OCZ’s Vertex 4 solid-state drive

OCZ’s original Vertex wasn’t the first SSD on the market, but it deserves a lot of credit for popularizing solid-state drives among enthusiasts. The first iteration debuted in late 2008 with an Indilinx Barefoot controller that offered much better performance than the flaky JMicron chip making the rounds at the time. Although the 120GB drive’s $470 price tag seems high by today’s standards, it was a bargain compared to Intel’s X25-M, which cost more but offered just 80GB of storage. The Vertex also gained TRIM support via a firmware update, while Intel famously reserved that feature for its second-gen X25-M.

Since its arrival, the Vertex has spawned multiple models distributed across a several generations. About a year after the first Vertex arrived, OCZ ditched Indilinx for a new muse. SandForce’s SF-1200 controller served as the basis for the Vertex 2, and one year after that, the SF-2200 controller brought 6Gbps Serial ATA connectivity to the Vertex 3. Now, it’s time for round four.

Rather than sticking with SandForce, the Vertex has fallen back into the arms of its first love. OCZ actually bought Indilinx last March, so the relationship is much closer now. Less than nine months after the acquisition, the OCZ Octane SSD arrived with a new Indilinx Everest controller onboard. A second-generation version of that controller, dubbed Everest 2, now provides the foundation for the Vertex 4.

The Octane is a solid drive, with competitive performance and a seemingly decent reliability record, so there’s reason to be optimistic about the new Vertex. This latest offering features numerous hardware and firmware changes that promise better performance and improved endurance. Let’s see what it can deliver.

Introducing the second child

Modern SSDs are made up of three primary components: the controller, the NAND flash, and the firmware that ties them together. OCZ controls the first and last elements, and those are the areas in which the Vertex 4 differs the most from the Octane. Let’s start with the controller, which takes center stage on the drive’s circuit board. Seriously, it’s right in the middle of the action:

Most SSDs put the controller closest to the SATA connector, with the NAND chips arranged in an orderly grid below. The Octane follows that layout, but the Vertex arranges the NAND in a sort of ring around the controller. This arrangement was designed to make the trace lengths between the controller and the NAND more consistent; in typical SSD configurations, some NAND packages are much farther from the controller than others.

Another notable feature of the board is the array of “golden fingers” located along its bottom edge. The pattern looks like it might be a match for the mSATA interface, but OCZ tells us this is simply a connector for its internal debugging tools. Alas, you won’t be able to overclock the drive by flipping a few DIP switches. There’s no need to, anyway, because OCZ has turned up the clocks itself.

The Everest controller in the Octane runs its dual ARM cores at “less than 300MHz,” according to OCZ. In Everest 2, those cores are clocked at a speedier 400MHz. The gen-one chip’s 6Gbps Serial ATA interface remains, as do its eight memory channels. The 65-nm fabrication process used to manufacture the chip is unchanged, as well. 256-bit AES encryption support? Check.

Everest 2 does feature a new error correction engine, which OCZ credits for improving the Vertex 4’s NAND endurance. This programmable ECC block can be tuned to match the “specific error characteristics” of different flash flavors. NAND built on finer fabrication processes is more prone to errors and can typically withstand fewer write/erase cycles, making robust error correction an increasingly important feature for SSDs.

The new ECC engine is a big part of nDurance 2.0, a collection of technologies focused on extending drive life. We first heard about nDurance when it was introduced with the Octane as a “highly integrated and complementary solution set of NAND flash management techniques.” OCZ wouldn’t say more at the time, but it contends that version 2.0 is better, of course. This latest iteration purportedly gives the Vertex 4 a much lower write amplification factor than the Octane. OCZ achieved this feat without resorting to compression. Instead, the nDurance 2.0 concatenates multiple write requests and attempts to reduce the number of unnecessary copy-back operations.

The Vertex 4’s actual write amplification factor will depend on the workload, but OCZ says it should be lower than on any other SSD with the same 7% overprovisioning percentage, including those based on SandForce controllers. That would be an impressive feat considering the compression voodoo built into the last couple of generations of SandForce chips.

Another component of nDurance 2.0 is RNA, otherwise known as Redundant NAND Array. This programmable redundancy scheme protects against physical flash failures by striping data and the associated parity bits across multiple flash cells. Like similar redundancy schemes offered by other SSDs, RNA can survive the death of up to a single flash die without losing the user’s data. The distributed parity bits do consume some of the NAND capacity, though. Redundancy is never free.

More on the memory

When coupled with the revamped controller, nDurance 2.0 makes the Vertex 4 quite distinct from the Octane that preceded it. There are more differences, too, like the amount of cache memory on the drive. Our 512GB drive has a whopping gigabyte of DDR3 RAM split between a pair of chips, one on each side of the circuit board. That’s double the cache size of the Octane and four times what you get in the Crucial m4 and Samsung 830 Series.

OCZ wouldn’t divulge why the cache needs to be so large, although we suspect the new nDurance features use some of the added capacity. To start, all Vertex 4 drives rolling off the production line will feature 1GB of cache. Over time, though, the lower-capacity models will migrate to smaller caches. It seems OCZ hasn’t validated any lower-capacity memory chips for use in the Vertex 4 just yet.

Although it features more cache memory than the Octane, the Vertex 4’s complement of NAND is pretty much identical. The drive uses 25-nm synchronous NAND packages bearing Intel’s name. The Vertex 4 may also be sold with Micron flash, but OCZ assures us the chips will meet the same specifications as the Intel ones. Intel and Micron do, after all, share a joint flash venture in IM Flash Technologies.

Like the Octane, the Vertex 4 will be available in 128, 256, and 512GB capacities. OCZ isn’t bothering with 64GB models for its high-end SSDs, which makes perfect sense. The market is already flooded with cheap 64GB SSDs, and these smaller drives tend to perform poorly because they don’t have enough NAND dies to exploit the parallelism available in modern controllers.

Capacity NAND

dies

Dies per package Max sequential (MB/s) 4KB random QD32 (IOps) Price
Read Write Read Write
128GB 16 x 64Gb 1 535 200 90,000 85,000 $179
256GB 32 x 64Gb 2 535 380 90,000 85,000 $349
512GB 64 x 64Gb 4 535 475 95,000 85,000 $699

Our handy comparison chart nicely illustrates the already substantial differences between the sequential write performance ratings for the models in the Vertex 4 lineup; the 128GB model’s write speed is less than half that of the 512GB drive. The sequential read and random I/O ratings are virtually identical regardless of the capacity, though.

The Vertex 4’s firmware purportedly bears little resemblance to what’s found on the Octane, which would explain the huge differences in performance ratings between the two drives. The Octane has lower stats across the board, and it looks particularly weak on the random I/O front. Indeed, OCZ says the Vertex 4’s firmware has been specifically optimized to improve performance with random reads and writes.

As one might expect, OCZ is positioning the Vertex 4 as a high-end solution. Although the Octane technically costs more for some capacities, its price should fall as the Vertex rolls out in volume. The two drives will coexist for at least the next little while. I wouldn’t be surprised to see the Octane replaced by an Agility 4 SSD that pairs the Everest 2 controller and its new nDurance magic with more pedestrian asynchronous NAND, though.

The last ingredient in the Vertex 4 is OCZ’s Toolbox software, which can download and apply firmware updates with just a couple of clicks. The toolbox can also be used to secure-erase the Vertex if you’d like to return the drive to its factory-fresh state.

The toolbox app works well enough, but I wish it had better monitoring options. As it stands, users can view the SMART attributes that track the number of written sectors and the drive’s remaining life, but those values are presented in a simple text window that looks a little low-rent compared to the SSD utilities offered by Intel and Samsung. Since OCZ has the keys to the controller and firmware, it should be able to do some interesting things on the utility front.

Reputation rehab

Let’s be honest; OCZ doesn’t have a stellar reputation for SSD reliability. Part of that comes from its early adoption of the SandForce SF-2200 controller. OCZ was the first drive maker to offer SSDs based on the chip, and firmware updates came fast and furious in the initial few months. Many more months passed before the SandForce controller’s infamous BSOD bug was finally squashed with a firmware fix.

Although the number of users afflicted by the bug appears to have been relatively small, OCZ SSDs have dominated much of the chatter surrounding the issue. That’s not necessarily an indication that the company’s last batch of SandForce SSDs was more prone to problems than competing drives based on the same controller, though. Early adopters flocked to the OCZ drives because they were the only next-gen SandForce models available for quite some time. Those folks are exactly the sort that one might expect to sound off about problems in forum posts and user reviews.

When the Octane was released, OCZ was quick to point out that the controller and firmware were both developed in-house. Any issues would be on OCZ’s head. So, have there been problems with SSDs based on the new Indilinx controller? Yes and no.

We’ve scoured the OCZ forums for evidence of issues with new Indilinx-based SSDs. We’ve also browsed the user reviews at Amazon and Newegg. Overall, there are very few complaints about the Octane, and seemingly no common thread between them. However, OCZ’s budget Petrol SSD, which uses a 3Gbps version of the Everest controller and less exotic NAND, appears to be a bit finicky. There are numerous reports of DOA drives, premature failure, detection issues, and even some instances of freezing and data corruption. Yikes.

To be fair, that sample size is still relatively small. Fewer than 50 user reviews of Petrol and Octane SSDs are spread across Amazon and Newegg. There are over a thousand user reviews of just the Vertex 3 between those two sites.

When we asked, OCZ confirmed that more users are having problems with the Petrol than with the Octane. The affected drives are being replaced, the company said, and it was adamant there are no issues with the Indilinx controller or the associated firmware. It sounds like the Petrol’s cheaper NAND could be the culprit, but OCZ is still investigating.

While I’m hesitant to draw any conclusions based on such limited evidence, it’s clear that OCZ’s new Indilinx-based SSDs aren’t entirely trouble-free. Fortunately, the problem child seems to be the budget Petrol model. The Octane appears to be solid, which bodes well for the Vertex 4.

If you do run into problems with the drive, some comfort can be taken in the fact that it’s covered by a five-year warranty. Most consumer-grade SSDs get only three years of coverage, so OCZ is going above and beyond here. Keep in mind that the longer warranty doesn’t guarantee greater reliability or the integrity of your data, though—just that you’re entitled to a free replacement if the drive goes bad during the warranty term.

Our testing methods

We have a full suite of performance results for literally dozens of different SSDs, but today, we’ve narrowed the field to include only the highest-capacity models available in the Benchmarking Sweatshop. Most of these drives offer 240-300GB of capacity, so they have much less storage than the Vertex 4 512GB. However, the higher-capacity flavors of those other drives don’t carry substantially higher performance ratings than the models we’ve tested. In most cases, the performance specifications are identical, so we should have a good look at how each family of drives performs in its optimal configuration.

We’ve included a Western Digital Caviar Black mechanical desktop drive for reference, which gives us more than enough fodder for overstuffed graphs. Our test methods and systems haven’t changed, so the Vertex 4’s scores can be compared to those in any of our storage reviews dating back to last September.

If you’re familiar with our test methods and hardware, the rest of this page is filled with nerdy details you already know; feel free to skip ahead to the benchmark results. For the rest of you, we’ve summarized the essential characteristics of all the drives we’ve tested in the table below. Our collection of SSDs includes representatives based on the most popular SSD configurations on the market right now.

