An inverted question: Are there any systems and GPU.cards that works with PL 9.1 or 9.2

DxO PhotoLab
1

We have seen for months reviewers on Youtube that have been reviewing the new Photolab 9 seemingly without any problems even using the new premade AI-masks and al the new different features without any problems.

I want to know their secret! What do they have that I and so many others seems to lack.

I would very much like people that have no problems at all with any of the AI-funtions what kind of gear they have.

For a change it would be nice to know what actually works and not just the things that don´t

3 Likes
2

Macs?

3

There are Macs too with issues with this version but Apple is a protected workshop and a much less complex and much more limited environment than the Windows-platform and in some cases that can definitely be an advantage.

I´m interested of the ones that don´t have any problems even using the premade AI-models in the drop down menu.

4

I ran a trial of PL9.1 without any problems at all using AI masks. Win 11, AMD CPU, 32GB RAM, RTX 3080 and the older nVidia driver that DxO was recommending. (I installed that when it was new, never updated it.) My trial’s over now, and I won’t be upgrading until around Black Friday.

1 Like
5

Now on PL9.2 with latest nvidia drivers (581.8).
No problems related to masking (including preset masks).
No errors with export, though it does seem slow (2 - 30 seconds for DP3).

I do have the occasional crash though, not sure why, my suspicion at the moment is something to do with stacked images.

Win11 Pro
Core i5 4600K
32 gig ram
Nvidia rtx 4060 with 8 gig

… so my PC is far from leading edge, but so far so good. I did have a lot of problems with 9.0 and earlier nvidia drivers (ai masking wouldn’t work at all, auto brush was a mess, etc).

Still on the trial, hoping for BF offer, but loving it at the moment.

6

Can anyone with an RTX 5060 8G card running Windows 11 say that PL9 works stably, and if so, can they share which driver version they are using?

7

It is a card in the RTX 50xx-series so I suggest you download the latest trial of Photolab (9.2) and test it with the latest driver and if that doesn´t work properly install 9.1 instead. I have definitely problems with 9.2. If the latest driver is the problem revert to the one recommended by DXO.

No one else has your specific computer, so you have to do this yourselves.

1 Like
8

I downloaded the PL9 trial and couldn’t be happier. This is my first go with PL, my only DXO experience until now is with Deep Prime. Experienced one crash, but no idea what caused it. I’ve run hundreds of images through it, including batch processing of almost 200 at a time. This is a fairly new computer that was built for me in the spring. The processor is the Core Ultra 7 265K, part of the Intel’s Arrow Lake group, which has a pair of Intel’s third-gen Neural Compute Engines, which Intel says has AI performance of 13 TOPS. Supposedly the first Processor built for AI home computer. All I know is it’s fast as heck.

Intel Core 7 Ultra 265K processor
Asus Prime Motherboard, can’t recall which
G.Skill Ripjaws S5 32 GB (2 x 16GB) DDR5-6000 CL36 Memory
Inland QN450 1TB M.2-2280 PCLe 4.0 X4 NVMe SSD
Sparkle Guardian OC Arc B570 10 GB Video Card
Montech X3 Mesh ATX Mid Tower Case

I can tell you it rips through Deep Prime 4 applications in seconds, and AI masks show up instantly if the area is recognized.

All I’m waiting for is the Black Friday sale!

9

Seems only the GPU VRAM amount important in AI masks (and in the case of nVidia driver version also important, but seems we over that days) and all related things to ‘not crash’- If you has more GPU VRAM (like 12-16) - crash not really come and works fine. 6-8GB GPU VRAM also can be quite fine if you Manual’ AI masks (and not ‘Pre-defined’ AI masks) and for Export DP3 used (and not DP XD2).

All based in my previous measurements (not fully exact values, but give an idea) checked in some previous PL. 9.x version - but i think its still valid at general. May some value is not correct (especially in pre-defined AI mask and Export)

PL has “two part”:

  1. DxO.Photolab.exe → Main application
  2. DxO.Photolab.ProcessingCore.exe → Its does the Export

PL Main app

  • PL ‘main app’ itself usually use somewhere 0.3-1+GB GPU VRAM
  • If Deeprime rendering: ON than depends on what NR you select (DP3 or DP XD2). In the case of DP3 its like GPU VRAM 1.5GB. If XD2 than its a bit more than DP, like +0.5-1GB GPU VRAM. Its may also depend on Zoom level (may 1:1 use more than ‘Fit’)
  • ‘Manual’ AI mask: AI model loaded to GPU VRAM may takes like 2-3GB GPU VRAM
  • ‘Pre-defined’ (keyword) AI mask (like Sky, etc.) → frankly i have no idea (i can’t test it) - but as i read forum colleagues post/comments i try to ‘re-calculate’, but seems somewhere 3-5GB GPU VRAM, may it has a ‘Peak’ that more than this.

PL Export app:
Its also depend on what NR selected. Like in DP3 use less than DP XD2. In the case of DP3 may use like 1.5-2GB.

Windows and other apps
Not forget, Windows and other apps like Web browsers also use some GPU VRAM. Usually something like 0.5GB

So, do some math in few examples. (GPU VRAM, and best/average cases, not exact values/sum)

PL client + ‘Manual’ AI mask + Windows/Other apps
→ 0.5+2.5+ 0.5 → 3.5GB VRAM

NR: DP3
PL client + NO AI mask + Export + Windows/Other apps
→ 0.5+0+1,5+0.5 → 2GB

PL client + ‘Manual’ AI mask + DeeprimeRendering +Windows/Other apps
→ 0.5+2.5+ 1.5+0.5 → 5GB

PL client + ‘Manual’ AI mask + Export +Windows/Other apps
→ 0.5+2.5+2+0.5 → 5.5GB

PL client + ‘Manual’ AI mask + DeeprimeRendering + Export + Windows/Other apps
→ 0.5+2.5+1.5+2+0.5 → 7GB

PL client + ‘Pre-defined’ (keyword) AI mask + DeeprimeRendering + Export + Windows/Other apps
→ 0.5+4+1.5+2+0.5 → 8.5GB

NR: DP XD2
PL client + ‘Pre-defined’ (keyword) AI mask + DeeprimeRendering + Export + Windows/Other apps
→ 0.5+4+3.5+3+0.5 → 11.5GB

May its give an idea.

In spoken words:

  • If you has only 4GB VRAM → you can use ‘Manual’ AI mask, but cant export with DP3 → as with export its required like 5.5GB
  • If 6GB GPU VRAM: ‘Manual’ AI mask + DP3 export works → As its use like 5.5GB → But you may can’t turn on ‘DeeprimeRendering’ and/or may can’t export with DP XD2 as its needs more than 6GB.
  • With 12GB GPU VRAM you can do pretty much everything, with 16GB GPU VRAM your a most safe side.
  • Pre-defined (keyword, like Sky) is the real ‘killer’. If you not use it, than far-far less ‘internal error’ (if any) raised. We already talk a tons about that in other threads/comments.

Others:

  • Standard (previously called HQ) NR not use GPU VRAM (as i see, or near nothing) - in export and also if “high quality preview” is turned ON.
  • Export performance (with AI masks, DP NR) seems depend mainly (near only) the GPU performance - not just GPU VRAM amount, but GPU clock speed, GPU VRAM bandwidth, etc. So, i think talking about ‘export speed’, doesn’t really matter as it’s ‘as it is’. Its Mpixel/second → more Mpixel photo → more processing time. If you has "faster’ GPU, than its “faster”. Ofcourse AI mask amount may count and Driver quality (driver performance - may nVidia only) also can count.
  • Manual AI mask ‘performance’ seems more-or-less linear. Means: If you have 5 ‘manual’ AI mask the processing speed like 5 ‘unit’, if 8 mask, than speed like 8 ‘unit’ (‘unit’ based on your computer/GPU performance). Seems ‘Manual’ AI selected (mask) area not really make a difference (mask area large or small). May complexity count, but very little → like one mask not 1 ‘unit’ but 2.
  • You can kill the DxO.Photolab.ProcessingCore.exe (the export process, if not used), its restart itself. May its helps out in the few cases.
  • PL app not really free up GPU VRAM. AI model loaded to GPU VRAM. May first seems to good idea if PL ‘release’ GPU VRAM, but in the end its can’t made better as when you export, its re-used again (at least my standpoint of view)
  • PL 9.2 release notes say like: minimum GPU VRAM 6GB for/with DP3 export and AI mask. Its NOT say DP XD2 and NOT say ‘pre-defined’ AI masks. With 6GB GPU VRAM you may ‘on the edge’, that’s why release notes say for ‘optimal’ 8GB GPU
  • In ‘CPU only’ mode everything works - of course as computer CPU performance.

