Please process images simultaneously in DxO PureRaw on MacOS. Trying to batch process right now is painfully slow as each image processes 1 after the other with very low resource utilization. Processing 8-10 images at a time would significantly cut down on processing time and make the product much more useful.
PureRAW uses DeepPRIME NR, which is extremely resource intensive. You can adjust the number of images to be processed in PhotoLab 5 but, due to the limitations of all but the most powerful of computers, you don’t really save much time.
I think the process could be optimized if PureRAW would accept further input in the development cue while it is already working on images.
The implementation of such a possibility would also not be too complicated, since it is already able to manage user interactions with the UI while the process is running.
Where do you see this?
I totally agree…
I’m sitting waiting on 400 images…
cpu %20
GPU %3
memory %20
disc %0
My machine could probably process 5 images at once and make the process 5x faster…
400 images is like 3 hours worth of processing time…
with parallel processing, it could be 30 minutes…
As Losart3d said, it is in Activity Monitor during batch processing. With an M1 Max I see about low cpu usage, and very spotty gpu usage.
I really wish this was an option. on a 10900K CPU w/3090 GPU, my 20 cpu cores never go above 15% (with only one core getting stretched to about 50%) and my gpu goes 30% then 0% then 30% when the next image is being processed. Seems like this would benefit from parallelism.
There’s an old saying that 9 women can’t give birth to a baby in 1 month. Parallel processing needs to be carefully implemented to be effective, but I’d be surprised if the developers hadn’t experimented.
My guess is that the application is written to support a wide range of older and newer machines with varying levels of computing power and memory.
As already suggested, if it’s feasible for high-end users, why not make it an option with the caveat that “Your mileage may vary” to prevent a flood of customer support issues?