| OCZ Agility-EX SLC SSD OCZSSD2-1AGTEX60G |
| Reviews - Featured Reviews: Storage | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Written by Olin Coles | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Wednesday, 30 September 2009 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
OCZ Agility-EX SSD ReviewOCZ has earned the pole position in the race for Solid State Drive consumer supremacy. Most of the OCZ SSD product line is sold at retail to enthusiast end-users, but the EX-series offers SLC construction and delivers premier performance to the Enterprise sector and an affordable price. Corporate environments offer vast potential for SSD technology by offering drive performance faster than SAS-based storage, and dramatically reducing heat output. In this article, Benchmark Reviews tests the 60GB OCZ Agility EX SSD OCZSSD2-1AGTEX60G, and Indilinx-based SLC Solid State Drive. The Indilinx Barefoot controller has become the backbone chip for over a dozen different SSD models, first introduced in the OCZ Vertex SSD and later in the Vertex-EX SLC SSD, the IDX110M00-FC reinvents itself as the OCZ Agility-EX SLC SSD series. Identical to the popular OCZ Vertex series in every aspect except NAND selection, the Agility EX series offers the same Indilinx 'Barefoot' controller and 64MB of DRAM buffer, but now boasts ultra-fast 270/200 MBps read and write speeds. Benchmark Reviews tests the reaction time and bandwidth performance for the 120GB OCZ Vertex Turbo SSD OCZSSD2-1VTXT120G model against over two dozen other storage products in this article. Very recently Seagate introduced the Barracuda XT hard drive, and while you might be wondering what that has to do with an SSD article I have a point: Seagate is the first to offer a product for the SATA 6Gb/s (SATA-III) interface. This is good news for hard drives, but it's even better news for SSDs, which have encroached upon the 300 MBps SATA-II throughput barrier for quite some time now. The first two SATA-6G motherboards have already been announced: the ASUS P7P55D Premium and GIGABYTE GA-P55-EXTREME, which will open-up desktop performance and eventually lead to improved SOHO server architecture. The competition among Solid State Drive manufacturers is reminiscent of the video card wars between AMD and NVIDIA a few years ago. The invention of new controller technology has slowed somewhat, while each company races to combine high-speed NAND with custom firmware tweaks to produce the fastest SSD speeds possible. The Indilinx Barefoot controller has exploited a dominant hold over the consumer SSD market during the past several months, and 'overclocked' Solid State Drive products are beginning to replace 'new' product designs. In this article, Benchmark Reviews compares the fastest SLC SSD we've discovered to-date: the 60GB OCZ Agility-EX OCZSSD2-1AGTEX60G.
Anyone familiar with articles published at Benchmark Reviews is very well aware of our obsession with Solid State Drive technology. They're complex, and every SSD is different than the next. Nothing like Hard Disk Drive technology, which improves as spindle speed and cache buffer are increased, SSDs are the future and because of this their internal architecture is constantly evolving. This is why we offer so much coverage on the topic: it's interesting and exciting. Plus, they can turn the average computer system into a roaring beast. It's not marketing hype; for once the truth is stranger than fiction. Since first making a public debut at the 2007 Consumer Electronics Show, Solid State Drives (SSDs) have been a topic of hot discussion among performance enthusiasts. These nonvolatile flash memory-based drives feature virtually no access time delay and promise a more reliable storage medium with greater performance while operating at a fraction of the power level. Moving into 2008, SSDs became a consumer reality for many performance-minded power users. Now that 2009 has proven how strong the industry support behind Solid State Drive technology is, we should hope that mainstream acceptance moves faster than it did for DDR3 SDRAM. Solid State Drive products are no longer restricted to bleeding edge hardware enthusiasts or wealthy elitists. Heading into 2009, SSD storage devices were available online for nearly $2 per gigabyte of storage capacity while the most popular performance desktop hard drive hovered just above $1/GB. While most consumers are waiting for that day when SSD costs the same as HDD, they seem to be forgetting how Solid State Drives have already surpassed Hard Disk performance in every other regard. Our collection of SSD reviews is a good starting point for comparing the competition. SSD Testing DisclaimerEarly on in our SSD coverage, Benchmark Reviews published an article which detailed Solid State Drive Benchmark Performance Testing. The research and discussion that went into producing that article changed the way we now test SSD products. Our previous perceptions of this technology were lost on one particular difference: the wear leveling algorithm that makes data a moving target. Without conclusive linear bandwidth testing or some other method of total-capacity testing, our previous performance results were rough estimates at best.
