Library module LrC bottlenecks

[Mulder]

Lightroom Version Number

14.0.1

Operating System

1. macOS 14 Sonoma

Here are my bottlenecks when working in Lightroom: 1) Library module 1:1 Preview Generation 2) DNG conversion 3) Denoise AI.

LrC 14 from subscription runs on MacOS 14.7.1, Hackintosh Gigabyte Z490 UD + i9-10850K OCed + Asus RX 6600 8GB.

If bottleneck #3 is clearly GPU-dependent, I wonder how modern CPUs (including latest Apple SoCs) handle bottlenecks #1 and #2. My rig is stalled at 100% CPU load (full 20 multithreaded cores) when generating 1:1 previews as I need them for fast culling on the 2-monitor setup (one screen for the whole image and another one for 100% magnification).

The same heavy CPU load goes for RAW→DNG conversion.

Is number of cores crucial for the machine to work efficiently? Is there anyone running the same tasks on a machine with enormous horse power (64 or 128 cores) and not having the bottlenecks I've mentioned?

 

[Jim Wilde]

I'm running those tasks on an Apple MacBook Pro with the M2 Max chip. It has "only" 12 cores - 8 x Performance cores, and 4 x Efficiency cores.

When running 1:1 preview generation, that largely depends on the Preferences>Performance tab setting for "Generate Previews in parallel". If that option is enabled, the 8 Performance cores are pretty much maxed out, but the 4 Efficiency cores are busy but not maxed out. The system still remains responsive, so browsing etc is still perfectly fine. I have no complaints about the speed of the preview generation.

If the option is unchecked, CPU usage drops significantly whilst time to generate the previews goes up significantly. I always have that option enabled so I get the best speed and my system can handle the increased CPU load.

It's a similar story with DNG conversion, though recent changes have made that process very much faster than it used to be.