Side observations:

  • Currently seems Export may run to few error if more than around 20-30 photos exported with a few ‘pre-defined’ AI mask and DP NR, like 2 error from 20. Based on other forum colleagues observations.

I would very much like people that have no problems at all with any of the AI-funtions what kind of gear they have.
For a change it would be nice to know what actually works and not just the things that don´t

I hope my comment helps on that. If you has 12-16GB GPU VRAM, than at general ‘no problem at all’.

What GPU type (‘gear’) used - IF its has enough GPU VRAM - doesn’t really matter regarding ‘no problem at all’. Its doesn’t matter its Nvidia or AMD, its doesn’t matter if GPU is very old and slow and under under ‘minimum specification’ regarding type/chip IF it has enough GPU VRAM. Of course more advanced GPU chip, faster GPU clock, faster GPU memory bandwidth speed up things. But that’s all.

1 Like
10

@andras.csore Sorry but sadly not a true statement. There are undoubtedly problems with AI and your analysis seeks to show the issues, but I do not believe that is the whole story.

DxO have changed the export mechanism between the two releases and I have made tests with NO NR and NO AI that have failed during export!

So there are issues caused by using AI when actually doing the editing and then when exporting those images but there are also issues with the export process itself that are causing export failures that have nothing to do with NR or AI and where VRAM in use is way below the maximum VRAM available !?s

My tests have been done on a 3060(12GB) and a 5060Ti(16GB) and I have had this

image
image1920×1500 298 KB
.

What do your calculations say about these?

What about the change in export behaviour between PL8 and PL9

image
image1920×2199 328 KB

AI has upset the balance on PL9 but it is not the whole story by a long shot!

PS:- While some users have problems and others don’t that isn’t necessarily an indicator of anything, other than perhaps luck but more likely what has gone before!?

I have run tests where I “predicted/expected” a failure and it worked fine and I have then run another test as a step towards an expected problem run and that preparation run has failed.

These comments are restricted to the actual process of exporting and have been conducted without any AI whatsoever and with and without NR.

11

i5 14600K 32GB RAM 4070 Ti Super Windows 11 24H2

High-quality previews with DeepPrime enabled.

Runs fine, pretty smooth, quick exports (just a feeling, haven’t timed this).

Does seldomly crash or fails an export here and there, but works fine otherwise.

12

Yes, and i even mentioned that! in the ‘Side observations:’

Currently seems Export may run to few error if more than around 20-30 photos exported with a few ‘pre-defined’ AI mask and DP NR, like 2 error from 20. Based on other forum colleagues observations.

I not summarize all in details (what you find out and test in depth. Great job!) about this Export issue, but i think at least give a notice about that! I also note: ‘May some value is not correct (especially in pre-defined AI mask and Export)’.

This thread at general about ‘what kind of gear they have.’ vs ‘Internal error’ - and i try my best. I think i summarize this VRAM amount things a few times, but may that’s the first when i goes thru on that in more readable / easier to understand format. And at general i think its still valid.

PS:- While some users have problems and others don’t that isn’t necessarily an indicator of anything, other than perhaps luck but more likely what has gone before!?

Yes and not.

Not
First of all: regarding my measurements - not fully exact science, but at least some math - i’m sure its valid at general. This also supported with PL 9.2 release note GPU VRAM amount and DP3 + ‘AI mask’ (as release not not say ‘pre-defined’ AI mask and not DP XD2) and also seams users with more GPU VRAM tends to see less ‘internal error’ (as least few of them add a bit more description). So i don’t think ‘luck’ can be helps on that. BUT ‘In the edge’ on GPU VRAM may seems ‘luck’. I think, some cases its ‘in the edge’. Like calculated 5.5GB GPU VRAM requirement can be ‘in the edge’ in 6GB GPU in the real world. Some (like me) dont really care export performance. Looks like ‘luck’, but its math. Anyhow, seems now 'pre-defined (keyword) AI mask is the ‘killer’ for most of the users.

