OCZ Vertex 60GB Review + RAID0

Solid state disks have been a topic of great debate among hardware enthusiasts for well over two years now. The thought of having a non-volatile storage medium which can read any piece of data with near-zero access latency would have seemed like nothing more than a pipe-dream not too long ago, but now it is a reality.

OCZ Vertex
OCZ Vertex SSD

Unfortunately JMicron controllers paired with Multi-Level Cell (MLC) disks burst everybody’s bubble: they were slow writing, caused freezing and lock-ups for many users and though technically fast, ended up leading to a very poor user experience. A new company is out to solve all that though: Indilinx. They have endeavored to design and manufacture a drive controller made specifically for solid state disks that would solve all our woes. But did they succeed?


Vertex Boxes
A couple of Vertex drives ready to be pulled out of their boxes

Now before I begin showing off some benchmark data we’ve collected, I think that a little more background is required into specifically the OCZ Vertex series of disks. If you’ve been poking around the net a little already, you’ve probably found that the 60GB Vertex is rated to be slower than its 120GB big brother that got sent out as samples to reviewers. I want to set out to answer whether the more “affordable” 60 gig model is worth its asking price or if OCZ pulled another fast one on us.

Vertex in Antistatic Bag
This is what you are presented with once the box is opened

Based on their specs, the 60 and 120 gigabyte models perform as follows:

60GB Max Performance:
Read: Up to 230 MB/s
Write: Up to 135MB/s
Sustained Write: Up to 70MB/s

120GB Max Performance:
Read: Up to 250 MB/s
Write: Up to 180MB/s
Sustained Write: Up to 100MB/s

Second, there is a lot of talk of firmware versions for these drives. A lot of other reviews are on old firmware revisions that you won’t be using and OCZ has done a great job of improving the Vertex over time. You can flash it yourself once you pick one up, so don’t worry about being left out in the cold if you get some older stock. The firmware that I’m using in this review is v1.10. Firmware and flashing instructions can be found on the product page for the Vertex. You’ll notice that on the back of these drives, there is the standard SATA power and SATA data, but also room for a jumper. That’s used for the flashing process.

Two Stacked Vertexes
Our two test drives stacked one on top of the other and just about ready to go

As for our testing platform, everything was done on a system with an Intel ICH9r drive controller, a Q6600 processor and 8GB of RAM. All testing was done on Windows Vista and no special attention has been paid to setting up these drives optimally. I merely dropped them into the machine and benched them. You won’t find cherry-picked results here; what you see is what you get with no fancy tricks to speed things up. Note that some users say that Windows XP can’t set proper offset when partitioning these drives while Vista can, so that may affect you if you use these under XP. I don’t know what the heck that means either, but we enthusiasts are all on Vista by now anyways, right?

Single Drive Benchmarks

Alright, so with that entire preface into the probably overly complex world of buying solid-state drives, let us get into the number games. First up is how the Vertex looks as most of you will be using it: in a single-drive configuration. Whether in a laptop or desktop, this is what you can expect to see.

First up in our barrage of tests is Atto which has become standard for testing MLC SSDs because it shows performance of small up to relatively large transfer sizes.

Single Drive OCZ Vertex Atto

In Atto, we see that the Vertex is writing up to 152MB/s and reading up to 233MB/s which is perfectly in line with what OCZ rates these drives for on read and well above the stated 130MB/s max write performance. This is the kind of stuff that I love to see – ratings being not only met but even exceeded. You’ll also notice that at lower transfer sizes, write speeds are higher than read speeds; presumably because of the 64MB onboard write-back cache. So far everything seems to be holding up very well.

Next, we have the mandatory HDTach benchmark. We performed the full test on both read and write.

Single OCZ Vertex HD Tach Full

In HDTach average read is much lower than we saw in Atto at an average of 173MB/s, while write averaged out at 106MB/s. These figures are much lower than we saw previously. Furthermore, the graph of the write test is a little strange as it drops off sharply at the end. Regardless of any of this, I’m just glad to see that write didn’t constantly drop off as JMicron MLC disks do. This bodes very well for these new Indilinx drives. Seek time is of course the lowest possible at 0.1ms as expected.

HD Tune is next up in our suite of tests performed. It simply measures sequential read and sequential write similar to HDTach.

Single OCZ Vertex HD Tune Read

Good stuff here, nice flat curve averaging out to 207.5MB/s on read. Again, I’m not particularly concerned about this since none of these utilities were designed for testing SSDs. Besides, we have already seen the Vertex exceed manufacturer ratings in Atto. Again, the lowest possible 0.1ms access time was achieved. Now on to write!

Single OCZ Vertex HD Tune Write

Looks like Indilinx is pretty good at this SSD stuff! Nice flat graph unlike what you’d see with a JMicron based disk which are literally all over the map. It also put out a strong performance in terms of average access time at 0.2ms.

And finally we have Everest benchmarks for you. Everest is much the same as the other benchmark suites in that it measures the average read and write speeds over the entire disk, but the most interesting benchmark is its average write access test which really separates the boys from the men.

Single OCZ Vertex Everest Write Access

At the beginning of the graph, we see that latency starts out notably lower and gradually climbs. I can only presume that this is the 64MB buffer being filled. It does finally stabilize at just over 0.3ms though, which I’m ecstatic about. It is clear that this drive can handle having a full write-back cache. I’m actually pretty much ready to declare this drive as “safe” for use in your computer.

Single OCZ Vertex Everest Random Read

Very good as expected.

Single OCZ Vertex Everest Write Access

Again, not much to say, it’s a pass and boy these things are much faster than mechanical drives!