Our test results were obtained after each SSD had been prepared using DISKPART or Sanitary Erase tools. As a word of caution, applications such as these offer immediate but temporary restoration of original 'pristine' performance levels. In our tests, we discovered that the maximum performance results (charted) would decay as subsequent tests were performed. SSDs attached to TRIM enabled Operating Systems will benefit from continuously refreshed performance, whereas older O/S's will require a garbage collection (GC) tool to avoid 'dirty NAND' performance degradation. It's critically important to understand that no software for the Microsoft Windows platform can accurately measure SSD performance in a comparable fashion. Synthetic benchmark tools such as HD Tach and PCMark are helpful indicators, but should not be considered the ultimate determining factor. That factor should be measured in actual user experience of real-world applications. Benchmark Reviews includes both bandwidth benchmarks and application speed tests to present a conclusive measurement of product performance. About OCZ Technology IncOCZ Technology Group, a member of JEDEC, designs, develops and manufactures ground-breaking, high performance memory and computer components that set industry standards. OCZ products are the first choice for users needing high-reliability, ultra-high performance solutions. In 2007, PC Power & Cooling and Hypersonic PC were brought into the OCZ Technology Group, forming a well-rounded, highly innovative organization that places the company at the forefront of high-end computing. All of OCZ Technology Group's products are available through its worldwide network of distributors, online resellers and retail stores. OCZ Agility EX FeaturesThe OCZ Agility EX Series makes SLC (single-level cell) NAND-based storage truly affordable in a solid state drive for the first time. The Agility EX provides the best of both worlds -the performance and advantages of SLC NAND technology at an incredible value. Based on the quality Indilinx controller, the Agility EX Series delivers an enhanced computing experience with faster application loading, snappier data access, shorter boot-ups, and longer battery life. With superior performance over both conventional hard drives and MLC (multi-level cell) SSDs, the SLC-based Agility EX features incredible sequential and random write speeds for an unparalleled computing experience across the complete spectrum of applications. In addition to exceptional performance the Agility EX series provides an improved total cost of ownership (TCO) with its superior write/erase cycle endurance. Agility EX SSDs feature industry-dominating speeds, up to 255MB/s read and 195MB/s write speeds, 64MB of onboard cache, and unique performance optimization to keep the drives at peak performance. The OCZ Agility EX drive feature a durable yet lightweight alloy housing, and because OCZ SSDs have no moving parts, the drives are more rugged than traditional hard drives. Designed for ultimate reliability, Agility EX Series SSDs have an excellent 1.5 million hour mean time before failure (MTBF) ensuring peace of mind over the long term. All Agility EX SSDs come backed a three year warranty and OCZ's legendary service and support.
As of October 2009 the OCZ SSD family includes the following products from top to bottom:
OCZSSD2-1AGTEX60G Specifications
First Look: OCZ Agility EX SSDWhen it comes to the appearance of notebook drives, it must be understood that the product you're looking at will be hidden away from plain view once it gets installed. Keeping in mind that this product is solid state, and therefore has no moving parts, it takes some extra-special attention to product presentation in order to help keep the consumer feeling comfortable with their premium purchase. OCZ does their very best to make each new SSD product as appealing as the last, and even though the Agility-EX is a premium high-performance SSD it's essentially identical to every other Solid State Drive product OCZ makes.
Unlike desktop computers which utilize a SATA cable system to connect drive to motherboard, nearly all notebooks allow the 2.5" drive to simply slide directly into a connection bay within the system. In addition to notebooks and desktop computer usage, this OCZ Agility-EX Indilinx SLC Solid State Drive can be utilized for enterprise servers or high-abuse data systems, but it's real focus is high-performance speed. Most Indilinx-based SSDs include a two-pin jumper for flashing firmware updates onto the controller, which is obsolete with newer firmware (that updates without a jumper shunt). One unfortunate omission from all recent OCZ SSD products is the integrated High-Speed USB 2.0 Mini-B last seen on their Solid-series of SSD products.
Unlike the average Hard Disk Drive (HDD) storage product, SSD's are nearly impervious to impact damage and do not require or benefit from any kind of special vibration dampening or shock-proof enclosures. OCZ utilizes a flat-black metal enclosure for nearly all of their Solid State Drive products, including the Agility EX series, which reveals the internal components after removing just four small counter-sunk screws on the underside.
Standard 2.5" drive bay mounting points are pre-drilled and threaded into the Agility-EX SSD, which allows for quick upgrade or addition into any existing notebook or desktop system. The mounting positions matched up to the drive bracket on my laptop, and after only a few minutes of drive cloning I was quickly loading Windows O/S. The OCZ Agility EX does not include a 2.5" to 5.25" converter tray with the kit, as we've seen other manufacturers begin to include, but OCZ offers this adapter in a separate kit if necessary. Now that you're acquainted with the basic exterior features of the OCZ Agility-EX SLC Solid State Drive, it's time to peek inside the OCZSSD2-1AGTEX60G enclosure and inspect the internal components... Indilinx Internal ComponentsOCZ has designed the Agility EX Solid State Drive to use the exact same architecture and components as their OCZ Vertex SSD series, with the NAND flash modules being the only difference between the two products. The NAND parts on the Agility EX use a Single Layer Cell (SLC) construction, compared to the MLC construction of NAND components on the Vertex. For this Internal Components section, we have re-used content from the original OCZ Vertex series review. Intermittent and delayed response cycles (stuttering) from Solid State Drive products is not entirely wide-spread among all MLC SSDs, but it has become a big-enough issue among most affordable SSD products that many are well-aware of the phenomenon. Consumers first experienced the bitter taste of stuttering SSD performance with the OCZ Core Series (v1) SSD, although it has also been reported with the G.Skill MLC SSD and Patriot Warp v2. The phenomenon occurs when the drives buffer is filled faster than it can read or write data, and was prevalent among first-generation JMicron JMF602 SSD controllers. In our OCZ Apex SSD article, Benchmark Reviews detailed how a single JMicron JMB390 RAID controller managing a pair of JMF602(B) SSD controllers in a RAID-0 striped array was able to overcome the 'stutter' experienced in MLC drives... at least in read-from functions. Indilinx utilizes more traditional methods of delivering performance with their renowned Barefoot ARM7 Solid State Drive micro-controller.
To the untrained eye, the OCZ Agility-EX SSD looks like every other Solid State Drive you've probably seen when the internal components have been exposed. There's a collective bank of NAND modules, usually with Samsung markings, followed by the SATA controller chip. Indilinx delivers the SATA controller interface, since their 'Barefoot' chip was production-ready to be paired with a large cache at a time when JMicron was (and still is) a few months away with their successor to the JMF602B chip.
While the internal NAND flash DRAM used in the original Indilinx SSDs, such as the OCZ Vertex, are comprised of Samsung K9HCG08U1M-PCB00 IC parts (pictured below) branded as K9HCG08U1M PCB0, the 'overclocked' Indilinx SSDs use several different manufactures to supply the fastest SLC and MLC NAND modules available. These lead-free RoHS-compliant 48-pin ICs are multi-layer, with one IC directly atop another to offer 64GB in 4x organization for the 64GB (60GB advertised) OCZSSD2-1AGTEX60G model.
