Updating your PC is good but don’t donit just for PL9
I updated my Laptop last December
I9 14th gen 24 core RTX4080 64gb RAM.
It made PL8 lighting fast…
But PL is slow as a 
Wait for a steady PL9 before you judge
2 Likes
Its very slow when using heic its near unusable the times it keeps re doing the preview. RAW is done by the speed of light, well after seeing heic being done. Closing the program when opened on a folder with heic takes for ever as it looks like it re renders before closing which is mad. Close on a RAW folder and it closes.
2 Likes
Seems ok on M4 iMac with 16 gig.
Sadly I already have these options disabled and it’s still slower than I’d like.
Something I have noticed is… I just applied a batch of edits to ~90 images in a folder. Rather than generate previews for all of these in the background, PL9 is still generating previews as I click into each image - another slow-down.
Again here I’d like the option to force generation of previews for everything in a working folder (or selected images) - either in the background or at the press of a button - so that I only need to experience one long wait, rather than ~90 small waits.
I’ll go and make tea or something.
1 Like
This is a PC that meets the specks
I agree! PL 9 is too unstable on too many platforms. And nothing about DXO fixing the problem. Don’t blame it on Windows, AMD or Intel or Nvidia.
3 Likes
I bought PL9, I should have waited…
I have a windows pc with a nvidea graphics card and am happy with photolab 8 so wont be trialling photolab 9 until i read these issues are resolved.
2 Likes
Exactly! I don’t know how stable nVidia drivers are now, because I’ve been using Radeons for a long time and haven’t encountered any problems in other applications that use hardware acceleration. I just use the latest drivers and it’s very stable. Everything except PhotoLab. I have the impression that even though the new features are really great, they just haven’t been tested properly…
3 Likes
I too have experienced problems using AI Sky and People masking in PL9.
My Win11 machine spec…
AMD Ryzen 9 3900X 12 core CPU
64 GB DDR4 memory
NVIDIA GeForce RTX 2070 Super 8GB GPU
Win11 Professional 64-bit build 26100
I updated the GPU driver to the most recent 581.29
Editing a photo and using AI Sky mask resulted in an ‘Internal Error’ with 7GB of GPU memory used.
Restarting PL9 release lots of GPU memory.
Tried to edit the same photo using AI People mask which produced the same error.
After reading all of the posts I decided to install GPU driver 572.83, doing a clean install.
Editing the same photo and using AI Sky mask worked with 7GB of GPU memory used.
I then added AI People mask which also works - still 7GB of GPU memory used.
I then exported to photo to .jpg which worked, taking 30 seconds.
I hope the above information helps others.
I’ve not tested PL9 extensively but at least I seem to be able to do some basic AI masking without encountering errors.
Also a big thanks to others that have helped me get to this stage.
Good luck!
2 Likes
I installed version 572.83, and as you experienced better stability and it mostly works OK, but rather slow to export, especially with AI masks involved.
The funny side of the matter, I just got the advice from support to install the newest driver for my NVIDIA GPU… The one I used was from March and very old.
Ah well I spoke too soon!
When I start PL9 again and select a photo with People and Sky AI masking set,
I get an Internal Error (Correction failed to execute stage) when the program
attempts to show ‘Full preview’.
At this stage, I can disable the two masks (using the eye symbol), select another photo, go back to the problem photo and then re-enable the masks, one at a time.
Running the test again with the masks enabled, sometimes it will open the image without erroring, and sometimes it will produce the same error.
Very random!
In my experience, PL9 needs a restart after I get the Internal Error message. The same photo will often get exported without error after that.
1 Like
I have the same experience - a restart is needed after every Internal Error message.
I’ve taken to just not using AI masks at all while trialling PL9. The Auto-Brush function works a lot better than in previous versions, so that is a benefit.
PL9 sorely needs fixing though. 2 weeks since release and I’ve not seen any signs of an update on a fix, far less a fix.
1 Like
I’ve seen a lot of speculation here about DxO PhotoLab 9 crashes being mostly due to limited system resources (like not enough VRAM). That can certainly cause problems - running close to the minimum requirements will always increase the risk of instability - but it doesn’t explain all cases.
My system has 64 GB of RAM and an NVIDIA GPU with 8 GB of VRAM, so resources shouldn’t be an issue. Yet PhotoLab 9 still hangs or crashes during export, even if I only apply a single small mask. This isn’t a case of pushing the system to its limits - it happens under very light usage.
As far as VRAM and system resources go, the whole point is that they are supposed to be used. You want base models and data loaded into RAM or VRAM so they don’t have to be constantly unloaded and reloaded. Having resources utilized is expected and desirable. So using up memory on its own doesn’t necessarily indicate a problem. A lot of the speculation around resource usage feels more like a misunderstanding of how things actually work.
I’ve been working in the software industry for over 25 years, and I know that bug-free software doesn’t exist, especially with complex desktop applications that have to run on endless hardware combinations. Still, it’s disappointing that the first impression is marred by instability. I’m running a standard platform (Windows 11 and a mainstream NVIDIA GPU), so it’s not some exotic setup.
So yes, not enough resources can explain some crashes, but there are also bugs and implementation issues in PhotoLab 9 that need to be fixed. I really hope they’ll fix this shortly - it’s part of professionalism, and the new AI masking looks really cool and useful. Masking has always been one of the weaker points in PhotoLab for me, so getting this right would be a very handy change. I’m definitely looking forward to this being fixed, because it could be a great functionality. Congrats to the team for the release regardless.
I can confirm, that limited system resources are not the problem:
I use PL9 as PL8 without a external GPU Unit on a Ryzen 5700U with Radeon Graphics 512MB VRAM, 64GB RAM and Windows 11 Pro 24H2.
So far I have no chrashes, DeepPrime XD2 needs about 4-6 min for a 50MB nef-File, a little bit longer as in PL8.
The AI Mask in Selection Mode and the predefined masks are really slow, the Area Mode with drawing a selection box is quite useful.
1 Like
Let me speculate a little: It could be that the VRAM on the graphics cards has not yet been “eaten up” from Photolab to the last bit. With the use of AI, as much of the VRAM as possible is now being used by the software. Memory management is in extremely high demand. Problems with crashes are often hidden in these extremes.
Running Win11 on AMD Ryzen 7 3700X 8-Core, 64Gb RAM, GeForce GTX-1660 6Gb.
PL8, Capture One, Affinity Photo run very fast on this setup. C1 and AF AI masking is also fast.
Installed PL9 and AI masking didn’t work. Regressed to NVIDIA driver 576.52, and it is now working but everything is very slow. The more I use AI masking the slower it gets. Did some tests between C1 and PL9 for AI subject masking and PL9 is x3 slower. And gets incrementally worse. Other actions such as cropping takes for ages with ‘full preview in progress’ …
I have reverted back to my PL8 workflow, with noise reduction in PL8, DNG to C!, and local editing in C1. Will await performance fixes in PL9 …
1 Like
That could be. However, if that is so, DXO should have warned about PL9 not working on anything else than a 10000$ machine… 
1 Like
Thaat iiis slooooow…