Windows 11 Pro, 64GB RAM, Nvidea GTX 1650 Super, AMD Ryzen 9 3900 XT. Mainly processing Nikon NEF files from D750, D7200 & Z7 II.
Using DXO V7, I would have the occasional file fail to save with the error message “Internal error (Correction failed on the Execute stage)” With Version 8 it is annoyingly common, more so if I use DeepPrime or DeepPrime XD/XD2s. I can edit the same files in V7 and they save without any issues. Or I can not apply either of the above and save. Then if it is ISO 4000 or above I end up having to use use Topaz Denoise. Obviously not ideal.
At the moment it is bad enough that I would not upgrade to V8 as the gains are outweighed by the repeated. Hopefully an update in the next 20ish days will resolve things
@Steve-Selwood I have been doing some tests with various combinations of the machines (5900X, 5600G, i7 4790K) and GPUs ((RTX 3060 and RTX 2060) ) at my disposal.
I swapped out the RTX 3060 from my 5900X to conduct tests on a (now redundant) i7 4790K and then conducted tests with the RTX 2060 with the 5900X and part way through the export I encountered this
The version of PL8 on the i7 was a trial version (now lapsed) and up-to-date and on the 5900X it is still PL8.0.0.
I reran the (complete) test again immediately after the failure and it was fine!?
I had made a mess of the image numbering when I created the batch at the point where the 2 failures occurred but as the exports show I have used this set of images on a number of releases with no problems
PS:- I have put in a support request related to my incident and I believe that it would be useful for you to do the same.
The reason for that suggestion @Steve-Selwood is that my incident has been a one off so far, i.e. the chance of me being able to reproduce it for any additional diagnostics are remote, whereas the chance of you being able to reproduce it are higher.
The images I was using were the Egypt image from the benchmark site times 60.
I have a batch of 40 images, created from .NEF images (D850) from the benchmark site (spreadsheet) so I will try a test or 2 with those images to see how they fair.
PPS:- The batch of 40 was fine so I created a batch of 200 which is currently processing.
The batch did not fail
If you have an image that you think might cause a failure please upload with the DOP and I will create batch and see if I can “break” PL8