Indilinx claims that their IDX110M00-FC 'Barefoot' chip offer a maximum read speed 230 MBps and supports the capacity up to 512GB with standard multi-level cell (MLC) NAND flash, but really this limit was based off of early generation chips. The Indilinx (IDX110) Barefoot SSD controller chip is touted as delivering bandwidth over 200 MBps and random input-output (IO) of 20,000/s into various servers such as on-line transaction processing (OLTP) and streaming server units. These figures are NAND-flash dependant, which is why the OCZ Vertex Turbo series and Corsair X-series SSDs offer improved speed when compared to the original Indilinx product varieties.
A single Elpida 64MB SDRAM module is marked with S51321CBH-7BTT-F, but the actual Elpida part number is EDS51321CBH, which is a 133MHz (CL3-3-3) mobile RAM component. This 64MB cache buffer helps improve small write-to performance and removes the 'stuttering' effect from all Indilinx Barefoot-based SSD products.
In the next section, Benchmark Reviews begins performance testing the OCZ Agility-EX OCZSSD2-1AGTEX60G SSD, and we determine just how well the new Indilinx Barefoot SLC Solid State Drive compares to the current best-performing competition. SSD Testing MethodologySolid State Drives have traveled a long winding course to finally get where they are today. Up to this point in technology, there have been several key differences separating Solid State Drives from magnetic rotational Hard Disk Drives. While the DRAM-based buffer size on desktop HDD's has recently reached 32 MB and is ever-increasing, there is still a hefty delay in the initial response time. This is one key area in which flash-based Solid State Drives continually dominates because they lack moving parts to "get up to speed". However the benefits inherent to SSD's have traditionally fallen off once the throughput begins, even though data reads or writes are executed at a high constant rate whereas the HDD tapers off in performance. This makes the average transaction speed of a SSD comparable to the data burst rate mentioned in HDD tests, albeit usually lower than the HDD's speed. Comparing a Solid State Disk to a standard Hard Disk Drives is always relative; even if you're comparing the fastest rotational spindle speeds. One is going to be many times faster in response (SSD's), while the other is usually going to have higher throughput bandwidth (HDD's). Additionally, there are certain factors which can affect the results of a test which we do our best to avoid. EDITORS NOTE: After November 2009 Benchmark Reviews will begin testing storage devices using the Microsoft Windows 7 Operating System. Although the Marvell SATA 6Gb/s (SATA-III) controller is available on select motherboards, our testing indicates that SSD devices perform better on the Intel ICH-10 Southbridge. SSD Testing DisclaimerEarly on in our SSD coverage, Benchmark Reviews published an article which detailed Solid State Drive Benchmark Performance Testing. The research and discussion that went into producing that article changed the way we now test SSD products. Our previous perceptions of this technology were lost on one particular difference: the wear leveling algorithm that makes data a moving target. Without conclusive linear bandwidth testing or some other method of total-capacity testing, our previous performance results were rough estimates at best. Our test results were obtained after each SSD had been prepared using DISKPART or Sanitary Erase tools. As a word of caution, applications such as these offer immediate but temporary restoration of original 'pristine' performance levels. In our tests, we discovered that the maximum performance results (charted) would decay as subsequent tests were performed. SSDs attached to TRIM enabled Operating Systems will benefit from continuously refreshed performance, whereas older O/S's will require a garbage collection (GC) tool to avoid 'dirty NAND' performance degradation. It's critically important to understand that no software for the Microsoft Windows platform can accurately measure SSD performance in a comparable fashion. Synthetic benchmark tools such as HD Tach and PCMark are helpful indicators, but should not be considered the ultimate determining factor. That factor should be measured in actual user experience of real-world applications. Benchmark Reviews includes both bandwidth benchmarks and application speed tests to present a conclusive measurement of product performance. Test System
Drive Hardware TestedThe following storage hardware has been used in our benchmark performance testing, and may be included in portions of this article:
Test Tools
System Speed TestI doubt that when DOS was put to rest, Vladimir Afanasiev ever thought he would see his System Speed Test software used again in professional reviews. This program offers comprehensive system information, but it also has a powerful benchmarking tool for memory, processor, and disks. In terms of disk performance, it measures interface and physical transfer rates, seek and access times at the hardware level, and it does so without delay or interference from Operating System software or running processes. This is why Benchmark Reviews will continue to use this test: it polls its results directly from the hardware layer without the need for an Operating System (such as Windows). System Speed Test does not require a partition to be present for testing, so all of our Random Access Time benchmarks are completed prior to drive formatting. To detect the Random Access Time, each device runs the full test routine a total of five times. The highest and lowest scores were ignored, and the remainder was averaged. This would be prove pointless however, because the access time benchmark for every single SSD recorded identical test results between runs.
