I was wondering if the GPU is actually used in any of the photomerge features? Because I'm not convinced it is.
I'm have several GPUs available to test from consumer low end cards (£25) to mid-end professional workstation cards (£450). I'd really like to try a high end gaming card...but think I'll give that a miss as I don't have access to one.
In running tests I've noticed a difference in some features, some rather noticeable, particularly with the GUI/ image response between the cards, so all is good there.
However, what I have noticed is not a great deal of difference, if any, in the photomerge features. Average processing times are the same, irrespective of what card is used.
GPU acceleration is enabled by the way.
And here's the killer: if I disable GPU acceleration, I get the same result (LR is restarted between changes BTW).
So hence that's my conclusion. I'd find it a little strange if it wasn't used. I thought something as mathematically intensive as merging images would be ideal for a GPU.
Anyone care to comment?
According to Eric Chan, the GPU is used exclusively in the Develop Module, and Photomerge isn't really the Develop Module
If I am understanding things properly, it wouldn't help with photomerge anyway, which probably requires CPU rather than GPU.
Thanks. It looks as though they haven't got that far yet.
On the plus side, it's still faster than PS by far.