3

Example:

  • Select 10x2GB files, then drag and drop to another location.
  • Drag and drop each 2GB file to the location individually, in succession (10 times).

I’ve tested this across multiple OS X versions and Apple computers, and the latter example always executes faster.

  • Why’s the drag and drop sequential fashion faster?
  • That said, why doesn’t Finder just execute batch operations in the same manner? I must be missing something(s)…

Note: This is a question that teeters on the edge of not being within the bounds of this site, yet nonetheless, one worth asking :)

The results are the same regardless of the order (batch/manual) in which it's performed.

Improve this question
10
  • Specify in your original question if you are running a multi-processor:  > About This Mac > System Report... > Hardware.
    – dan
    May 15, 2016 at 8:34
  • Specify in your original question how you avoided between your 2 first copies to take advantage of the system cache?
    – dan
    May 15, 2016 at 8:41
  • @danielAzuelos Yes…quad core 2.6Ghz i7. Same result on a dual core iMac from 07.
    – njboot
    May 15, 2016 at 8:44
  • Then if your system is nearly empty (just the Kernel, the Finder plus a few screenfull of standard daemons) parallel operations might perform nearly 4 times faster than sequential ones.
    – dan
    May 15, 2016 at 8:48
  • 1
    Why have you not answered daniel Azuelos question, "Specify in your original question how you avoided between your 2 first copies to take advantage of the system cache?"? Are you rebooting between the batch test and individual files test? Are you copying just to another place on the internal drive or to an external drive? If external, what type is it?
    – user3439894
    May 15, 2016 at 13:59

1 Answer 1

Reset to default
1

I think in your example you are copying all files to the same disk. If the finder is trying to write to multiple places (or reading from multiple places) on the same hard drive, it will be slowed down by drive access times. It is easier (faster) to read (or write) one whole file at a time rather than ten files at the same time.

You can observe similar behavior if you are compressing video to a smaller codec. If you choose to export to the same disk that you are reading your source file from, your export will go much slower than exporting to another drive.

You could re-do your test using 3 separate disks (provided those disks were not being accessed across the same bus (USB3/Thunderbolt, etc.). That might be interesting to test.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .