• Global community
    • Language:
      • Deutsch
      • English
      • Español
      • Français
      • Português
  • 日本語コミュニティ
    Dedicated community for Japanese speakers
  • 한국 커뮤니티
    Dedicated community for Korean speakers

High memory usage

Participant ,
Feb 20, 2024 Feb 20, 2024

Copy link to clipboard


I have been having a lot of trouble with both CPU and GPU ram allocations when trying to use Lightroom, Photoshop and Topaz apps in the workflow: Edit in Lightroom, then from there, do an "Edit in Photoshop" and, from there, call Topaz Sharpen AI or Topaz 2 as a filter. This is a nice workflow. Calling Topaz as a filter on a layer from Photoshop means when it returns to Photoshop, I can use Photoshop opacity and masking to apply the Topaz changes where I want. And when I save in Photoshop, the new TIF image gets added automatically to my Lightroom catalog. There is only one thing wrong with this workflow: LR and PS are memory hogs. They take and take but never seem to release again, causing Topaz apps to fail or at least work very slowly.


First some info: I am now using a new PC, Windows 10, 64 GB motherboard RAM, Ryzen 9 7900 CPU, and a 12 GB Gigabyte RTX 3060 graphic card. I have Lightroom use of graphic processer set to "auto", and Photoshop is also using the graphic card for processing. Topaz apps use the graphic card too. For Sharpen AI, there is the option to insist on using the CPU only but that is really slow.


On my previous PC (which I just migrated from) I had only 32 GB motherboard ram and that was simply not enough for the above workflow. I had to shut down Lightroom if I went to Photoshop with the intent to use Topaz apps. But with my new PC, with much more modern tech, I am still having problems.


There are two things that I notice. I wonder if someone has found a fix for these, or if Adobe would please do some development here to fix these. The first is that when Photoshop gets allocated memory to work on an image, it does not release it again even if I close down all photos in Photoshop. Here I need to shut down Photoshop completely. So it seems once this memory is grabbed by Photoshop, it will not be available for Topaz (or my Spotify, or anything else) until I shut Photoshop down completely.


The second is that if I set the Preferences in Lightroom to make use of the GPU for processing, I get really high performance with Lightroom tasks (making previews, building masks, etc). However, Lightroom then takes a claim to a huge percentage of my GPU ram. And never releases it again. Photoshop also takes a hold of a good bit of the GPU ram, I notice. But the Lightroom claim on GPU ram is the biggest problem for me. It is this which I believe causes Topaz apps to fail with either a crash-silent-shutdown (Topaz 2) or an error (Sharpen AI).


My move to 64 GB instead of 32 GB motherboard ram was a blessing with the new PC. But the 12 GB ram I have in my GPU doesn't seem to be enough to support my desired workflow. And looking at the prices of graphic cards, I just do not want to purchase one with 16 or more GB, they are far too expensive.


I have a few options. One is to go back to the workflow I used last for my older PC, and just get used to breaking the workflow chain and shutting down Lightroom when I go on to edit in Photoshop. This works but I then have to sync my folder after to find my new TIF file from Photoshop to add to my catalog. Another is to set the Prefernces in Lightroom to not use the GPU (Lightroom will use it anyway for its newer denoise feature, but this doesn't seem to hold onto any GPU ram). And the third option is to wait patiently for Adobe to add some kind of dynamic memory management. Perhaps I should add, it is not the high memory needs that I am concerned about, but the apparent hold on the GPU ram. To my knowledge, Windows does not free up GPU memory by paging to hard drive like it does with motherboard ram.


But any tips, workarounds, advice would be so nice!







Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
no replies

Have something to add?

Join the conversation