Intel Core i7 930
Noctua NH-D14 Cooler
OCZ Flex 12GB DDR3
Sparkle GTS 250
BFG Tech 550w
NZXT Premium Cables
Corsair Obsidian 800D Chassis
ASUS optical drive
Windows 7 Home Premium 64-bit
Since this is a new test system and a new storage platform, unfortunately we don't have anything comparable to test against just yet. That being the case, I will just report the benchmarks by themselves.
I will be using the following benchmarks with the RevoDrive:
CrystalDiskMark 3.0 x64
ATTO Disk Benchmark 2.46
AS SSD 1.5.3784.37609
SiSoft Sandra Lite 2010 SP3
Real world file transfer
And here are the results, in no particular order.
For the real world file transfer test, I measured how long it took to copy two sets of files from one location to another on the drive. For the small files test, I transferred 557 MP3 files, totaling 2.56GB. For the large file, I zipped these same MP3s into a single file measuring 2.52GB in size. Results are below:
Small File Transfer Time: 39 seconds
Large File Transfer Time: 32 seconds
Things start off very strong in CrystalDiskMark on the read side but write numbers are a little low. ATTO shows the true speed of the RevoDrive, posting the highest numbers we have tested to date from a solid state drive. The subpar write numbers continue in AS SSD. Everything looks great in Sandra and although not at the 70k IOPS rating, IOMeter shows good results as well.
So, why are the write numbers a bit lower than what you may be seeing at other sites? There are a few possible reasons that I can think for this.
First, as we have discussed in other reviews, many benchmarking programs that are commonly used for solid state drives were actually designed for use with traditional spinning hard drives. This could explain why some of the results are through the roof and some don't seem as fast as they should be.
Second, I tested this drive as an operating system drive, and not as an empty storage drive. It's my believe that nobody in their right mind is going to buy a solid state drive with the intent purpose of using it as a storage drive, at least not while SSDs cost so much more than hard drives.
Finally, the more data you write to a solid state drive, the more it degrades its performance. This normally isn't a huge issue with most SSDs that support TRIM, but with a RAID0 setup, there is no TRIM. The SandForce controllers are pretty good about keeping things running smoothly under normal Windows operations, but when you hammer a drive with multiple benchmarks, numbers are going to decrease. For example, IOMeter fills the drive to capacity each time it runs. Do this once for Read tests and again for Write tests and you already have a pretty "dirty" drive.
Furthermore, I am the second reviewer to work with this drive. OCZ initially sent the RevoDrive to Zahn but after working with it for a few days, it was determined that the drive wasn't fully compatible with his system and thus, he sent it to me.
The SandForce controllers should work this out over time, but for the purpose of benchmarks, it makes it difficult to show the true speed of the drive. After filling the drive a few times, I reran CrystalDiskMark and the write scores were much lower. Point proven.
Let' move ahead and wrap things up with a conclusion.