Yes
Yes, hard to know what exact others do. If someone export just 5-10 photos, may say: ‘works okay’. If someone export 30 photos seems encounter export issues. If someone not turn on ‘deeprime rendering’, use only ‘manual’ AI mask and DP3 may also say: ‘works okay’. I suggest a few time in this forum to try this and try that - but frankly i not receive too much feedback. So, hard to know.

13

I still can’t export any images after using AI, everything else works as it should.

There’s obviously an issue somewhere, for now I’ll just continue to use PL 9 without using the AI feature until it’s sorted.

I still think it’s dissapointing that it was released before it has been properly tested though.

14

What your GPU VRAM amount? 6GB? 8GB? etc.
May worth to check thru my comment in this thread a few comment before about GPU VRAM usage:

15

I run PhotoLab 9.2.1 without problem with a NVIDIA RTX 4060 8GB external GPU.
My computer is a HP Spectre x360 with 16GB RAM. Windows 10 Pro.
The GPU driver is NVIDIA Studio version 581.57, published Oct. 14th, 2025. NVIDIA says: “What’s Fixed: DxO PhotoLab 9: stability issues when using the AI masks”. I have not made extensive tests with AI masks, but up to now I had no error.
I have run several exports of about 400 images, taking around 45 minutes each, and without any error. The RAW files are from an OM-1, with several corrections, and XD/XD2s denoising.
It runs approximately at the same speed than PhotoLab 8. See results on the spreadsheet located here: PL4 GPU benchmarking? - #21 by jch2103 where I have tested several PhotoLab versions.

1 Like
17

Great news! Is this the “final” fix?

18

@Stenis Absolutely not, for you and your nice shiny machine, it might be but there are still problems lurking in the export process which you may never encounter but that doesn’t mean that they are fixed, just that you are lucky enough not to encounter them.

So PL9.2.1 is still not as reliable as PL8 when it comes to exporting non-AI images let alone what it might ot might not do when exporing AI images.

19

Ok Bryan, I will keep my eyes open.

But are you really having problems with your newest GPU 5060 Ti now?

So far I think my new machine handles Photolab 9.2.1 really well even with the predefined AI-models. It really feels like a relief to be able even to use the premade AI-presets. It is so much faster and isn´t the masking in the last version a little bit more precise? Sometimes it even seems to fix a sky properly at least with some trees at the horizons.

It has worked pretty good even with my old machine when avoiding the premade AI-presets and working picture by picture but now I also can use the AI-presets to more intelligently process series of pictures which will boost productivity a bit.

20

@Stenis No I am having problems with the export worker and its new structure I believe, i.e. I did a test with 1,000 images with NO Dops, NO edits, No NR and NO AI, i.e. effectively just export a RAW as a JPG, and it fell over (2 images marked as problems + a DUMP message) on image 615 on the 5900X with the 5060TI(16GB)

I have had other problems occurring with much smaller runs but those had edits and/or NR but no AI.

So the problem is not (necessarily) with AI but it is with the changes DxO have made to the export worker task, which is now a single program with a maximum of 6 export streams down from the previous 11 (or 12!?) rather than one export program per stream.

My concern has been that AI might add to the problems that appear to be inherent in the new export worker structure. That doesn’tt mean that we won’t encounter other issues with the new AI code but I am hoping that DxO will fix the worker and some of the differences users are noting with export times, i.e. PL8 versus PL9 which may well be a consequence of the new structure or the new structure + bug(s), will vanish!?

As for the quality of the AI I am afraid my image of Brighton Pavilion is still as bad as ever

image
image1920×1080 211 KB

Sky preset on the left and selection on the right arguably identical except not quite, for some reason

image
image1920×1237 354 KB

Please note that the Local Attributes applied to the SKY/sky selection are not there for artistic reasons but there to make the mask obvious in exported images!

The side by side and Beyond Compare comparisons were done on exported 100% JPGs.

21

An artistic reason can be a very good excause :slight_smile:

It looks like it still can use some improvements.