Using the System Speed Test software, the top Random Access Time benchmarks places the OCZ Agility-EX, Vertex-EX, Mtron Pro 7500, and ACARD ANS-9010 at the very top of our results. The 0.9 ms performers include: MemoRight GT, Intel X25-E Extreme, Mtron Pro 7000, Mtron MOBI 3500, Intel 80GB X25-M, and the OCZ Summit. With a Random Access Time of 10ms, the following SSDs were included: Corsair X256, OCZ Vertex Turbo, OCZ Vertex, Patriot Torqx, Super Talent UltraDrive ME, Crucial M225, OCZ Agility, and lastly the Mtron MOBI 3000. With a lightning-fast sub 0.1 ms access time, every other SSD is forced to live in the shadow that these nearly-instant products have just created. Some of the slower Random Access Times include the following SSD products: Kingston SSDNow V+ Series (Samsung PB22-J) with 0.14 ms, The Silicon Power SLC SATA-II SSD SP032GBSSD750S25 with 0.17 ms, the OCZ Apex at 0.18 ms, the Patriot Warp v2 with 0.19 ms, and finally the G.Skill Titan finishing at 0.21 ms. There were slower SSD products, but obsolescence removed them from our results. The DRAM cache buffer is common link between request and response time. Solid State Drive devices have the advantage of a nearly instantaneous NAND storage bank responding to an even faster DRAM buffer. Conversely, Hard Disk Drive products depend on a fast spindle speed to reduce the delay before the buffer transmits data. Although the SSDs at the slower end of our Response Time chart may seem less impressive, in reality you couldn't begin to perceive these subtle differences in real-world applications. Still, the slowest SSD product (0.51ms) is 14x more responsive than the fastest desktop hard drive. Hard Disk Drive alternatives are much slower to react, regardless of spindle speed and cache buffer size. Even the very best of the desktop hard drive products, such as Western Digital's VelociRaptor, only produced a best response time of 7.15 ms. The older Western Digital Raptor took 8.53ms to respond, followed by 12.99ms for the Seagate 7200.11, and 15.39ms for the 7200 RPM Hitachi Travelstar 7K100 notebook drive. The worst performer was the standard 5400 RPM notebook drive (Hitachi Travelstar 5K160 HTS541640J9SA00), which recorded a painfully slow 17.41ms Random Access Time. Drive Hardware
ATTO Disk BenchmarkThe ATTO Disk Benchmark program is free, and offers a comprehensive set of test variables to work with. In terms of disk performance, it measures interface transfer rates at various intervals for a user-specified length and then reports read and write speeds for these spot-tests. There are some minor improvements made to the 2.34 version of the program, but the benchmark is still limited to non-linear samples up to 256MB. ATTO Disk Benchmark requires that an active partition be set on the drive being tested. While the bandwidth results are not realistic for determining the maximum drive speeds, ATTO Disk Benchmark is still a good tool for illustrating bandwidth using various file size chunks. Please consider the results displayed by this benchmark to be basic bandwidth performance indicators.
Our basic I/O bandwidth tests begin with the OCZ Agility-EX SSD connected to the integrated Intel ICH10R Southbridge chip, as the ATTO Disk Benchmark tools performs file transfers ranging from 0.5 KB to 8192 KB. The 64GB OCZ OCZSSD2-1AGTEX60G model reveals a 261 MBps read plateau from 128-8192 KB file chunks, while the 175 MBps write performance plateaus from 128-8192 KB.
Drive Hardware
In our next section, Benchmark Reviews compares random access IOPS performance among high-end storage devices using HD Tune Pro... HD Tune Pro BenchmarksIn the past, Benchmark Reviews has avoided HD Tune benchmarks because the software was so similar to others already being used in our articles. However, EFD Software has released several versions of this program, which now adds functionality and features not available in previous revisions. The latest edition of HD Tune Pro allows random access read and write testing, a feature not available to other similar software benchmark tools. HD Tune is a low-level test that will not operate on a drive which contains a partition, so Benchmark Reviews uses DISKPART to prepare hardware and remove any partitions before conducting these tests. Random Access tests are divided into 512b, 4KB, 64KB, 1MB and random size test files sizes. The Random Access test measures the performance of random read or write operations. The amount of data which will be read varies from 512 bytes to 1 MB. Performance is reported in operations per second (IOPS), average access time, and average speed. Because it is our intent to compare one product against another, Benchmark Reviews has focused on random transfer size IOPS performance.
Benchmark Reviews has tested the 64GB OCZ Agility-EX SLC SSD OCZSSD2-1AGTEX60G against a collection of top-performing desktop storage drives for our random IOPS benchmarks. The 4 KB random IOPS performance in HD Tune measured 7482 for read IO, and 9903 for write. In comparison, the 4KB read IOPS was only slightly better than the Corsair X256 SSD which scored 7444, and the Vertex Turbo's 7272. Both the Vertex Turbo (17358 4KB write IOPS) and Corsair X256 (16244) SSDs were well ahead of write performance.
The tight range of IO is an indicator of operational bottlenecks. For example, the WD VelociRaptor WD3000HLFS SATA Hard Disk Drive indicates a read-IOPS range of 10-150 whereas the average SSD might offer 200-1,000. As a direct result, in most cases SSDs will offer a much higher IO over their hard disk counterparts. The random read/write operations per second is charted below:
The OCZ Vertex EX (firmware 1.20) enjoys the benefit of SLC construction that delivers traditionally better IOPS performance than MLC counterparts, and also offers the best measured random IOPS performance of the group with 384 read and 477 write IOPS. The Corsair X256 and OCZ Vertex Turbo (both MLC), along with the SLC Agility EX, all lead the pack of Indilinx-based SSDs which follow behind (Vertex, CT128M225, Torqx, UltraDrive ME, Agility) and performs at near-SLC levels with 327 read and 429 write IOPS. The Intel X25-E Extreme SSD performed well, and produced 337 random read IOPS, and 394 write. Our test results were obtained after each SSD had been prepared using the DISKPART program, and in the case of products using the Indilinx Barefoot controller they were further prepared with the Sanitary Erase application. In our tests, we discovered that the maximum performance results (charted) would decay as subsequent tests were performed. As a word of caution, applications such as Sanitary Erase (SE) and Wiper offer immediate but temporary restoration of original 'pristine' performance levels. Drive Hardware
Benchmark Reviews measures I/O Response Time and IOPS performance using the Iometer tool in our next section... Iometer IOPS PerformanceIometer is an I/O subsystem measurement and characterization tool for single and clustered systems. Iometer does for a computer's I/O subsystem what a dynamometer does for an engine: it measures performance under a controlled load. Iometer was originally developed by the Intel Corporation and formerly known as "Galileo". Intel has discontinued work on Iometer, and has gifted it to the Open Source Development Lab (OSDL). Iometer is both a workload generator (that is, it performs I/O operations in order to stress the system) and a measurement tool (that is, it examines and records the performance of its I/O operations and their impact on the system). It can be configured to emulate the disk or network I/O load of any program or benchmark, or can be used to generate entirely synthetic I/O loads. It can generate and measure loads on single or multiple (networked) systems. Benchmark Reviews has resisted publishing Iometer results because there are hundreds of different configuration variables available, making it impossible to reproduce our tests without having our Iometer configuration file. To measure random I/O response time as well as total I/O's per second, Iometer is set to use 4KB file size chunks over a 100% random sequential distribution. The tests are given a 50% read and 50% write distribution. Our charts show the Read and Write IOPS performance as well as I/O response time (measured in ms). Iometer was configured to test for 120 seconds, and after five tests the average is displayed in our benchmark results. The first tests included random read and write IOPS performance, where a higher I/O is preferred.