Interface Cache Flash controller NAND
Corsair Force Series 3 240GB 6Gbps NA SandForce SF-2281 25-nm Micron async MLC
Corsair Force Series GT 240GB 6GBps NA SandForce SF-2281 25-nm Intel sync MLC
Crucial m4 256GB 6Gbps 256MB Marvell 88SS9174 25-nm Micron sync MLC
Intel 320 Series 300GB 3Gbps 64MB Intel PC29AS21BA0 25-nm Intel MLC
Intel 510 Series 250GB 6Gbps 128MB Marvell 88SS9174 34-nm Intel MLC
Intel 520 Series 240GB 6Gbps NA SandForce SF-2281 25-nm Intel sync MLC
OCZ Octane 512GB 6Gbps 512MB Indilinx Everest 25-nm Intel sync MLC
OCZ Vertex 4 512GB 6Gbps 1GB Indilinx Everest 2 25-nm Intel sync MLC
Samsung 830 Series 256GB 6Gbps 256MB Samsung S4LJ204X01 2x-nm Samsung Toggle DDR
WD Caviar Black 1TB 6Gbps 64MB NA NA

We used the following system configuration for testing:

Processor Intel Core i7-2500K 3.3GHz
Motherboard Asus P8P67 Deluxe
Bios revision 1850
Platform hub Intel P67 Express
Platform drivers INF update 9.2.0.1030

RST 10.6.0.1022

Memory size 8GB (2 DIMMs)
Memory type Corsair Vengeance DDR3 SDRAM at 1333MHz
Memory timings 9-9-9-24-1T
Audio Realtek ALC892 with 2.62 drivers
Graphics Asus EAH6670/DIS/1GD5 1GB with Catalyst 11.7 drivers
Hard drives Corsair Force 3 Series 240GB with 1.3.2 firmware

Corsair Force Series GT 240GB with 1.3.2 firmware

Corsair m4 256GB with 0009 firmware

Intel 320 Series 300GB with 4PC10362 firmware

Intel 510 Series 250GB with PWG2 firmware

WD Caviar Black 1TB with 05.01D05 firmware

OCZ Octane 512GB with 1313 firmware

Samsung 830 Series 256GB with CXM03B1Q firmware

Intel 520 Series 240GB with 400i firmware

OCZ Vertex 4 512GB with 05.10.30 firmware

Power supply Corsair Professional Series Gold AX650W
OS Windows 7 Ultimate x64

Thanks to Asus for providing the systems’ motherboards and graphics cards, Intel for the CPUs, Corsair for the memory and PSUs, Thermaltake for the CPU coolers, and Western Digital for the Caviar Black 1TB system drives.

We used the following versions of our test applications:

Some further notes on our test methods:

  • To ensure consistent and repeatable results, the SSDs were secure-erased before almost every component of our test suite. Some of our tests then put the SSDs into a used state before the workload begins, which better exposes each drive’s long-term performance characteristics. In other tests, like DriveBench and FileBench, we induce a used state before testing. In all cases, the SSDs were in the same state before each test, ensuring an even playing field. The performance of mechanical hard drives is much more consistent between factory fresh and used states, so we skipped wiping the HDDs before each test—mechanical drives take forever to secure erase.

  • We run all our tests at least three times and report the median of the results. We’ve found IOMeter performance can fall off with SSDs after the first couple of runs, so we use five runs for solid-state drives and throw out the first two.

  • Steps have been taken to ensure that Sandy Bridge’s power-saving features don’t taint any of our results. All of the CPU’s low-power states have been disabled, effectively pegging the 2500K at 3.3GHz. Transitioning in and out of different power states can affect the performance of storage benchmarks, especially when dealing with short burst transfers.

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at a 75Hz screen refresh rate. Most of the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.

HD Tune — Transfer rates

HD Tune lets us present transfer rates in a couple of different ways. Using the benchmark’s “full test” setting gives us a good look at performance across the entire drive rather than extrapolating based on a handful of sample points. The data created by the full test also gives us fodder for line graphs.

To make the graphs easier to interpret, we’ve greyed out the mechanical drive. The SSD results have been colored by drive maker, with the Vertex 4 set apart from OCZ’s other offering in a yellow shade halfway between mustard and gold.

The Vertex 4 gets off to a sluggish start in HD Tune. With an average read speed of less than 200MB/s, the drive is slower than all the other SSDs. Even Intel’s 320 Series, which has an old-school 3Gbps Serial ATA interface, has a higher sustained read speed in this test. The Octane is ahead by almost 150MB/s.

When we encountered unexpectedly low HD Tune read speeds with OCZ’s RevoDrive 3 X2 PCIe SSD, increasing the block size from 64KB to 8MB returned substantially better performance. That trick doesn’t work with the Vertex 4, whose average speed remains in the 200MB/s range even with larger block sizes.

Whatever’s slowing the Vertex’s read performance in HD Tune doesn’t affect the write speed test. The Vertex 4 tops the field, edging out Samsung’s 830 Series while offering a much higher average write speed than the current crop of SandForce-based SSDs. This time around, it’s the Octane that’s nearly 150MB/s slower than the Vertex.

Shortly after we received our Vertex 4 sample, OCZ warned us that benchmarks run on unpartitioned drives might not show drive’s performance in the best light. The Vertex 4 has a prefetch mechanism that only works on partitioned drives, but we couldn’t pry more details from the company about what’s going on under the hood. We do know that this prefetch functionality isn’t part of the Octane. For what it’s worth, running HD Tune on a formatted partition failed to produce higher read speeds on the Vertex 4.

Curious to see how another benchmark might react to the Vertex’s prefetch scheme, we ran CrystalDiskMark’s 1GB sequential transfer rate tests on a few of the drives. CrystalDiskMark runs on formatted partitions and uses fully random data by default.

OCZ’s latest can’t match the read speeds of the Samsung and Intel SSDs, but it does manage to push over 450MB/s—more than double the throughput it achieved in HD Tune, and nearly enough to catch the Octane. The Vertex 4 is faster in the write speed test, where it has a comfortable lead over the Samsung 830 Series and a huge advantage over the Intel 520 Series and Octane SSDs. Keep in mind that the Intel drive is based on a SandForce controller that relies on easily compressible data to achieve peak write performance. The Octane has has no such excuse.

HD Tune’s burst speed tests are meant to isolate a drive’s cache memory.

