Sync app tests

[TP109]

Below are the results for the Sync app tests. I dropped a few more apps from the initial list because of incompatibility issues, system driver problems, or because they just wouldn't run on my system. Because these apps take time to test, I didn't do multiple runs on each one. There is some variability between runs and the mode used (backup, copy, mirror, sync, etc), but because these tests just synced to an empty directory, variability should be under 10% regardless of mode.

Basically, these tests were performed to identify the fastest and lightest weight sync apps. Some of the results were surprising. Many only appear to be fast because they scroll files as they are processed, but that is deceptive. Most of the fastest apps don't even display what files are being processed and use progress bars to indicate the progress of the current file and/or of the overall process.

The fastest 2/3rds have acceptable speed, but the slowest 1/3rd are noticeably slower. Some of the apps that I expected to freeze or crash didn't. Also, many of them are surprisingly lightweight.

These were the test conditions:
No Compression. Backup, one-way sync/copy, or mirror mode, depending on app.
If scan and copy/sync are 2 separate processes, only copy/sync measured. If a single process, then total time measured. Pre-scan bypassed where possible.
Sync results verified with Sysinternal's DU command-line utility.
RAM usage measured on completion of the sync operation.
Source - 129,819,285 Bytes comprised of 141 directories, and 6786 files.
Destination - an empty directory on a separate partition.

Here are the charts. Lower is better for both.

[Midas]

Great work, TP109!

Just to extract the highlights, here are the referenced top and bottom 3 listings:

= Top 3 for Speed =

1. ozSync (https://www.portablefreeware.com/?id=1314)
2. SyncFolders (https://www.portablefreeware.com/forums ... hp?t=22039)
3. ZBack (https://www.portablefreeware.com/?id=1683)

= Bottom 3 for Speed =

1. Toucan (https://www.portablefreeware.com/?id=984)
2. Allway Sync Portable (https://www.portablefreeware.com/forums ... php?t=1646)
3. DirSyncPro (https://www.portablefreeware.com/?id=2643)

= Top 3 for Least RAM Needed =

1. Bart (https://www.portablefreeware.com/forums ... php?t=6556)
2. EzBackup (only distinct TPFC mention I could find: https://www.portablefreeware.com/forums ... 099#p78099)
3. Back4Sure (https://www.portablefreeware.com/?id=1684)

= Bottom 3 by Least RAM Needed =

1. Allway Sync Portable (https://www.portablefreeware.com/forums ... php?t=1646)
2. ozSync (https://www.portablefreeware.com/?id=1314)
3. FreeFileSync (https://www.portablefreeware.com/?id=1475)

[TP109]

@midas,

Correct, but I think you meant Bottom 3 by Most RAM Needed for your last category. (edit 11-09-16: Never-mind. I misunderstood the way you worded your category. Either way would be correct.)

I think the top 18 fastest apps are roughly equal. Variability could account for much of the differences between them although the differences could be more substantial for larger sync tasks, for differential backups, or for other more complex backup modes. As for the usage tests, lower is always better.

[TP109]

Because of variation between runs and from the test system, I will make multiple test runs for each app to obtain an average sync time. I've already done this for about 1/2 and there will be some changes, but nothing really significant so far. I want to be as accurate as possible though since I didn't do that with all the apps the first time.

[TP109]

Updated chart changes:
Updated SyncFolders to v3.4.111.
Replaced FreeFileSync 7.9 with v6.2 (as recommended by TPFC).
Included Shadowcopy after correcting VSS writer registry error.

There is some jockeying in position from the previous results - possibly due to errors. It's easier to make an error if each app is only tested once. For the following, at least three runs were performed for each app. App changes highlighted in yellow in App column.

[TP109]

Tested some command-line apps as a comparison. Just a table, no charts since there are only a few.

[TP109]

Updated one of the command-line backup tools and removed duplicates (highlighted in yellow). Re-ran each app three times to get an average. There is variation between runs averaging about 10%, but some apps have more variation than others. Placed the updated data table under the previous one for easier comparison.