In the Random IOPS performance tests the single layer cell (SLC) OCZ Agility EX (3982/3988), Intel X25-E Extreme (3543/3548), and OCZ Vertex EX (3106/3091) outperformed all other products by a wide margin. The OCZ Vertex Turbo SSD rendered 1774 read/1770 write I/O's, while the Corsair recorded 1769/1773, both of which are just slightly ahead of a single Vertex 1.3 SSD that produced 1702 for read and write IOPS. Sharing the same Indilinx Barefoot controller and NAND architecture, the Crucial CT128M225 SSD (Firmware 1571) reported a read and write I/O of 1695. The mainstream OCZ Agility SSD trails behind with 1625/1618 I/O's while the OCZ Summit MLC SSD completed 730/733 I/O's. Every other product thereafter performs far beneath the above-listed products, and are not suggested for high input/output applications. The 64GB Kingston SSDNow V+ (which is a rebranded Samsung PB22-J SSD) produced a meager 150 I/O's with Iometer. While offering better IO than any other desktop hard drive (and most first- and second-generation SSDs), the Western Digital VelociRaptor still fell short on IOPS performance compared with several current-generation SSDs and produced only 134/138 IO's. The Mtron MOBI 3000 performed 107 read and write IOPS, while the Western Digital WD5001AALS rendered 86 and the Seagate 7200.11 completing 77. The Seagate Momentus 5400.6, which is a 5400 RPM notebook hard drive, produced 60/59 IO's. The newer Mtron MOBI 3500 rendered 58 IOPS, which was worse than the older 3000 model. The OCZ Apex struggled to complete 9 IOPS, and its identically-designed G.Skill Titan managed only 8 IOPS. Clearly, the twin RAID-0 JMicron controllers are built for speed and not input/output operations. Next comes the average I/O response time tests...
The Iometer random IOPS average response time test results were nearly an inverse order of the IOPS performance results. It's no surprise that SLC drives perform I/O processes far better than their MLC versions, but that gap is slowly closing as controller technology improves the differences and enhances cache buffer space. The Read/Write IOPS performance for the 64GB OCZ Agility-EX SLC SSD was 0.19/0.06 ms, with the Intel X25-E Extreme SSD measuring 0.22/0.06 ms, while the OCZ Vertex EX (firmware 1.20) achieved 0.26/0.06 ms. Both of these premium MLC products share a dramatic lead ahead of the other SSDs tested. The Corsair X256 and OCZ Vertex Turbo SSD both scored 0.50/0.06ms, while the Vertex 1.30 and Crucial CT128M225 SSD both offered 0.52/0.06ms. The mainstream Agility SSD produced 0.55/0.06ms, and the OCZ Summit responded to read requests in 0.78ms while write requests were a bit quicker at 0.59ms. Kingston's SSDNow V+ (Intel X25-M) produced 3.50/3.14 ms. The Western Digital VelociRaptor did very well compared against SSD products, producing 6.59/0.82ms. These times were collectively the best available, as each product measured hereafter performed much slower. The Mtron MOBI 3000 offered a fast 0.42ms read response time, but suffered a slower 8.97ms write response. Both the WD5001AALS and Seagate 7200.11 hard drives performed around 11ms read and 1.2ms write. The Seagate Momentus 5400.6 offered 15.3/1.36ms response times. Mtron's newer MOBI 3500 offered great read response times at 0.19ms, but suffered poor write responses at 17.19ms. The worst was yet to come, as the G.Skill Titan and OCZ Apex offered decent 0.42ms read response times but absolutely unacceptable 127ms write times. Drive Hardware
In our next section, we test linear read and write bandwidth performance and compare its speed against several other top storage products using EVEREST Disk Benchmark. Benchmark Reviews feels that linear tests are excellent for rating SSDs, however HDDs are put at a disadvantage with these tests whenever capacity is high. EVEREST Disk BenchmarkMany enthusiasts are familiar with the EVEREST benchmark suite by Lavalys, but very few are aware of the Disk Benchmark tool available inside the program. The EVEREST Disk Benchmark performs linear read and write bandwidth tests on each drive, and can be configured to use file chunk sizes up to 1MB (which speeds up testing and minimizes jitter in the waveform). Because of the full sector-by-sector nature of linear testing, Benchmark Reviews endorses this method for testing SSD products, as detailed in our Solid State Drive Benchmark Performance Testing article. However, Hard Disk Drive products suffer a lower average bandwidth as the capacity draws linear read/write speed down into the inner-portion of the disk platter. EVEREST Disk Benchmark does not require a partition to be present for testing, so all of our benchmarks are completed prior to drive formatting. The high-performance storage products tested with EVEREST Disk Benchmark are connected to the Intel ICH10R SATA controller resident on the Gigabyte GA-EX58-UD4P motherboard. Using the 1MB block size, read performance of the OCZ Agility-EX SLC SSD OCZSSD2-1AGTEX60G measured an average 245.8 MBps with a maximum peak of 246.3 MBps. Linear write-to tests were next...
