[section_title title=”4.75GB Transfer”]
4.75GB Transfer
We have devised a method to test real-world performance via a scenario of a 4.75GB file transfer from my test bench SSD (Crucial MX200 250GB M.2 SSD) to the target device; the one on test at the time. The file itself contains a selection of music, multi-media files and word documents to give real world data; not synthetic files etc.
The test will be over the SATA3 (6GB/s) interface as this is the most commonly found on the latest motherboards/SATA compatible devices. So the test itself? Well, it involves transferring via drag and drop into the target drive from the test bench and the time taken will be divided by the amount of data giving us the average MB/s achieved. The lower the better of course!
Going from 1 drive to 2 in RAID 0 yields the biggest improvement with a 20-second jump. The same can’t be said between 2, 3 and 4 drives in RAID 0 however as only 6 seconds separates them. Still, going from single to RAID 0 with 2 drives will always offer near double performance due to the way the technology works.
I believe WD’s “Red” series HDDs have a dynamic balancing feature in the main bearings.
Also, when one of the drives in a RAID-0 array fails, it must be replaced and the
entire array re-built: this is not much different from the steps that must be taken
when a single “JBOD” HDD fails. We use a lot of RAID-0 arrays, for speed, and
we LUV them! Lastly, in future reviews, you might want to mention the ratio
of (price) / (warranty years). Most often, the 5-year warranties excel on this metric.
Why? Why only synthetic, worthless benchmarks? Of course 2 disks in RAID0 are going to have 2 times everything! We don’t need benchmarks to prove it, logic proves it!
Why can’t there be one benchmark on the whole Internet, of HDD RAID0 vs no RAID testing real-world scenarios, with a RAID controller (and comparing hardware and software controllers)? And by real-world I don’t mean copying speed as that’s as obvious as synthetic benchmark results.
I know this thread is olddd .. but the response above by Karol is brilliant .. there is a definite anti RAID thing generally these days… what I find hilarious is they state RAID 0 gives minimal performance increase. Then they go into RAID 10 and then talk about how much faster it is, and it’s awesome blah blah .. RAID 10 is by its nature slower than RAID 0 but despite that they say it’s crap I one sentence then talk about how fast it is in another!
They also love to bring in drive failure .. I have owned many very high performance PC’s over 35 years, and have had precisely 1 drive failure – BUT that was on a single external HDD .. So I have always used hardware RAID on all my setups, with 0 failures .. What’s more I was State Manager of a company who imported RAID cards and massive pre-built. RAID 5 drives. We had only 1 instance of a failure, the RAID 5 array lost 3 drives simultaneously – which obviously was a huge problem .. the fact it was Australia’s largest advertising agency at the time made it more fun, and I put in 36 hour day to get it sorted .. BUT that problem came to be because of a well known issue with the Seagate Barracudas at that time.
So this doubles your chance of failure is an incorrect understanding of real world statistics. It doubles nothing. In either case the chance of 1 drive going boom is the same and if that happens in any case then both RAID 0 and normal setup is gone.
Even if it doubled your chance of failure your talking .02% vs .01% or some stupid comparison like that. RAID failures are usually software RAID or non matching drives or other plain bad setup