The Vertex 4’s burst speeds are a little faster than those of the Octane. However, both Indilinx-based drives are well off the pace set by the fastest SSDs in these tests.

HD Tune — Random access times

In addition to letting us test transfer rates, HD Tune can measure random access times. We’ve tested with four transfer sizes and presented all the results in a couple of line graphs. We’ve also busted out the 4KB and 1MB transfers sizes into bar graphs that drop the mechanical drive and should be easier to read.

The line graph plainly illustrates the massive difference in access times between solid-state and mechanical storage. The gaps between the SSDs are much smaller in comparison.

In the 4KB random read test, the Vertex 4 sits in the middle of the pack, just ahead of the Octane. The 1MB test proves somewhat troublesome, sending the Vertex to the back of the field. Its 1MB random read access times are more than twice as long as the Octane’s.

The picture looks different with writes, although our lone mechanical drive is still horribly overmatched.

Among the SSDs, the Vertex 4 fares well. The OCZ drive has the shortest access times in the 4KB test, and it’s just shy of the three-way tie for the lead in the 1MB test. In both cases, the Vertex offers a substantial performance improvement over the Octane.

Once again, the Vertex 4 fares better with writes than it does with reads. That dynamic doesn’t change when HD Tune is run on a partitioned drive, suggesting that OCZ’s prefetch mechanism perhaps just doesn’t like HD Tune.

TR FileBench — Real-world copy speeds

Concocted by resident developer Bruno “morphine” Ferreira, FileBench runs through a series of file copy operations using Windows 7’s xcopy command. Using xcopy produces nearly identical copy speeds to dragging and dropping files using the Windows GUI, so our results should be representative of typical real-world performance. We tested using the following five file sets—note the differences in average file sizes and their compressibility. We evaluated the compressibility of each file set by comparing its size before and after being run through 7-Zip’s “ultra” compression scheme.

Number of files Average file size Total size Compressibility
Movie 6 701MB 4.1GB 0.5%
RAW 101 23.6MB 2.32GB 3.2%
MP3 549 6.48MB 3.47GB 0.5%
TR 26,767 64.6KB 1.7GB 53%
Mozilla 22,696 39.4KB 923MB 91%

The names of most of the file sets are self-explanatory. The Mozilla set is made up of all the files necessary to compile the browser, while the TR set includes years worth of the images, HTML files, and spreadsheets behind my reviews. Those two sets contain much greater numbers of smaller files than the other three. They’re also the most amenable to compression.

To get a sense of how aggressively each SSD reclaims flash pages tagged by the TRIM command, we’ve run FileBench with the solid-state drives in two states. We first test them in a fresh state after a secure erase. The SSDs are then subjected to a 30-minute IOMeter workload, generating a tortured used state ahead of another batch of copy tests. We haven’t found a substantial difference in the performance of mechanical drives between these states.

When it’s copying the larger, highly compressed files in the movie, MP3, and RAW sets, the freshly erased Vertex 4 is faster than any other SSD. However, it’s stuck in the middle of the pack in the other two tests.

The Mozilla and TR sets are made up of relatively small files that are highly compressible, which is why the SandForce-based Force Series GT and Intel 520 Series do so well. Interestingly, the Octane boasts higher copy speeds than the Vertex in those tests. Its advantage is only 3-7MB/s, which is much smaller than the 32-40MB/s gaps in the movie, MP3, and RAW tests.

When running FileBench in a used state, the Vertex 4 offers slower copy speeds across the board. Its performance drops by only 1MB/s in the Mozilla and TR tests, which allows the Vertex to retain its position in the middle of the field. However, there’s a 20-33MB/s penalty in the other tests. That’s enough of a hit to allow the Samsung 830 Series to take over the lead when copying movies, MP3s, and RAW images.

The Vertex 4 suffers a bigger performance drop than the Octane when transitioning to a used state, suggesting that a more relaxed approach to reclaiming trimmed flash pages might be a part of nDurance 2.0.

TR DriveBench 1.0 — Disk-intensive multitasking

TR DriveBench allows us to record the individual IO requests associated with a Windows session and then play those results back as fast as possible on different drives. We’ve used this app to create a set of multitasking workloads that combine common desktop tasks with disk-intensive background operations like compiling code, copying files, downloading via BitTorrent, transcoding video, and scanning for viruses. The individual workloads are explained in more detail here.

Below, you’ll find an overall average followed by scores for each of our individual workloads. The overall score is an average of the mean performance score with each multitasking workload.

The Vertex 4’s showing in DriveBench 1.0 is disappointing to say the least. We should, however, note that DriveBench runs on unpartitioned drives. Only the sectors required by the test are formatted, and that may confuse the Vertex 4’s prefetch mojo. Let’s see if the individual test results shed any light on the situation.

At least the Vertex isn’t the slowest SSD across the board. The fact that the Vertex 4 scores lower than its Indilinx-based sibling in each and every test suggests that something is amiss, though.

TR DriveBench 2.0 — More disk-intensive multitasking

As much as we like DriveBench 1.0’s individual workloads, the traces cover only slices of disk activity. Because we fire the recorded I/Os at the disks as fast as possible, solid-state drives also have no downtime during which to engage background garbage collection or other optimization algorithms. DriveBench 2.0 addresses both of those issues with a much larger trace that spans two weeks of typical desktop activity peppered with multitasking loads similar to those in DriveBench 1.0. We’ve also adjusted our testing methods to give solid-state drives enough idle time to tidy up after themselves. More details on DriveBench 2.0 are available on this page of our last major SSD round-up.

Instead of looking at a raw IOps rate, we’re going to switch gears and explore service times—the amount of time it takes drives to complete an I/O request. We’ll start with an overall mean service time before slicing and dicing the results.

Not bad, especially considering that DriveBench 2.0 requires an unpartitioned drive. The Vertex 4’s mean service time across our two-week trace is slightly improved over the Octane and nearly quick enough for a spot on the podium.

Splitting the read and write service times puts things into perspective. While the Vertex 4 offers the shortest write service times, its read service times are much less competitive. In both categories, though, the Vertex improves upon the mean service time of the Octane.

There are millions of I/O requests in this trace, so we can’t easily graph service times to look at the variance. However, our analysis tools do report the standard deviation, which can give us a sense of how much service times vary from the mean.

The Vertex 4 exhibits comparatively little variance in its read and write service times. Only the Samsung 830 Series has less variance in its results with both reads and writes. The SandForce-based Corsair and Intel SSDs have an edge over the Vertex with reads, but not with writes.

If I haven’t already scared you off with too many graphs and statistics, this next pair will do it. We’re going to close out our DriveBench analysis with a look at the distribution of service times. I’ve split the tally between I/O requests that complete in 0-1 milliseconds, 1-100 ms, and those that take longer than 100 ms to complete.

We’re most concerned with the number of extremely long service times, and the Vertex fortunately has a very low percentage above 100 ms. That’s true for both reads and writes. As we saw in our other DriveBench 2.0 metrics, though, the Vertex’s write performance is more competitive than its read performance. OCZ’s latest has a higher percentage of sub-millisecond write service times than the others, but it falls a little bit behind the leaders when we look at the 0-1 ms read service times.

IOMeter

Our IOMeter workloads feature a ramping number of concurrent I/O requests. Most desktop systems will only have a few requests in flight at any given time (87% of DriveBench 2.0 requests have a queue depth of four or less). We’ve extended our scaling up to 32 concurrent requests to reach the depth of the Native Command Queuing pipeline associated with the Serial ATA specification. Ramping up the number of requests also gives us a sense of how the drives might perform in more demanding enterprise environments.

We run our IOMeter tests using the fully randomized data pattern, which presents a particular challenge for SandForce’s write compression scheme. We’d rather measure SSD performance in this worst-case scenario than using easily compressible data.

Our IOMeter tests are run with a partition in place, and that seems to suit the Vertex 4. The drive offers higher transaction rates than its peers in the database test, and it’s only beaten at one of the six load levels in the file server and workstation tests, both times by the Samsung 830 Series.

Those three access patterns are made up of read and write requests, but the web server test is a read-only affair. There, the Vertex 4 trades blows with the Crucial m4 and a couple of SandForce-based SSDs. OCZ’s latest doesn’t come close to matching the performance of the Samsung 830 Series in the web-server test as the load ramps up.

Speaking of not coming close, check out the Octane, which struggles mightily. The Vertex 4 boasts substantially higher transaction rates across all four tests.

Boot duration

Before timing a couple of real-world applications, we first have to load the OS. We can measure how long that takes by checking the Windows 7 boot duration using the operating system’s performance-monitoring tools. This is actually the first time we’re booting Windows 7 off each drive; up until this point, our testing has been hosted by an OS housed on a separate system drive.

Only about a second separates the fastest SSD from the slowest in our Windows 7 boot test. Take away the pack-leading Crucial m4, and the rest of the solid-state drives are squeeze into a half-second span. Vertex 4 is certainly competitive here, but it’s not any faster than its rivals.

Level load times

Modern games lack built-in timing tests to measure level loads, so we busted out a stopwatch with a couple of reasonably recent titles.

The results of our level load tests very close. Although the Vertex 4 isn’t among the leaders, it’s never more than a second off the fastest load time.

Power consumption

We tested power consumption under load with IOMeter’s workstation access pattern chewing through 32 concurrent I/O requests. Idle power consumption was probed one minute after processing Windows 7’s idle tasks on an empty desktop.

The Vertex 4 draws only slightly more power than the Octane at idle, but it nearly doubles the power draw of its Indilinx-based sibling under load. Samsung’s 830 Series actually draws more wattage than the Vertex under our IOMeter load despite offering only half the capacity.

The value perspective

Welcome to our famous value analysis, which adds capacity and pricing to the performance data we’ve explored over the preceding pages. We used Newegg prices to even the playing field, and we didn’t take mail-in rebates into account when performing our calculations. Since the Vertex 4 isn’t widely available online just yet, we’ve had to use OCZ’s suggested retail price for that drive.

First, we’ll look at the all-important cost per gigabyte, which we’ve obtained using the amount of storage capacity accessible to users in Windows.

The Vertex 4 looks pretty competitive here. It’s cheaper per gig than the Octane by a good 30 cents and costs less per gig than SandForce SSDs with similar synchronous NAND.

Lest you think that only the highest-capacity model has an attractive cost per gigabyte, the 128GB and 256GB versions of the Vertex 4 both cost less than $1.40/GB. Given the rabid competition between SSD vendors, we wouldn’t be surprised to see the Vertex 4 selling for even less than OCZ’s suggested price when it hits virtual shelves in volume.

Our remaining value calculations use a single performance score that we’ve derived by comparing how each drive stacks up against a common baseline provided by the Momentus 5400.4, a 2.5″ notebook drive with a painfully slow 5,400-RPM spindle speed. This index uses a subset of our performance data described on this page of our last SSD round-up. The Intel 510 Series was actually worse than our baseline in one of the tests—100+ ms writes in DriveBench 2.0—so we’ve fudged the numbers a little to prevent that result from messing up the overall picture. We’ve also had to tweak the score for the RevoDrive in that test; it didn’t log any extra-long write service times, and the resulting zero is problematic for our baseline comparison formula.

Not bad, all things considered. Our overall score includes the results from HD Tune’s read speed test, in which the Vertex 4 tanked, plus scores from our DriveBench tests, which run on unpartitioned drives that may not play nicely with OCZ’s prefetch trickery. Nevertheless, the Vertex just edges out the Octane and ends up about 100 percentage points behind the leader.

Now for the real magic. We can plot this overall score on one axis and each drive’s cost per gigabyte on the other to create a scatter plot of performance per dollar per gigabyte.

