU computation, particle IDs are inherited across different timestamps, allowing for tracking.
Description
Please refer to the attached image and code.
I attempted to perform a TNSA simulation and tracked some particles that moved to the back of the simulation area (x > 50) using their IDs, as shown in the figure. It can be seen that the results from the CPU and GPU are completely different. In the GPU version, these IDs seem to be reassigned based on position at each step, so they always appear together. In the final time step, they suddenly moved to the back of the simulation area.
CPU
GPU
input_gpu.py.txt
input_cpu.py.txt
plot.py.txt
Parameters
The text was updated successfully, but these errors were encountered: