Size of queue of photos determines processing time

Dear staff,

I am working with the latest M2 PRO MacBook Pro and am currently using the most recent version of Topaz Photo AI.

I have detected a very curious and annoying issue. Specifically, I’ve observed that the processing time for each photo in Topaz Photo AI increases proportionally to the number of photos I have in the queue.

Here’s the breakdown of what I’ve noted:

  • With 1,000 photos in the queue, each photo takes 1 second to process.
  • With 2,000 photos, each photo takes 2 seconds.
  • With 10,000 photos, the time increases to 10 seconds per photo.
  • And, with 35,000 photos, it escalates to a whopping 35 seconds for each photo.

This seems to follow a linear relationship.

I have tried various methods to circumvent this behavior, but to no avail. It becomes quite inconvenient as I find myself having to work in smaller batches to prevent the overall process from slowing down. This isn’t efficient and becomes a significant disruption in my workflow.

I hope this issue can be looked into and addressed soon. I believe that having a stable and consistent processing time regardless of the batch size would be greatly beneficial to many users.

Thank you for your attention to this matter. I look forward to hearing from you soon.

Best regards,

Mario Hernández

Thanks for reaching out and reporting this. I’ve made a task for my team to investigate this behavior.

My intuition so far is that loading so many images in the application takes up resources. We can certainly try to minimize this, but that load is to take up some resources.

We’ll see if there is a better option to allow the processing to function better.

1 Like

Thank you for looking into this. I appreciate your understanding of the situation. I’ll await any further updates on this matter.

Hello Lingyu and the Topaz Support Team,

I hope this message finds you well. I’m writing to follow up on the persistent and problematic bottleneck issue I’ve been encountering with Topaz Photo AI.

To recap the specific challenge: the application is experiencing a linear increase in processing time per photo that is directly proportional to the size of the photo queue. This is symptomatic of a buffering or queuing bottleneck where the system seems unable to handle the load effectively as the queue grows. Here’s a quick reminder of the scaling issue:

  • 1,000 photos in the queue: ~1 second per photo
  • 2,000 photos in the queue: ~2 seconds per photo
  • 10,000 photos in the queue: ~10 seconds per photo
  • 35,000 photos in the queue: ~35 seconds per photo

This suggests a linear degradation in processing times, which seems unusual, especially since the GPU processing should handle parallel tasks efficiently. It feels as if Topaz Photo AI is becoming overwhelmed, possibly due to how it manages memory or handles the queue internally.

The main concern is that the program should ideally be able to maintain consistent processing times regardless of the batch size, to ensure workflow efficiency. As it stands, the current situation forces users to work with smaller batches to avoid significant slowdowns, which is not ideal.

Could you please provide any updates on your investigation into this issue? Additionally, if there are any workarounds or optimizations I could implement on my end while you work on a more permanent solution, that information would be greatly appreciated.

Thank you for your attention to this critical workflow efficiency matter. Your efforts to optimize Topaz Photo AI for all users are very much appreciated.

Warm regards

PS. the issue remains using 2.1.0

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.