Even if it scored higher on our performance index, the Vertex 4 would need to drop in price to match the Samsung 830 Series’ enviable position on our scatter plot. Still, the Vertex 4 offers slightly better performance than the Octane while being much cheaper per gigabyte.

To look at the value perspective from a slightly different angle, we’ve divided our overall performance score by the sum of our test system’s components. Those parts total around $800 before we add the cost of the SSDs.

With a $700 street price, the Vertex 4 512GB costs quite a bit more than the 240-300GB drives that dominate the scatter plot. That pushes the Vertex off the right, but it’s at least in a more attractive position than the Octane, which costs $150 more and is no faster overall.

Conclusions

OCZ has been one of the most aggressive SSD makers when it comes to rolling out drives based on new controller technology, and the Vertex 4 fits nicely with that tendency. Not even six months after the Octane gave us our first taste of Indilinx’s Everest controller, there’s already a new version of the chip in the Vertex 4. When combined with other hardware changes, a firmware overhaul, and the NAND-friendly features that fall under the nDurance 2.0 umbrella, this new Vertex is very different from the Octane SSD that came before it.

As we’ve seen over all-too-many pages of rainbow-colored graphs, these two drives have very different performance characteristics. The Vertex 4 is definitely the faster of the two, and in some cases, it beats every other 2.5″ SSD we’ve tested. Particularly impressive are the drive’s copy speeds with compressed data like movie, music, and image files. This latest Vertex also delivers exceptionally high IOMeter transaction rates with workloads that mix reads and writes, teasing the potential of an enterprise-oriented offering based on similar technology.

However, the Vertex 4 also struggled in some tests, possibly due to a prefetch mechanism we unfortunately know little about. This prefetch scheme is purportedly optimized for partitioned drives, which is a fair choice to make for a consumer-oriented SSD, but one that may also handicap performance in some benchmarks, including our own real-world DriveBench simulations. The Vertex 4 looks like a very fast SSD, but it may not perform up to expectations in all applications.

The fact that OCZ sent us a 512GB version of the drive also presents problems. Most of our readers are shopping for smaller SSDs, and those capacities carry much lower sequential write ratings according to OCZ’s specifications. In tests where the Vertex 4 did well against SSDs in the 240-300GB range, I worry that the 256GB model might be less competitive, especially given how inexpensive drives like the Samsung 830 series have become. We’ll have to see about getting some other models in the Vertex 4 lineup to test, and it may be time for another update to our storage test suite.

Those reservations aside, the few days I’ve been able to spend testing the Vertex 4 have left me impressed. The drive may not be the undisputed performance king, but it adds new features to take care of the NAND and backs up claims of improved reliability with a five-year warranty. The fact that the Octane has gone several months without any major issues also gives me some faith in the Indilinx controller and the team charged with handling its firmware.

Comments closed
    • rjseo1
    • 7 years ago
    • rjseo
    • 8 years ago
    • merryjohn
    • 8 years ago
    • Coyote_ar
    • 8 years ago

    [b<]"Alas, you won't be able to overclock the drive by flipping a few DIP switches."[/b<] That was an old school comment 😛

    • Ruiner
    • 8 years ago

    U2 reference liked.

    • Compton
    • 8 years ago

    TBH, the one time I had to RMA one of the seven OCZ drives I own, the RMA process was pretty damn awesome. The ticket process leaves something to be desired, but my turnaround time was about 6 business days. YMMV, but they sent my drive back 2 day air.

    Also, I once had to RMA a Mushkin, and their support was fantastic as I could actually talk to a human being and get the RMA done quickly.

      • indeego
      • 8 years ago

      TBH 6 business days is a lot for a critical resource like a storage drive.

        • BobbinThreadbare
        • 8 years ago

        Shouldn’t a business using drives for mission critical stuff have it’s own backups on hand anyways?

        Just to be clear, I mean backup drives ready to go, not backups of data.

        • Bensam123
        • 8 years ago

        Seagate provides active swaps for like $10 and a deposit. You send the drive and they send you one in the mail immediately. As long as the drive you RMA’d checks out, you get your deposit back.

      • Deanjo
      • 8 years ago

      Best RMA service I ever had was from Leadtek. Had a bum GForce 4400Ti with a bad fan, RMA’d it, sent me a air shipping label for free, they accidently replaced it with a 4200 Ti by mistake, called them again, sent me a 4600 Ti for the inconvienence and told me I can keep the 4200Ti as well all over the course of 1 week.

        • ImSpartacus
        • 8 years ago

        What bros!

    • Rza79
    • 8 years ago

    I don’t understand why so many websites test these drives unpartitioned. The performance of a SSD is influenced by different filesystems, so I would say that these drives need to be tested in at least two different filesystems (and not at all unpartitioned). I don’t know any use or anybody that uses a SSD unpartitioned.

    • Bensam123
    • 8 years ago

    Holy crap, first thought when seeing the interface on the other end of the board was “It can be used as a PCI-E device too?!?!” only to realize I was mistaken. This could’ve been a ridiculously cool drive if you could take off the case and slip it into a PCI-E slot if you wanted. Given they probably wouldn’t want that as it would compete directly with their Revodrives.

    Curious that the 1GB of cache doesn’t do more. It makes me wonder if the caching scheme actually measures usage over time, so you wouldn’t be able to replicate or fully exploit it without the drive being used for weeks on end. I’m sure if that was the case OCZ would tell you though so they would have a better looking review.

    Guess Samsung is still the goto choice… With the reads the way that they are the read performance of these drives might improve immensely in the future with firmware updates. It may be worth a revisit in like six months.

    Nice bit on reputation. It’s nice when people take a level headed approach to debunking common conceptions and basically lay everything on the table in a non-biased fashion. Definitely added to the article.

    Your price graphs most definitely could still use a best fit line. They help with reading and interpreting results in graphs with a meaningful variable on one of the axis (money in this case as well as performance). Most statistic programs have an option to turn it on; I know SPSS does.

