Read Speed Raid 1 vs raid 0

The source you read is not quite correct. The problem with read aheads on RAID 1 volumes is the drive are constantly pre-fetching data. They do not know what to "pre-fetch" in this case. So when a driver attempts to "stripe" the reads on a mirror volume, there is a lot of lost efficiency because you cannot take advantage of "striped reads" effectively.

I would guess the author meant multi-threading, also, not multiplexing.

SoftRAID does have several proprietary techniques to enable mirror reads to be nearly as fast as stripes, however.

You should not be seeing mirror reads the same as single disk with SoftRAID over thunderbolt. Mirror reads should be significantly faster than single disk performance.

So I do not have an explanation for your results without more information.
We have to assume this is a freshly created volume in both tests?

Hi,
thanks for reply. Here the source http://www.diffen.com/difference/RAID_0_vs_RAID_1

I have a Lacie Thunderbolt, I switched to softraid as the old apple raid is no longer available.
I just installed two brand new Seagate 3TB disk.
No raid, read around 170-180 MbS
Raid 0 - Striping 380-400
Raid 1 - Mirroring 170-190 (peak 200)
So, in case of raid 1 I can see a marginal performance increase compared to the no raid config (JBOD)
What should be the performance with mirroring in your opinion?
The Softraid where configured for photo editing use

My testing shows:

Writes: 185.56 Reads: 292.46

This test was done on a MacBook Pro, 2TB Toshiba drives and an OWC Thunderbay enclosure, with SoftRAID 5.0.7 (driver is the same as 5.1) The result above is from an average of ten runs of AJA Performance benchmark tool, with file cache enables, 16GB file size and 4K frame rate. I obtained identical results in another Thunderbolt enclosure that was used for testing.

When mirroring, the optimization setting has no impact on benchmark tests.

I do not know why you are not getting better read results.