RevoDrive 120GB PCI Express SSD Review
Windows on our PCI-E SSD?
OCZ Technology is on a roll. While most consumer SSD manufacturers are content to just slap the latest controller and some NAND into a 2.5-inch enclosure and call it a day, OCZ has been pumping out innovative products, from top-of-the-heap SATA SSDs to the blistering-fast (and stylish) USB 3.0 Enyo drive. Now it has introduced the RevoDrive, a PCI-E SSD in capacities from 50GB to 480GB. Though it’s not the first PCI Express SSD (Fusion-io’s been making enterprise-level PCI-E SLC devices for years), it is the first bootable consumer PCI-E SSD. OCZ claims the RevoDrive can hit up to 540MB/s reads and 450MB/s writes, which sounds like nonsense. But is it?

It's the first bootable consumer PCI-E SSD, but lack of Trim hurts performance over time.
The RevoDrive is a x4 PCI Express card containing a Pericom PI7C9X130 PCI-E-to-PCI-X bridge, a SiliconImage SiI3124 PCI/PCI-X-to-SATA controller, two SandForce SF-1200 controllers, and 120GB of NAND flash—it’s effectively two 60GB Vertex 2 drives in RAID 0 on a single PCB. Installation is easy, though as of press time, the drivers lack an executable file and need to be installed via Device Manager, unless you’re installing Windows on the drive, in which case they can easily be F6’d at install. The SiliconImage BIOS is accessible during POST, so you can wipe and restore the RAID manually should you so choose. The default stripe size is 64KB as all our tests were run at that size.
Because the Trim command doesn’t pass through RAID controllers, you’ll have to rely on the SandForce controllers’ built-in garbage collection utilities. In our tests, repeated abuse did slow the RevoDrive in some tests. After several days of heavy (and unrealistic) use, average sustained reads in HDTune dropped from 300MB/s to 240MB/s, while average sustained writes dropped from 267MB/s to just 175MB/s—worse than a single Vertex 2 drive. However, as OCZ points out, HDTune is a queue-depth 1, low-level hardware benchmark for unformatted drives that doesn’t deal well with RAID. Our Premiere Pro encoding times slowed from 337 seconds to 358 seconds. PCMark Vantage HDD subscores remained above 44,000, and our IOMeter 4KB random write test, at queue-depth 32, hit above 80,900 IOPS—that’s 316MB/s, or 65 percent faster than the 48,900 IOPS we saw from a single Vertex 2.
Where are the advertised 540MB/s reads and 450MB/s writes? We didn’t see them in the low-level benchmarks we typically use. Instead, we had to look to ATTO Disk Benchmark, which tests drive performance over a variety of read and write sizes from 500 bytes to 8,192KB. Lo and behold: For larger file sizes (512KB and above), ATTO recorded read speeds above 540MB/s and writes above 460MB/s. Our 100GB Vertex 2, by comparison, got around 285MB/s read and 274MB/s write on the same test.
Is the RevoDrive a practical solution for home users? High queue-depth IOPS are more useful for servers than for day-to-day use, and the absence of Trim is palpable, though OCZ claims to be to working on adding Trim support. Depending on the benchmark, the RevoDrive’s performance ranges from nearly twice as fast as a single Vertex 2 to slightly worse. But its performance at queue depths greater than 1 never falters, and in those scenarios it crushes all comers.
The RevoDrive comes in capacities from 50–480GB. The 120GB version we tested currently retails for $370; a 120GB Vertex 2 is $310. Given the lack of Trim and the fact that most home use doesn’t require high queue-depth performance, most people should go for a single SATA SandForce drive. A price drop and Trim support, though, could turn this from a decent and intriguing product to a must-have.
RevoDrive 120GB PCI Express SSD

Science
Great queue-depth performance and random read/write; bootable; trumps single SandForce.
Magic
No Trim; drive slows down after heavy use.