    [url<]http://en.wikipedia.org/wiki/Linear_regression[/url<]

      • Voldenuit
      • 8 years ago

      [quote<]Curious that the 1GB of cache doesn't do more.[/quote<] I thought that the cache was probably why the Vertex 4 performance was so good at writes.

        • Bensam123
        • 8 years ago

        Also more then likely true… Not sure why it can’t be used for both though.

      • Airmantharp
      • 8 years ago

      +1 for the best-fit line suggestion.

      Maybe a light color-gradient in the background?

        • Bensam123
        • 8 years ago

        Yeah that would be a good idea too. There are a lot of things you can do to graphs to improve readability and how well they convey information.

    • south side sammy
    • 8 years ago

    wish you guys would have included a Seagate Hybrid drive in the review. Be nice to know where it stands. And any dope on when other manufacturers are going to come out with theirs ?

      • Chrispy_
      • 8 years ago

      I’ve used a couple of hybrid momentus drives, I’d rate them as good laptop drives where you’re limited to one physical drive, and need the capacity.

      In terms of testing though, they’re hard to get accurate numbers from because they vary a lot between different test runs, and on top of that, they’re still in the same ballpark as a caviar black, which is so far off the graphs in some cases that it’s just plain excluded.

    • internetsandman
    • 8 years ago

    After seeing the Revodrive and now this, it’s gonna take me a while to start thinking OCZ is a great brand again. I’m not saying they’re absolutely horrible, but being presented with one underwhelming product after another becomes incredibly boring and disappointing, especially when these products are supposed to represent the absolute top-of-the-line OCZ products. It kinda reminds me of when Fatal1ty was slapping his name on every PSU and CPU heatsink possible, and it turned out nearly all of them were crappy products. OCZ has a long way to go in my eyes to earn back a decent reputation

    • tootercomputer
    • 8 years ago

    “OCZ doesn’t have a stellar reputation for SSD reliability”

    That is a kind comment. More accurate would be to substitute “history” for “reputation”, as the latter implies perception whereas the former connotes actual fact. OCZ has a history of poor SSD reliability. Just say it. I got the Vertex 3 after the supposed firmware fix that was going to solve BSODs, and it BSODed all over the place anyway. I tried and tried, and still got BSOds. I got a replacement drive and have had but one BSOD in 5 months, and that was on the first day.

    So I hope their drives are good. But personally, I avoid OCZ at this time. It will take a while before I would consider buying their product. That said, I remember years ago when they were slammed for the reliability of their RAM, but eventually that changed, I purchased some of their RAM and it worked perfectly. I hope the same for their SSDs.

    • Compton
    • 8 years ago

    OCZ and Marvell do have a partnership. Openly, they’ve announced the Kilimanjaro project together (as reported here on TR). Surely they have more of a relationship than just one controller, and I’m virtually 100% that one OCZ product has used a Marvell on the sly already. OCZ provided the FW, and Marvell provided the silicone for Kilimanjaro, but as I said, I’m certain that another such Marvell/OCZ collaboration has already happened and shipped.

    [url<]https://techreport.com/discussions.x/22288[/url<] Products using the new Marvell controller (the 88SS9187 announced a few weeks back) are "imminent". The Marvell products already on the market, like the Plextor M3/M3P are already quite fast and stack quite well against the V4. If OCZ had waited to "fix" the V4 FW's power consumption and low QD performance, then they'd lose any advantage getting a faster drive out to market first would bring. The V4 is quite fast and very interesting for a variety of reasons, but I expect it will clearly be outclassed by new contenders over the next 3/4 months or so. If they had waited any longer to get the FW validated and "fixed" it would launch at a time when it wouldn't look as good compared to newer releases. And, it couldn't have launched on 4-4 2012. I do expect the V4 will have some new competition (soonish) from drives with similar performance, but not as lopsided toward high QDs. OCZ should get that fixed in the next few months, but I can sympathize with getting the drive out now and fixing it later.

    • indeego
    • 8 years ago

    [i<]Nadir 1[/i<] sounds like a better name for this. Burn, OCZ, burn.

      • Voldenuit
      • 8 years ago

      Have you seen the Vertex 4 benchmarks on Anand? It is positively scorching there, so there’s definitely something going on with the performance discrepancy between the two reviews.

      It’s possible that the drive is very sensitive to usage model (as seen in TR’s unpartitioned bench and Anand’s low-queue benches), and OCZ has said that the firmware needs work* (cf idle power and low queue depth performance).

      I think it’s too early to write off the Vertex and Everest 2, but I can certainly understand being underwhelmed by it when Samsung has the excellent 830 out.

      * Are there words more chilling to an enthusiast than OCZ saying these things?

        • indeego
        • 8 years ago

        Good on writes and bad on reads, it matches TR’s benches fairly closely. On light loads it’s horrid, certainly evidence of OCZ’s lackluster firmware testing facilities for real-world usage patterns.

        Maybe on a workstation/server? I wouldn’t trust OCZ on a workstation/server, where a write-fast drive belongs.

        Fine, adjust the firmware, but the drive is seriously borked right now. I wouldn’t begin to consider it with many strong stable competitors like Intel, Samsung, Crucial…

    • Voldenuit
    • 8 years ago

    Per [url=http://www.anandtech.com/show/5719/ocz-vertex-4-review-256gb-512gb<]Anand's review[/url<] there are a couple of things worth mentioning: * Write amplification is not as good as m4 or intel 320 * 128 GB and 256 GB drives will only use half of the available 1 GB cache

    • Deanjo
    • 8 years ago

    [quote<]I worry that the 256GB model might be less competitive, especially given how inexpensive drives like the Samsung 830 series have become.[/quote<] Sure but have you tried to source one outside of the US? Newegg.ca still doesn't have them and I can guarantee you that you won't be able to visit a non-us site without tripping over Vertex 4's.

