| OCZ Vertex 3 Max IOPS Edition SSD | |
| Reviews - Featured Reviews: Storage | |
| Written by Olin Coles | |
| Tuesday, 24 May 2011 | |
|
Page 1 of 13
OCZ Vertex 3 Max IOPS SSD Review
Manufacturer: OCZ Technology Group, Inc. Full Disclosure: The product sample used in this article has been provided by OCZ. Not so long ago, Benchmark Reviews tested the OCZ Vertex 3 SSD, which delivered the best transfer speeds and operational performance we had seen from any SATA-connected drive. Not to be outdone, the OCZ Vertex 3 Max IOPS Edition solid state drive debuts with the same SandForce SF-2281 controller but switches to Toshiba TH58TAG7D2FBAS9 3nnm toggle NAND flash components. OCZ suggests up to 120,000 combined IOPS are possible, with speeds reading 550 MB/s. In this article, Benchmark Reviews compares the OCZ Vertex 3 Max IOPS Edition SSD to other high-speed SATA 6Gb/s storage devices. OCZ Technology first launched their Vertex SSD series shortly after the 2009 Consumer Electronics Show, and delivered impressive performance to performance hardware enthusiasts using the latest Indilinx Barefoot storage controller. Indilinx would go on to have a phenomenal year, and capture the consumer market with their Barefoot controller. But market dominance is a short-lived success, and at the 2010 CES there was a new player in the market: SandForce Inc. This new company promoted the SandForce SF-1200 SSD Processor, which operated faster and more efficiently than the competition, and would be the foundation for OCZ's Vertex 2 SSD. Like clockwork, OCZ debuted their next-generation enthusiast storage platform at the recent 2011 CES, and introduced the Vertex 3 SSD to the market a short while later. For many within the industry, SandForce was seen to control the 2010 market in much the same way that Indilinx did in 2009. The difference now is that SandForce's platform offers several technical benefits that the Indilinx platform was not capable of. Already into the Q2-2011, the landscape is approximately the same, but with some interesting new twists. OCZ Technology has recently acquired Indilinx, and while there's no word on any new project developments they've moved forward with the Vertex 3 SSD based on the second-generation SandForce's SF-2281 SATA 6Gb/s controller. OCZ promises 60,000 IOPS at up to 550 MB/s transfer speeds with the Vertex 3 SSD, and Benchmark Reviews confirms that they're keeping this promise with tests of the new storage device on a B3-stepping Sandy Bridge platform. The second-generation SF-2281 SSD processor maintains all of the original core technology SandForce originally introduced in the SF-1200 series, but now improves SSD performance with 20% faster IOPS and 40% faster sequential read/write throughput. They've enhanced BCH ECC capability, and the new processor now supports ATA-7 Security Erase. Finally, the new SF-2200 series implements cost-effective 20nm-class NAND flash from all leading flash vendors with Asynch/ONFi1/ONFi2/Toggle interfaces.
Even after decades of design improvements, the hard disk drive (HDD) is still the slowest component in any personal computer system. Consider that modern desktop processors have a 1 ns response time (nanosecond = one billionth of one second), while system memory responds between 30-90 ns. Traditional hard drive technology utilizes magnetic spinning media, and even the fastest spinning mechanical storage products still exhibit a 9,000,000 ns / 9 ms initial response time (millisecond = one thousandth of one second). In more relevant terms, the processor receives the command and must then wait for system memory to fetch related data from the storage drive. This is why any computer system is only as fast as the slowest component in the data chain; usually the hard drive. In a perfect world all of the components operate at the same speed. Until that day comes, the real-world goal for achieving optimal performance is for system memory to operate as quickly as the central processor and then for the storage drive to operate as fast as memory. With present-day technology this is an impossible task, so enthusiasts try to close the speed gaps between components as much as possible. Although system memory is up to 90x (9000%) slower than most processors, consider then that the hard drive is an added 1000x (100,000%) slower than that same memory. Essentially, these three components are as different in speed as walking is to driving and flying. Solid State Drive technology bridges the largest gap in these response times. The difference a SSD makes to operational response times and program speeds is dramatic, and takes the storage drive from a slow 'walking' speed to a much faster 'driving' speed. Solid State Drive technology improves initial response times by more than 450x (45,000%) for applications and Operating System software, when compared to their mechanical HDD counterparts. The biggest mistake PC hardware enthusiasts make with regard to SSD technology is grading them based on bandwidth speed. File transfer speeds are important, but only so long as the operational IOPS performance can sustain that bandwidth under load. Bandwidth Speed vs Operational PerformanceAs we've explained in our SSD Benchmark Tests: SATA IDE vs AHCI Mode guide, Solid State Drive performance revolves around two dynamics: bandwidth speed (MB/s) and operational performance (IOPS). These two metrics work together, but one is more important than the other. Consider this analogy: bandwidth determines how much cargo a ship can transport in one voyage, and operational IOPS performance is how fast the ship moves. By understanding this and applying it to SSD storage, there is a clear importance set on each variable depending on the task at hand. For casual users, especially those with laptop or desktop computers that have been upgraded to use an SSD, the naturally quick response time is enough to automatically improve the user experience. Bandwidth speed is important, but only to the extent that operational performance meets the minimum needs of the system. If an SSD has a very high bandwidth speed but a low operational performance, it will take longer to load applications and boot the computer into Windows than if the SSD offered a higher IOPS performance.
|
|





Comments
The only "con" to your review, one of few I've ever had, is the slight hit you issued in your conclusion regarding "expensive enthusiast" product. OK, it is contrary to the masses and their hard earned dollars, no question. The 240 Gb version at $2.58/Gb is a goodly sum. But looking at it from a performance standpoint, and thinking of the terminology "enthusiast" makes me think of it not as a "con" but pretty mainstream, if not reasonable as far as bang-for-buck goes.
There's my two cents worth, raising the price to $2.60/Gb. But no one will pay, right?
80,103 IOPS Random Write (4k)
36,120 IOPS Random Read (4k)
556MB Sequential Write
515MB Sequential Read
Really solid build aswell, which perhaps isn't suprising since these guys apparently are better known in the military/rugged world. Shaping up to be one of best purchases I ever made!
The 240GB drive is much faster than the 120Gb drive. It gets even worse with 60/64 Gb(like vertex 2 series) drives which can be 70% slower than the 240Gb drive with incompressible write speed.
This review gives people the impression that a 120GB drive will perform like a 240GB drive but they do not.
I know its part of the marketing to send only the big fastest drives to reviewers but the performance difference between capacities should be more clearly explained
comment just sent - unclear as to why the above 2 fields have bounced back yet contents of message was sent?
Olin, my response was not directed at your valid comment.
I simply sent a lengthy comment to Benchmark Reviews about the relevance
of the performance of the slowest component in a system - the hard drive.
It was not published.
I run 2 older Sata2 SSD's (A-RAM 32Gb) in Raid 0 to kick start the system and run a few apps. When I was using a single SSD and Ubuntu 9.04 the read/write speed was 270/253 Mb. In terms of the 'IOPS' I had no idea and didn't appreciate the importance as an equal indicator of overall system performance.
All of my other mid range bits that make up the system are more than adequate for what I do....6 core AMD,a GTX460, a Raptor storage drive, a SATA3 & USB3 motherboard, blue-ray burner,2 LED 22" screens etc.
My point and I'm sure many of you have read that the money spent on the hard drive replacement 'SSD' should be the priority.
When I get the bucks OCZ will be my next upgrade.
Question : when in Raid 0, is the read/write & IOPS doubled also?
And, as Olin wisely adds, a LOT depends on the controller.
The closest we've come to near-linear scaling was the combination of a Highpoint RocketRAID 2720 and 4 x Hitachi 15,000 rpm 2.5" SAS HDDs each with 64MB of cache:
3 x HDDs in RAID 0 topped out at 441 MB/second
4 x HDDs in RAID 0 topped out at 618 MB/second
We would not have ever tried 3 x HDDs, except that one was
defective from the factory, so we configured a RAID 0 with 3
of these Hitachis, while we waited for the replacement to arrive.
But, this near linear scaling has been the exception, NOT the rule, in our experience with RAID 0 arrays (which we use often, primarily for speed).
Sincerely yours,
/s/ Paul A. Mitchell, Instructor,
Inventor and Systems Development Consultant
All Rights Reserved without Prejudice
Thanks again for another good review.
A question keeps coming up, as I read these various reviews:
if TRIM does not work with multiple SSDs configured in a RAID array,
do you happen to know if TRIM will work if each SSD is configured
in AHCI mode, and then joined to an OS "software RAID"?