8
| OCZ 120GB RevoDrive (PCI-e x4) | OCZ 100GB Vertex 2 (3Gb/s) | |
|---|---|---|
| Capacity | 120GB | 100GB |
| HDTune 4.01 | ||
| Avg Read (MB/s) | 243.0 | 196.3 |
| Random-Access Read (ms) | 0.0 | 0.1 |
| Burst Read (MB/s) | 175.2 | 228.0 |
| Avg Write (MB/s) | 175.3 | 221.9 |
| Random-Access Write (ms) | 0.2 | 0.1 |
| Burst Write (MB/s) | 207.4 | 207.5 |
| 4KB Read (IOPS) | 10,600 | 11,045 |
| 4KB Write (IOPS) | 10,275 | 10,066 |
| IOmeter Random-Write IOPS (4KB, Queue Depth 32) | 80,958 | 48,958 |
| Premiere Pro (sec) | 358 | 359 |
| PCMark Vantage HDD | 44,450 | 39,309 |
Comments
![]()
EricX2
November 12, 2010 at 7:52am
I can't believe you never mention that this would technically mean you could have a PC with no internal wiring besides the motherboard power.
Referring to something like a business PC or any small form factor style PC that traditionally do not have large drives anyway. The PC's we get at work at SFF Dell OptiPlex's with 160gb drives without DVD/CD drives. Putting this in the PCI-E Slot would also eliminate the loudest part of the system.
![]()
allantang
November 12, 2010 at 1:33am
I have this drive in 120gb and am very happy with it. Yeah the lack of trim is a bit of bummer but, there are manaul ways of "cleaning up" the drive so that it runs better again.
![]()
Pureoverclocking
November 11, 2010 at 5:38pm
- Sequential Access - Read:
- up to 540MB/s
- Sequential Access - Write:
- up to 490MB/s
- after reading 5 review all over internet not one them show this drive hiting those speed waste of money.
![]()
Keith E. Whisman
November 11, 2010 at 7:05pm
Actually the article shows that those benchmarks are only achieved on one benchmark suite.
![]()
Bullwinkle J Moose
November 11, 2010 at 12:27pm
Here's the problem
Synthetic Benchmarks won't show you how fast this drive is for uncompressible data
Only a copy and paste of several hundred megabytes to and from the same drive under XP will show you what this drive will do under actual load
First off, due to Windows 7's caching scheme, ALL drives (Slow or Fast) seem to finish a copy and paste in the same amount of time and cannot be used for this test
In a worst case scenario, using an ATOM computer with Windows XP and Zero SSD Tweaks, a OCZ VERTEX 2 will copy and paste data at only 3.6 Megabytes per second
A 5400RPM laptop drive was faster than the Vertex 2 in this test because OCZ drives require massive amounts of Tweaking and highly compressible data to get the numbers they are advertizing
A 7200RPM desktop drive was A LOT faster than the Vertex 2 in this type of test
Anyone working with uncompressible data "already on the drive" such as video editors should avoid these SSD's and stick with the much faster desktop platter drives
Using a slower ATOM computer for these tests will amplify the difference between slower and faster drives and give you a better idea of the "Relative" speed difference between drives
You should use this test for ALL SSD's and compare the results to common hard drives so that end users can get a feel for the "Actual" throughput of these drives on uncompressible data
Remember, Data on the Vertex drive's is already compressed and cannot be compressed again during a copy/paste to show you the actual throughput of the drive under XP
Worst case scenario testing under XP is the way to go with SSD's to see what they will really do under actual workloads without endless tweaking and without getting bogus results due to Windows 7's caching scheme
![]()
COMMANDER_COOK
January 27, 2011 at 2:05pm
The reason the SSD is so slow in XP is because XP doesn't recognize the physical alignment that SSDs use. The SSD is actually being commanded to rewrite each block several times instead of just once.
And no, SSDs do not require a massive amount of tweaking and 7 does not have a "caching" scheme.
![]()
Keith E. Whisman
November 11, 2010 at 3:07pm
I don't know, for my OS and games an single SSD seems to be really fast, a lot faster than a 7200RPM HDD. The OS boots faster than any platter hdd can and games load in a few seconds compared to the long wait you get with a spinning platter. Itunes pops open just like that with nearly 20,000 songs.
Are you referring to .raw when you mention uncompressible files?
How does any limitation like the ones you noted affect the majority of users?
![]()
Bullwinkle J Moose
November 11, 2010 at 4:44pm
You are correct!
A vertex 2 boots really fast 7-12 seconds on my ATOM
A 300X compact Flash boots in 12 seconds but "Seems" much slower than the Vertex in accessing and loadin programs on the same machine
I was not referring to how fast it seems to be in loading Windows or games
The access times are great but handling data that is already compressed is horrible
ALL data on a Vertex 2 drive is already compressed and will not be compressed again during a copy and paste function to and from the same drive!
You are talking about apples and I'm talking about oranges
Too simplify this, Sandforce based SSD's have excellent load times for data already on the drive that never changes and make great boot drives, but have horrible throughput during a copy and paste function to and from the same drive
So, as long as you only use it to boot your computer and load games, it's all good
But if you want to manipulate data already compressed or that is uncompressible, it's a completely different story
Anyone doing video editing might have better results with a platter based drive!
A copy and paste of 200MB with 919 files in 85 folder got the following results>
55 seconds (3.636MB per second) on a Vertex 2
54 seconds on a 640GB Western Digital Laptop drive @ 5400RPM
17 seconds on a 640GB Western Digital Desktop Drive @ 7200RPM
You do the math!
![]()
ptolemyx
November 12, 2010 at 11:23am
Look, I don't know what you've done to your Vertex 2 to get such awful performance, but I tried your copy/paste test with a 500MB folder of 2500 files (a software project tree; ~160M is compressed distribution packages) and got around 80MB/sec.
To reiterate, this was copy/pasting across a single drive. It _was_ on Windows 7, which you claim has caching tricks that make judgement difficult -- which doesn't make sense considering W7 also shows copy bandwidth in the progress dialog; I suppose that's fake?
Anyway, I ran the first test 8 times in a row, and the copy speed never slowed down. That's 4G of data; I have less than 2G of free RAM. I'm not saying this is a scientific measure, but it's also over 20x the performance you suggest I should be getting.
![]()
Bullwinkle J Moose
November 12, 2010 at 2:12pm
I too got better results on a vertex 2 in Windows 7 but I also got identical results in Windows 7 with platter based drives due to that OS's funky caching scheme
Based on these results, you might as well use the cheapest platter based drive with Windows 7
To get real numbers based on the actual speed of the drive, you need to use XP
If all drives show faster but identical results in Windows 7, then the results are WRONG!!!!!
You also OBVIOUSLY used a faster computer than an ATOM to get your results as the speed of this test is also determined by the CPU and chipset
I simply used an ATOM to get the slowest times in a worst case scenario
That way, I could see the difference between one drive finishing in 54 seconds and another finishing in 55 seconds
A faster CPU would reduce the time to finish these tests, but I might not have noticed the 1 second difference between those two drives because of it
![]()
Keith E. Whisman
November 11, 2010 at 5:22pm
OK... I think I got it.. That's a pretty big limitation. Would it be different if the data was divided up on two or more drives in a Raid 0 array of SSD's? Or is it the same thing just spread across three drives? And would this limitation be evidenced on the new Hybrid drives now hitting the market? 500Gig HDD with a 4Gig SSD build in.
![]()
Bullwinkle J Moose
November 11, 2010 at 6:32pm
I don't use RAID and never tried the hybrid drive
I would supect that the hybrid drive would perform as well as any other platter based 7200RPM laptop drive though as the flash is only used for read caching and would not affect the writes
two drives however WOULD affect the copy paste times
two seperate drives that ARE NOT connected in RAID should give you MUCH better copy and paste times but that setup would not show you the internal throughput of a single drive as I have attempted to illustrate
Copying from one SSD and pasting to another should give excellent results but I have yet to try it
Sorry, I have no idea how RAID would perform with these drives under XP
Log in to MaximumPC directly or log in using Facebook
Forgot your username or password?
Click here for help.


















