This project has moved and is read-only. For the latest updates, please go here.

VeraCrypt as a performance tool?

Topics: Users Discussion
Sep 24, 2015 at 9:34 PM
I tried an experiment where I copied 4600 small files (average length 1K) to a flash drive. This took 309 seconds, which is abysmally slow, but not unexpected, given that flash drives are notoriously poor performes for small file writes.

I then tried creating a 100 MB FAT container using VeraCrypt 1.14, and mounted it, and copied the same files to the mounted container. The time was 2.5 seconds, or about a 120-1 speedup.

The read times through the container are also better, but not dramatically so, for example about 2-1.

I'm a software engineer, and I understand where the dramatic differences might be coming from -- issues of buffering and caching and OS overhead and hardware latency and so on. Coalescing lots of small files into one big one is typically always a major improvement for flash drive performance.

But the speedup still seems dramatic, and it seems like VeraCrypt could be a useful tool even if a person is not particularly interested in encryption.

Does anyone else have experience with this?
Sep 24, 2015 at 11:34 PM
Thanks for creating this very interesting experimental result! This may well be worth publishing in the VeraCrypt documentation... :)