Linear disk benchmarks are superior tools in my opinion, because they scan from the first physical sector to the last. A side affect of many linear write-performance test tools is that the data is erased as it writes to every sector on the drive. Normally this isn't an issue, but it has been shown that partition tables will occasionally play a role in overall SSD performance (HDDs are more mature products and don't suffer this problem). The 64MB buffer and fast NAND flash memory on the OCZ Agility-EX Indilinx-based SLC Solid State Drive helped improve results in our linear testing, as shown in the waveform chart below. Although the chart makes the linear write performance appear unsteady, the results seen here are actually decent compared to most other SSD products we've tested in the past. The OCZ Agility EX SSD recorded an average linear write-to speed of 234.0 MBps , with a maximum performance of 241.5 MBps.
The chart below shows the average linear read and write bandwidth for a cross-section of other SATA drives attached to the Intel ICH10 Southbridge:
Linear bandwidth certainly benefits the Solid State Drive, since there's very little fluctuation in transfer speed. Hard Disk Drive products decline in performance as the spindle reaches the inner-most sectors on the magnetic platter. I personally consider linear tests to be the single most important comparison of storage drive products, although hard disk drive products decrease performance as they reach the edge of the spindle, SSD products operate at a relatively smooth speed from start to finish. Drive Hardware
In the next section, Benchmark Reviews tests sequential performance using the CrystalDiskMark software tool... CrystalDiskMark TestsCrystalDiskMark is a very basic read and write benchmark tool by Crystal Dew World that offers performance speed results using sequential, 512KB random, and 4KB random samples. For our tests, sequential read and write performance was measured using a 1000MB file size, with 50, 100, and 500MB being the other available options. CrystalDiskMark requires that an active partition be set on the drive being tested, and all drives are formatted with NTFS. Benchmark Reviews uses CrystalDiskMark to confirm manufacturer suggested bandwidth speeds. In addition to our other tests, the sequential read and write benchmarks allow us to determine if the maximum stated speed of any storage product is within reasonable specification. In the chart below illustrated below, our sequential read and write performance speeds are organized from highest to lowest based on total bandwidth. Enjoying a noticeable lead atop of our sequential performance chart, both the Intel X25-E Extreme (261/206 MBps) and the OCZ Vertex EX Single-Layer Cell SSD (256/182 MBps) and OCZ Agility EX (258/172 MBps) offer the highest read and write bandwidth performance. Followed closely behind was the OCZ Summit MLC SSD with 128MB cache buffer that produced 227/191 MBps, and the Corsair X256 which scored 255/156. Not far behind the top leaders was the OCZ Apex, and it's twin cousin the G.Skill Titan (not shown). The OCZ Vertex Turbo offered 248/145, which is slightly faster than the Kingston SSDNow V+ and standard Vertex SSD performance of 230/138. The Crucial CT128M225 is essentially identical to the OCZ Vertex, as is the Patriot Torqx and Super Talent UltraDrive ME. The OCZ Agility followed closely behind the other Indilinx Barefoot SSDs and shared the same sequential write speed penalty. Although the remaining drives are much older models, the SLC construction keeps the OCZSSD2, Mtron MOBI 3500, and MOBI 3000 all within respectable range. Even though the high-performance VelociRaptor hard drive is made to look low-end by the results charted below, it's actually very encouraging to see that a hard disk can still keep pace with a few SSDs. The WD VelociRaptor offers sequential read and write performance comparable to the Mtron MOBI 3500 and the first-generation OCZ SLC SSD.
Drive Hardware
Benchmark Reviews tests and compares buffered transaction speed using HD Tach in the following section... HD Tach RW ResultsAlthough HD Tach (and also HD Tune or Crystal Disk Benchmark) are all excellent tools for measuring Hard Disk Drive products, they fail to offer the same precision with Solid State Drive products. These programs offer only an approximate estimate of bandwidth speed through their quick-result sample-testing mechanisms, as I have proven in the Solid State Drive (SSD) Benchmark Performance Testing article published not long ago. Nevertheless, HD Tach is still useful for offering an alternative perspective at performance, even if it isn't precisely correct when used with SSD architecture. HD Tach is a software program for Microsoft Windows that tests the sequential read, random access and interface burst speeds of the attached storage device. For the record. every single product tested was brand new and never used. HD Tach allows write-bandwidth tests only if no partition is present. Additionally, each and every product was tested five times with the highest and lowest results removed before having the average result displayed here. The graphical user interface (GUI) of the Windows-based benchmark tool HD Tach is very convenient. and allows the test product to be compared against others collected on your system or those registered into the Simpli Software database. HD Tach will not test write performance if a partition is present, so all of our benchmarks are completed prior to drive formatting. In the tests below, Benchmark Reviews utilizes the HD TachRW tool to compare the fastest collection of desktop hard drives and competing SSD's we can get our hands on. Using the Intel ICH10R SATA controller on the Gigabyte GA-EX58-UD4P, HD Tach was used to benchmark the test SSD five times with the best results displayed below. It's important to note that HD Tach's Burst Speed result should be ignored for Solid State Drives due to the cache methods inherent to each memory controller architecture. There are times where this number will be extremely high, which is a result of the optimized cache used for SSD's. The important numbers used for comparison are the sustained read and write bandwidth speeds, which indicate an approximate performance level of the product. Our featured test item, the 64GB OCZ Agility-EX OCZSSD2-1AGTEX60G Indilinx-based SLC Solid State Drive, performed at an average 232.1 MBps best sustained average read speed and a best average sustained write bandwidth of 205.0 MBps. For comparison, these results were nearly identical to the Corsair X256 MLC SSD. Using the Secure Erase tool after each test kept our results consistent and reliable, however used 'dirty' SSD will show significant performance degradation.
