• Global community
    • Language:
      • Deutsch
      • English
      • Español
      • Français
      • Português
  • 日本語コミュニティ
    Dedicated community for Japanese speakers
  • 한국 커뮤니티
    Dedicated community for Korean speakers

What hardware upgrades to a Mac Studio (cores, GPUs, RAM, etc.) would help LrC face recognition?

Community Beginner ,
Jun 26, 2022 Jun 26, 2022

Copy link to clipboard


Having been caught on the wrong end of a major dealer's sudden shutdown, I am back to square one on buying a Mac Studio. I had previously ordered an M1 Ultra (48 graphics cores, 64Gb ram, 1Tb disk), but now that it looks like there will be a significant wait for anything but a base Max, I am beginning to rethink specs.


My biggest drain on the speed of Lightroom Classic has been the Face Recognition function, specifically "similar." I have been running this against 300,000 images with an i7 iMac (4ghz) and maxed out GPU for 2014 (4Gb) with my library on the internal SSD (PCIe), but the iteration between runs has been slow - so much so that I feel like I am never finishing this project.


That said, I don't have a good sense of how the Faces function is working - is it against Smart Previews? Regular Previews? Tiles used only for the purpose?


I do know having had all this time to salivate that on the new machine, the images will be on 16tb of NvME running at least 2500Mb/s - but I'm a little stumped on how to spec the actual processor (because I don't have a great understanding of LrC's core utilization (how many/what kind of cores does Faces actually use?), storage (I assume that more onboard is better, but 2Tb would be more than enough for the library database and previews), and memory.


Ideas? Right now, my two top choices would be:

- M1 Ultra retail (64Gb / 48 GPU / 1Tb)

- M1 Max BTO (64Gb / 32 GPU / 2Tb)









Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
no replies

Have something to add?

Join the conversation