      • Voldenuit
      • 8 years ago

      Anand [url=http://www.anandtech.com/show/5719/ocz-vertex-4-review-256gb-512gb/3<]tested the 256 GB Vertex 4[/url<] and it is neck to neck with the 512 GB drive in nearly all tests. I agree with you that TR should have tried to source a drive before delving deep into conjecture.

      • indeego
      • 8 years ago

      > I can guarantee you that you won’t be able to visit a non-us site without tripping over Vertex 4’s.

      So we will constantly be tripping over Vertex 4’s on non U.S. sites?

        • Deanjo
        • 8 years ago

        Heck would be surprised if you could’t pick one up at the nearest gas station.

      • kuraegomon
      • 8 years ago

      Second this – neither Newegg.ca or NCIX have stocked _any_ Samsung 830’s. So, in Canada at least, the 830 is quite difficult to come by.

      • Firestarter
      • 8 years ago

      FWIW, there seem to be plenty 830’s available in Europe.

        • smilingcrow
        • 8 years ago

        You can claim the 20% VAT back if you buy one in the UK which brings it down to ~£200 which makes it seem the obvious choice for a 256GB drive.

    • Thatguy
    • 8 years ago

    I keep getting excited to read these reviews and coming away with a, Meh. Are we seeing the limits of the Sata interface at this point? The advances are not leaps and bounds anymore. Just incremental.

    Nice review, though.

      • stdRaichu
      • 8 years ago

      We’ve been bumping up against the limits of the SATA interface since the days of the Vertex 2; older SATA 3Gbps SSD’s could never get past 270MB/s, and pretty much as soon as 6Gbps SATA came out we saw SSD’s that could saturate that (for reads) too. On top of that, the SATA protocol still contains a lot of “cruft” in the ATA protocol that was designed to minimise the limitations of spinning platter media.

      Last I read a consortium called SATA-IO was working on a protocol that would allow seamless integration of discs into the PCIe protocol (SATA Express?), making PCIe-based SSD’s a reality without having to faff about with boot ROMs and special drivers. Then things like a laptop with an integral 64/128GB blisteringly fast PCIe SATA drive and a 2.5″ slot for mass storage could become commonplace.

      Edit: Bit Tech article about SATA Express – [url<]http://www.bit-tech.net/news/hardware/2012/03/08/ssds-faster-sata-express/1[/url<]

        • Thatguy
        • 8 years ago

        Interesting link. I expected the PCIe drives to be much faster than TR’s review found them to be. Oh well there is always the next revision to salivate over!

          • stdRaichu
          • 8 years ago

          PCIe drives are crazy fast if you look in the right places and are prepared to reduce your bank account to a smoking crater. We have a few of Fusion-IO cards at work (http://en.wikipedia.org/wiki/IO_Accelerator) and they’re mind-numbingly fast, but their entire architecture was developed around PCIe from the start, rather than the approach most “prosumer” PCIe SSD manufacturers make which is to take a load of SATA SSD controllers and RAID0 them together.

          Hopefully once PCIe SSD’s become more common we’ll see more controller designs designed specifically for it.

            • Thatguy
            • 8 years ago

            My bank account is already in crater status 🙁

    • Compton
    • 8 years ago

    The problem is this:

    The Samsung 830 is a bad ass package. So is the new Marvell 9174 + 24nm Toggle NAND paradigm.

    The Vertex 4 jumps in with what I’m convinced is unoptimized firmware, and puts in a good show, but I don’t know that I could recommend it without having had laid hands on one. Given the tradeoffs it could be a tough sell for some, but it looks like a much improved package over Octane.

      • DPete27
      • 8 years ago

      Agreed on the unoptimized firmware. If some of those unusually low scores dont change with firmware updates, I would be surprised and it doesn’t look like much of a step up from their Vertex 3. Price is higher, performance is lower, reliability is assumedly somewhat similar with the Sandforce BSOD patch….

      This article really reiterates how impressive the Samsung 830 is. Its got the performance, reliability, value…Incredible.

        • Compton
        • 8 years ago

        I think it’s unoptimized because they had to hurry to validate it and get it out ahead of the rest of the Marvell drives. They could have sorted this stuff out ahead of time, but it would be faced with several competitors as fast or faster.

        And by God, I’m sick of the OCZ plastic shell chassis. I was hoping they’d move to something that doesn’t look and feel as cheap, but it does do what it’s supposed to do. But it wouldn’t have killed them to make the V4 look a little more like the 830, at least on the outside.

          • kuraegomon
          • 8 years ago

          The rest of the Marvell drives? What Marvell relationship are you implying? This is OCZ’s own controller – no one else will be using it anytime soon.

            • Compton
            • 8 years ago

            Marvell and OCZ do have a “relationship”. Openly, they’re working on Kilimanjaro together (though that’s surely not the only product they’ve been working on together). Another product comes to mind.

            But presumably, new products using the Marvell 88SS9187 are coming. Marvell announced the controller release would be followed by shipping products from partners within weeks of the announcement. So OCZ probably decided to just use the FW they had rather than tweak it for low QD performance.

            • kuraegomon
            • 8 years ago

            Fair enough. I wasn’t aware of that. That being said, I found your original comment a bit misleading: i.e. ‘the “rest” of the Marvell drives’ implied that the Vertex 4 was one such, where it clearly is not – for better _and_ for worse.

            • Compton
            • 8 years ago

            Well, I was referring to the more recent Marvell releases, especially the Plextor M3 and M3P which would fare quite well in relation to the V4 I think. Now that those drives are out, there has to be more around the corner, ones which could be exceedingly fast.

    • drfish
    • 8 years ago

    Meh. The warranty is indeed awesome but otherwise I’d take the 830. This coming from someone who has bought dozens of OCZ drives in the past couple years…

      • 5150
      • 8 years ago

      I could’ve written this exact statement. Lots of OCZ drives, have given up after the Vertex 3. The Samsung 830 256GB I have is wonderful and has a great utility that comes with it.

Pin It on Pinterest

Share This