I have not been updating the table as it was made using LRC 12.4 and there were changes to times when Adobe release 13.* It would be good to see what an Intel GPU does though, maybe add it as a footnote. Thanks ...
I was curious to test this file. I have LR 14.0 release candidate. I'm using M1 Max 16" MBP with 32GB RAM and 24GPU.
Just wanted to know what your procedure is for the test? Are you running at the default setting of 50? Are you stopping the timer when the progress bar disappears.
Thanks.
I did just run it at a setting of 50 and stopped the timer when the progress bar disappeared and got ~26s.
arbitrage wrote:
I was curious to test this file. I have LR 14.0 release candidate. I'm using M1 Max 16" MBP with 32GB RAM and 24GPU.
Just wanted to know what your procedure is for the test? Are you running at the default setting of 50? Are you stopping the timer when the progress bar disappears.
Thanks.
I did just run it at a setting of 50 and stopped the timer when the progress bar disappeared and got ~26s.
pr4photos wrote:
How did you get LR 14.0? I'm on 13.5
A few years ago I was complaining about things on the Adobe forums and the admin gave me access to the LR Pre-release program to test potential fixes they were working on and I've had access ever since. So I will typically get many dot versions of a main version before the next main version goes out to everyone.
arbitrage wrote:
I was curious to test this file. I have LR 14.0 release candidate. I'm using M1 Max 16" MBP with 32GB RAM and 24GPU.
Just wanted to know what your procedure is for the test? Are you running at the default setting of 50? Are you stopping the timer when the progress bar disappears.
Thanks.
I did just run it at a setting of 50 and stopped the timer when the progress bar disappeared and got ~26s.
Yikes! ... it's been a while.
Pretty sure we stopped the timer when the image showed up, but that's probably only about a second after the task bar completes. Pretty sure we used default settings but I also used a new catalog (no files other than the test file). And also if you want/need to run it again either clear the cache or reboot LRC.
EDIT : Sorry, I had to edit the post (deleted last para) as I commented on times for Denoise in the GPU vs Time graph I'd posted. That was using Art's data for his 36MP Nikon file, not the 60MP Sony file we use
PS. If 26 secs is right then that's a good improvement over LRC 12.3. As you see from the table the same configuration of M1 Max and GPU cores, but with 64GB ram came in at 37 secs (data from my own 14" MBP).
