Had an update to PL 6 last week and now denoising a batch of photos takes hours, and when nearly complete I get a message ‘Your computer has run out of application memory’, followed by a crash that loses all the work. I have deleted several applications but I think it’s PL that is the culprit. Anyone had similar problems? I will have to ditch DXO if this goes on because it’s unusable.
I’m seeing the same thing after a recent PL6 update. I’m not sure if it was last week, but it is in the last couple weeks.
Version 6.18.0 build 74 seems to be a problem.
MacBook Pro
M2, 16GB mem
OSX Sonoma 14.5
With no other changes, I noticed after the last PL6 update, when I export photos, there is a memory leak. It seems related to the number of photos in the batch vs the setting for number of simultaneously processed images.
This is 100% repeatable for me.
Select 38 images → export jpeg 95%, I can watch swap (via activity montor) just start climbing steadily from 1GB → 40GB at which point I get notification about computer running out of memory and then the laptop becomes unresponsive to mouse/keyboard and a few minutes later the laptop restarts.
The crash is right around image 20-21. If I set simultaneous images down to 4 and repeat with the same photos, the crash happens at the same spot ~image 20-21, it just takes longer to get there.
Fix for me has been to select less photos in the export, but this is quite a workflow hindrance as I have to keep babysitting to start the next batch vs letting it run through all of them while I do something else.
I do use DeepPrimeXD. I haven’t tried a lesser DeepPrime setting (or off), but I really don’t want to, it is why I bought PL in the first place.
Is there a way to roll back versions? There is nothing I really needed in the last update. Up to the last version has been troublefree for me.
Very interesting - sounds exactly like what’s happening to me. I’ll try reducing the batch numbers. I’m only using Deepprime and not XD but I might as well if all are affected. I wish there was someone at DXO who would have the understand to deal with this, but my previous experience of their support is not encouraging. Meanwhile I’m tempted to move to Topaz.
Can confirm this issue.
PhotoLab does not terminate the XPCCore processes when exporting has finished. After using about 3/4 of available RAM, further exports need swapping. Only when I quit DPL are the processes terminated and the memory freed.
Second export. Swapping is active
PhotoLab 6 terminated, memory pressure is reduced
Same issue in DPL 7.8, DPL 7.7.1 seems to be okay in this respect.
I have created a ticket for DxO support.
Also check out this topic:
2 Likes
It looks like PhotoLab 6.18.1 could solve this issue.
Suggestion #1 - call Apple Tech Support, and they will tell you what to do, as they share your screen. They have several tools that might lead to a solution.
Phone: 1 800 MYAPPLE
Suggestion #2 - Do you really need to edit so many images simultaneously? Maybe try just editing fewer images simultaneously. But if they see this happening, they may have other suggestions. On my MacBook Pro, M2 chip, Apple found things to change to free up more memory.
If the regular advisor can’t help, ask them to put on a “Senior Advisor”. That’s what I needed to do.
@mikemyers , updating to PhotoLab 6.18.1 should eliminate the problem. It did on my Macs.
Looking at how memory usage develops while exporting images clearly indicates what is happening. Again, updating to PhotoLab 6.18.1 should be all there is to fix the issue.
Thank you; that’s what I am now running. When PhotoLab says they have an update, I’ve always just installed it. Maybe I should wait a while, looking for negative feedback.
Thanks all for the replies. I am also going to start archiving a few copies of DXO going forward, that’s a great idea.
6.18.1 build 75 has fixed the problem for me. I can see it releasing memory as it goes vs just going up and to the right until boom.
I also did submit a ticket for this, it is my first time using DXO support. I am very happy with the speed that a patch was released from DXO support (same week).