The chart below illustrates the collected averages for benchmark results using HD Tach RW on the Intel ICH10 SATA controller, with the read and write bandwidth results added together to determine rank placement. The first group is a collection of high-performance storage products. With an improved write performance, the SLC Vertex EX finds itself positioned in first place with the Corsair X256, Intel X25-E Extreme, and OCZ Agility EX SSDs all tied second, ahead of the MLC Vertex Turbo and several other Indilinx Barefoot-based SSDs. (Crucial M225, OCZ Vertex, Patriot Torqx, Super Talent UltraDrive, and Agility). A set of Western Digital VelociRaptor hard drives paired into a RAID-0 stripe array come close, but just aren't enough to stave off SSD performance. The ACARD ANS-9010 RAM-Disk, OCZ Summit SSD, OCZ Agility SSD, OCZ Apex SSD, Intel X25-M, and Kingston SSDNow V+ SSD all trail behind in overall performance and finish out the top five positions. Nearly every other storage product trails distantly behind these leaders, which all recorded a combined HD Tach bandwidth to over 300 MBps for each. A single (non RAID-0) VelociRaptor and Seagate 7200.11 hard drive begin the next segment of upper midrange performers, offering nearly 200 MBps of combined bandwidth. Trailed by a closely-packed group consisting of the Patriot Warp v2 SSD, Silicon Power SP032GBSSD750S25, and G.Skill FM-25S2S-64GB, are SSDs generating between 168-172 MBps of combined average bandwidth.
The lower-midrange SSD products begin with the Western Digital Raptor, scoring a combined total bandwidth of 154 MBps delivering half the performance of the leaders. Yesterday's high-performance SSD is today's low-performance drive, and the Mtron MOBI 3500, OCZ OCZSSD2-1S32G SSD, Super Talent MasterDrive MX SSD and Mtron MOBI 3000 all comprise products with less combined performance than Hard Disk Drive alternatives (except in regard to response time). Drive Hardware
Please continue on to the Final Thoughts sections, where the state of SSD testing and Data Storage market sector are put under analysis... SSD vs Hard Disk DriveThe last days of old technology are always better than the first days of new technology. Never has this saying been more true than with the topic of storage technology, specifically in regard to the introduction of Solid State Drive technology a few years ago. The only things standing in the way of widespread Solid State Drive (SSD) adoption are high storage capacity and affordable price of Hard Disk Drive (HDD) devices. Because NAND flash-based SSD technology costs more per gigabyte of capacity than traditional magnetic hard drives, the benefits of immediate response time, transfer speeds, and operational input/output performance often get overlooked. Like most consumer products, it wasn't a question of how much improvement was evident in the new technology, it was price. I'll discuss product costs more in just a moment, but for now consider how each new series of SSD product employs greater performance than the one before it, convincing would-be consumers into waiting for the right time to buy.
There's also a gray area surrounding SSD performance benchmarks that has me concerned. You might not know this, but SSDs can be very temperamental towards the condition of their flash NAND. My experience testing dozens of Solid State Drives is that a freshly cleaned device (using an alignment tool) will always outperform the same device once it's been formatted and used. A perfect example is Indilinx Barefoot-based SSDs, which suffers severely degraded performance when writing to 'dirty' flash NAND. The reason that all of this will matters is simple: the performance results reported to consumers in product reviews (such as this one) often report the very best performance scores, and the process used to obtain these results is not applicable to real-world usage. This is where garbage collection techniques such as TRIM become important, so that end-users will experience the same performance levels as we do in our tests.
Garbage Collection (GC) is the current solution for keeping flash NAND in 'clean' condition, while maintaining optimal performance. Windows 7 offers native TRIM support, and most retail SSDs also include this special GC function or at least offer a firmware update that brings the drive up-to-date. For anyone using an Operating System or SSD that does not offer Garbage Collection functionality, you'll be using 'dirty' flash NAND modules and suffering sub-optimal performance for each write-to request. A few SSD manufacturers offers free tools to help restore peak-level performance by scheduling GC to 'clean' used NAND sectors, but these tools add excessive wear to the NAND the same way disk defragmenting tools would. SLC flash modules may resist wear much better than MLC counterparts, but come at the expense of increased production cost. The best solution is a more durable NAND module that offers long-lasting SLC benefits at the cost of MLC construction. Adoption is further stalled because keen consumers aware of this dilemma further continue their delay into the SSD market. Getting back to price, the changes in cost per gigabyte have come as often as changes to the technology itself. At their inception, high-performance models such the 32GB MemoRight GT cost $33 per gigabyte while the entry-level 32GB Mtron MOBI 3000 sold for $14 per gigabyte. While an enjoyable decline in NAND component costs forced consumer SSD prices down low in 2009, the price of SSD products has been on the rise during 2010. Nevertheless, Solid State Drives continue to fill store shelves despite price or capacity, and there are a few SSD products now costing only $2.03 per gigabyte. Although the performance may justify the price, which is getting dangerously close to the $1.00 per gigabyte WD VelociRaptor hard drive, costs may still close some buyers out of the market. Price notwithstanding, the future is in SSD technology and the day when HDDs are obsolete is nearing; but there are still a few bumps in the road to navigate. OCZ Agility EX ConclusionBenchmark Reviews begins our conclusion with a short summary for each of the areas that we rate. The first section is performance, which considers how effective the OCZ Agility EX performs in operations against direct competitor products. For reference, OCZ specifies that the 64GB OCZSSD2-1AGTEX60G Solid State Drive should offer a maximum 255 MBps read and 195 MBps write performance. In many of our benchmark tests, the Agility-EX SSD performed at or above this rating. The Agility EX offered its highest performance of 246/234 MBps read/write in Everest, followed by 258/193 MBps in Crystal DiskMark, then 232/205 MBps in HD-Tach, and finally ATTO Benchmark scored 261/175 MBps. It's important to note that 'clean' NAND offered vastly better performance results than those with used NAND sectors on this SLC drive. IOPS performance was extremely high in Iometer, with the Agility EX (firmware 1.31) out-producing all other SSDs tested (including the Intel X25-E Extreme). In HD Tune there was better 4K IOPS out of the Corsair X256 or OCZ Turbo, but the Agility EX prevailed by a large margin in the random IO tests.