Do you have enough instrumentation at your lab to answer this question?
I'm aware that an OS "software RAID" cannot host the OS,
but I have an XP software RAID running well with 2 x 6G WD HDDs,
hosting a dedicated data partition, and it appears to work just fine.
RSVP if/when you have the time.
Thanks again! Keep up the good work.
Sincerely yours,
/s/ Paul A. Mitchell, Instructor,
Inventor and Systems Development Consultant
All Rights Reserved without Prejudice
So long as the drives are in AHCI/IDE mode with an OS that supports TRIM (Microsoft Windows 7), then the commands will receive the SSD. Hardware RAID does not use AHCI/IDE mode, and will not pass along TRIM commands. So if you're using Windows 7 (not XP) and configure mirrored sets in the Disk Management console using AHCI/IDE SSDs, then it's not truly RAID and you'll have the benefit of TRIM support... with the penalty of a CPU/RAM hit for software mirroring.
Thanks again, Olin!
By "mirrored sets" do you mean RAID 1, for example, and not RAID 0?
Also, just to clarify, do you predict the same for RAID 0 / striped arrays using OS software RAID?
This could be the topic for another Review, perhaps.
p.s. With so many quad-core CPUs proliferating, now with hyper-threading too, the CPU-RAM hit does not seem to be a huge penalty to pay, particularly for desktop and workstation users.
Also, users who have shrunk their C: system partitions have probably also created dedicated data partitions e.g. D:+ Those dedicated data partitions would be good candidates for a software RAID, particularly if TRIM will work reliably, as described above.
Sincerely yours,
/s/ Paul A. Mitchell, Instructor,
Inventor and Systems Development Consultant
All Rights Reserved without Prejudice
XP/Pro supports software RAID 0 also:
we have it running on a P45 chipset
with 2 x 6G Western Digital 1TB HDDs.
I'm not currently a Windows 7 user,
but I would expect that the same is true
for Windows 7.
The partitions need to be changed to "Dynamic" first;
but after that, it's a piece o' cake to configure software RAID 0
as long as the drive letter is NOT C: .
One needs to use the OS to effect this configuration,
so structural changes to C: are not possible at this point.
Do you have some spare hardware laying around,
with which to try it? Maybe 2 x 6G SSDs (my preference)?
OCZ might be particularly interested in this experiment,
with your usual battery of benchmark tests.
You be the "Benchmark Man" -- to be sure!!
Sincerely yours,
/s/ Paul A. Mitchell, Instructor,
Inventor and Systems Development Consultant
All Rights Reserved without Prejudice
##softwaresecretweap ons.com/jspwiki/windows-7-software-raid-disk-partition-and-volume-context-menus
"New Striped Volume" is an option.
Sincerely yours,
/s/ Paul A. Mitchell, Instructor,
Inventor and Systems Development Consultant
All Rights Reserved without Prejudice
somehow "softwaresecretweapon s.com"
became "softwaresecretweap ons.com" above:
just eliminate that extra space after "weap"
Sincerely yours,
/s/ Paul A. Mitchell, Instructor,
Inventor and Systems Development Consultant
All Rights Reserved without Prejudice
The experiment I'm suggesting would need Windows 7
to configure a software RAID 0 with 2 x 6G SSDs
initialized in AHCI mode.
The question at hand: will Windows 7 pass TRIM commands
to 2 x SSDs configured in an OS software RAID 0?
Sincerely yours,
/s/ Paul A. Mitchell, Instructor,
Inventor and Systems Development Consultant
All Rights Reserved without Prejudice
[begin quote]
The Professional/Enterprise/Ultimate editions of Windows 7 officially support the following dynamic disk modes:
* Simple
* Spanned
* Striped (RAID-0)
* Mirrored (RAID-1)
These are the officially supported modes. It has been possible in previous Windows releases to enable unsupported modes through DLL modifications and/or registry changes, and this will probably be the case in Windows 7 as well. At your own risk, of course.
RAID-5 dynamic disks are only available in Windows Server editions.
Note that dynamic disks are not available in the Standard/Home editions of Windows 7.
[end quote]
Sincerely yours,
/s/ Paul A. Mitchell, Instructor,
Inventor and Systems Development Consultant
All Rights Reserved without Prejudice