RAID 0 Benchmarks

Alright boys and girls, are you all ready to see what you can do with two ridiculously fast drives in RAID? This should prove to be interesting regardless of the results. What we hope to see is 120GB of really fast storage. Let’s jump right in!

Two OCZ Vertex in RAID
Two test drives in RAID 0 and ready to go!

First up in our barrage of tests is Atto which has become standard for testing MLC SSDs because it shows performance of small up to relatively large transfer sizes.

RAID OCZ Vertex Atto

This is actually a little disappointing; I would have loved to see a doubling in performance across the board. While these MLC drives are now slightly used after our initial slew of tests, I’d still expect to see results better than this. And in fact, small writes are slower.

Next, we have the mandatory HDTach benchmark. We performed the full test on both read and write.

RAID OCZ Vertex HD Tach

Again we are only seeing small increases in performance rather than the doubling that would be expected in a RAID 0 configuration. This test actually makes it look like the ICH9r is limiting us. That shouldn’t be happening since I’ve personally seen this board and chipset run up to about 350MB/s of sequential read in RAID 5 on normal hard drive.

HD Tune is next up in our suite of tests performed. It simply measures sequential read and sequential write similar to HDTach.

RAID OCZ Vertex HD Tune Read

Yeah, we ran this twice… For some reason, the RAID 0 array on the ICH9r didn’t like this test at all. The RAID array performed much worse than the single-drive configuration. Not looking good and this is really very disappointing. You can actually see a couple of peak points where you see the drives are pushing through briefly but it just doesn’t seem to “stick”.

RAID Vertex HD Tune Write

Running in RAID 0 smoothed our write latency quite nicely. Logically I’d imagine this is because there is more write-back cache with two drives and the Intel controller can actually system memory as write-back cache as well. Oddly though, the average write performance, while smoother, is within 1MB/s of the single-drive configuration.

And finally we have Everest benchmarks for you. Everest is much the same as the other benchmark suites in that it measures the average read and write speeds over the entire disk, but the most interesting benchmark is its average write access test which really separates the boys from the men.

RAID OCZ Vertex Everest Write Access

Kind of looks like the rib cage of a dinosaur don’t you think? Seriously though, aside from artistic merit this graph sucks. Write access times are not smooth here and we are seeing over 1.2ms of latency at peak – very un-SSDish.

RAID OCZ Vertex Everest Read

What is there to say? The single-drive configuration was faster. While interesting, these results are really very disappointing to say the least.

RAID Vertex Everest Write Access

Absolutely hideous… While a single drive was able to hold a constant speed, the RAID configuration is dropping as low as 25MB/s at times. I was so excited to see an MLC disk that was able to get relatively flat write curves, but now I’m once again sad.

Overall, I really can’t recommend these drives in a RAID array. These drives were set up in a system which normally runs a five hard drive RAID 5 array with absolutely no issues. I literally removed that array, dropped these drives in, created the array and booted Windows off the OS drive again. I’m not saying that it’s impossible to configure a RAID array which performs well with these drives, but if you can it’ll probably take a lot of work. Really, what we saw was just a bunch of very inconsistent results that tell me that something in here isn’t ready for prime-time, and I don’t think it’s the ICH9r. While it’s not an amazing controller, I know it can work better than this.

Long Term Testing and Subjective Analysis

Now that I’ve been using the Vertex for nearly a month in my daily-use laptop in a single-drive setup, I can tell you that the difference is like night and day compared to the original 80GB hard drive I had in here. The CPU in this machine is just a 2.0GHz Core 2 Duo T7250 with 2GB of RAM backing it, but things are significantly smoother than they used to be. Multitasking is a very different experience. When I first put Windows Vista on this machine with the Vertex installed, it installed in just over 15 minutes off of a DVD. The Vertex really is that fast. Better yet, my laptop actually seems to respond to most tasks much more quickly than my desktop computer with a quad core processor and 8 GB of RAM. All things considered, it’s really quite incredible. If you ever doubted that SSDs made much of a difference then doubt no more, it really is true!

Now I will admit that since this is my primary work computer, I did carry around my original 80GB drive for a couple of weeks just in case, but never needed it. The Vertex has worked flawlessly and performance has remained high. I’ve actually noticed that it doesn’t feel as fast as when new because MLC performance does degrade after lots of read/write cycles which can only be fixed by writing over every sector with something like the full HDTach write test which of course destroys all data. If you are concerned about this problem, you aren’t alone and OCZ actually has a response.

Single OCZ Vertex Used
Atto results on a drive with 3 weeks of use on it!

One reason I jumped onboard with OCZ for these SSDs is because they offer great support and allow you to flash the firmware. Since I benched these drives a few days ago, OCZ has already released their beta trim utility which can be found right here. It is really not recommended that you use that if you are on a 64 bit OS since it has caused corruption for some people. It is 100% safe on 32 bit operating systems and it will restore drive performance to near-new levels almost instantly with no data loss. You must be using firmware 1.10 for it to work. As for the results? Check it out:

Single OCZ Vertex Atto Trimmed
How beautiful is this to see? OCZ really has done it with Indilinx!

Overall, the Vertex series is massively impressive. After long term use, I’ve reached the point where going back to systems with mechanical hard drives is somewhat of a painful experience: you have to go back to listening to a disk drive scratch away while you wait for it to find the data you need. If you can afford one of these 60GB or larger disks then I guarantee you won’t be disappointed, especially with OCZ’s great product support. If you can afford two of them, get a 120GB drive and skip RAID: it’ll spare a lot of problems and headaches. If all you want is a system upgrade that really will make a huge difference in performance the Vertex series is simply it!

OCZ Vertex Editor's Choice

One thought on “OCZ Vertex 60GB Review + RAID0”

Leave a Reply

Your email address will not be published. Required fields are marked *