Manufacturers have been forced to use some creative ideas to help lure consumers to their product. SSD technology already carries a premium price tag over the alternative, which is why product presentation becomes so important. OCZ takes the mainstream Agility SSD series packaging back to basics with a simple-yet-catchy green appearance, while at the same time delivering critical product information and specifications for the untrained consumer. Solid State Drives are a lot like spark plugs: you see them just long enough to install, and then they're forgotten. OCZ keeps production costs down on the Agility EX-series SSD with a uniform black painted enclosure identical to all of their other SSD products, and uses an adhesive label for each product series with underside specifications. Because Solid State Drives are like their Hard Disk Drive counterparts in that they place function before fashion, anything above and beyond a simple metal shell is more than what's expected of the appearance. To this end, I still prefer that manufacturers utilize sealed plastic enclosures (such as the MOBI 3000) to prevent moisture or electrical shock damage; an unlikely trend to return to the industry. Construction is probably the strongest feature credited to the entire SSD product line, and OCZ products have never been an exception. Solid State Drives are by nature immune to most abuses, but add to this a hard metal shell and you have to wonder what it would take to make this drive fail. If an OCZ Agility EX SSD Series product does happen to fail during the 3-year warranty period, end-users can contact OCZ via the company website or extensive support forums. Fortunately, there's also a toll-free telephone number for support or customer service questions (800-459-1816). Although I would personally rate the Indilinx IDX110M00-FC chip among the best desktop SSD controllers available (price and performance are superior to the latest Intel and Samsung controllers), the 'Barefoot' product series is the most particular controller Benchmark Reviews has ever tested. A perfect example is the difference between performance benchmark results, which are only consistent if you ensure every test uses 'clean' NAND. Benchmark Reviews re-tested the collection of Indilinx-based SSDs after NAND cleaning tools such as Sanitary Erase (SE) were made available, and there was a night-and-day difference between results. While garbage collection (GC) functionality is in development for upcoming Indilinx firmware updates, there currently isn't any TRIM support available at the SSD level, which means you'll be using 'dirty' NAND and suffering sub-optimal performance even in a Windows 7 environment. There are tools like SSD Wiper which are meant to help restore peak-level performance by scheduling GC to 'clean' used NAND sectors, but this kind of tool adds the same excessive wear to NAND components that disk defragmenting poses. In the end, controllers incorporating a native TRIM function will be the solution, and OCZ will likely be the first to offer it. Unfortunately, it isn't here yet. As of October 2009, only the 64GB (60GB advertised) capacity is available for the Agility EX series. The OCZ OCZSSD2-1AGTEX60G Solid State Drive is currently sold at NewEgg for $399. With the exception of NAND speed the OCZ Agility EX shares identical construction with the Vertex EX, which performed at or below the Agility-EX despite a rating that gives it a 5/15 MBps read and write advantage. Regardless, the Agility EX is the obvious choice over the Vertex EX when you consider price, which has the 64GB OCZSSD2-1VTXEX60G selling for $659, amounting to an unrealistic $260 expense for similar performance results. Unless SLC is a mission requirement, there are several SLC options to consider. The 32GB OCZ Turbo OCZSSD2-1VTXT30G model sells for $155, while the 64GB OCZSSD2-1VTXT60G model is listed for $249. A larger 128GB Vertex Turbo OCZSSD2-1VTXT120G lists for $439, allowing double the capacity of the Agility EX at the expense of SLC NAND. The added expense of SLC construction is considered minimal in comparison to the added NAND lifetime it provides for mission-critical enterprise server systems. Since the Agility EX is now the most affordable SLC Solid State Drive on the market, it seems logical that corporate environments will take advantage. Nevertheless, more than a few retail consumers might also choose to pay for the added security of Single Layer Cell construction. In terms of performance and functionality, OCZ needs to hustle and offer firmware-based TRIM support, thereby solving 'dirty' NAND performance issues for the early adopters of Microsoft's Windows 7 O/S. Regardless, OCZ has brought the price of SLC SSDs down to nearly $6/GB with the Agility EX, something many MLC products have yet to do. The performance is on-par with the fastest MLC SSDs, such as the Corsair X256 and OCZ Vertex Turbo, but the Agility-EX also dramatically outperforms the ultra-premium SLC Intel X25-E Extreme. The Indilinx Barefoot controller has proven itself to be a very popular chip for SSD makers, and the OCZ Agility EX exploits the highest order of SLC performance. It may take some extra effort to maintain pristine-level NAND performance on a full-time basis until native TRIM support in enabled, but the SLC SSD market has never seen a faster portable storage product for the money. The OCZ Agility EX-series SSD is a high-performance storage product recommended for Enterprise servers and performance enthusiasts wanting the absolute best performance from their computer system. Pros:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
