Kingston HyperX Savage 240GB SSD Review

[SKYMTL]

With both their enthusiast and entry level models recently refreshed it comes as little surprise to see Kingston round out their new lean and mean lineup with a more mainstream orientated option. That option is new new HyerX Savage. In fact it is safe to say that their mainstream refresh was long overdue as their mid-tier models were flagging in both comparable performance and sales. What was not expected was the controller that Kingston has opted for.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/cont_sm.jpg" border="0" alt="" /></div>

In previous years Kingston opted for a wide variety of controllers ranging from Intel to SandForce and everything in between. The one thing all their new models had in common however was a relatively conservative controller selection as they avoided suppliers who didn't didn't have a proven track record. After the HyperX Predator and its ground breaking AltaPlus controller surprised everyone, Kingston has now turned to Phison for their Savage series.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/intro.jpg" border="0" alt="" /></div>

This isn't the first time we have come across Phison powering a mainstream SSD, but it is still a very unusual choice since these controllers have typically powered value-oriented products. This time however they have introduced a new controller named the PS3110 which is completely unique in the way it goes about things. It is the result of outside the box thinking that on the surface certainly seems to have merit. We will go over it in greater detail a bit later but suffice to say its approach to load balancing is radically different.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/top_sm.jpg" border="0" alt="" /></div>

Kingston's HyperX line of SSDs has always been quite aesthetically pleasing and the new Savage is no exception. Everything from its 7mm form-factor (with a 2.5mm adapter included) to its large red 'X' with 'Savage' imprinted in the black border simply exudes a sense of performance.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/board1_sm.jpg" border="0" alt="" /></div>

With so many types of NAND to choose from we were pleasantly surprised to see that Kingston opted for Toshiba A19 NAND ICs.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/nand_sm.jpg" border="0" alt="" /></div>

Toshiba's second generation 19nm Toggle Mode NAND's performance characteristics make it a great match for mainstream consumers and it is one of our top recommendations. With 16 ICs, the PS3110 controller not only has all 8 channels fully populated but also has good NAND interleaving per channel.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/ram_sm.jpg" border="0" alt="" /></div>

This NAND in conjunction with the Phison PS3110 controller, backstopped by a moderately sized 256MB DDR3-1600 ram buffer is precisely how Kingston intends to take back some of their lost market share. With that being said an asking price of $140-$150 for the 240GB model the Savage does indeed have its work cut out for it. This corner of the mainstream market is filled with excellent drives with more proven controllers but if any company can take on the likes of Crucial and their MX200 series, or even OCZ and their Vertex 460 series, it is Kingston's Savage.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/access_sm.jpg" border="0" alt="" /></div>

Before we move on to the meat of the review we do feel it necessary to comment on the included accessories that accompany certain Savage models. As with previous generations Kingston offers a 'bare bones' model that consists of only the drive, and a model that is oriented towards making the transition from your old OS drive to your new one as painless an experience as possible.

Our sample is the latter and while it costs $10 more than the barebones version you certainly do get value for your money. For ten bucks you get a 2.5mm z-height adapter, a new SATA 6Gb/s cable, a 2.5" to 3.5" adapter plate, Acronis True Image HD imaging software, a micro-screwdriver set, and a USB 3.0 enclosure with USB 3.0 cable. This last addition means you can simply plug the Savage into the enclosure, plug the enclosure into your system, run Acronis HD and clone your existing drive to the new SSD. This will be a great boon to consumers whose systems only have one SATA port - such as laptops.

Of course, Acronis HD does have a few issues with cloning UEFI systems (especially Windows 8 systems) so things may not go as smoothly as Kingston hopes, but this is still one damn fine accessories list.

 

[SKYMTL]

Introducing the Phison PS 3110 Controller

Introducing the Phison PS 3110 Controller

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/Phison-logo.jpg" border="0" alt="" /></div>

Phison may not be a household name but that does not mean they are newcomers to the solid state storage industry. Rather they are a company that has been around for a long time but only recently decided to divert their focus from the OEM market. Up until their PS3108 controller, Phison were best known for either being the being the first company to be able to offer OEMs a single-chip USB flash drive IC, or being one of the founders of the Open NAND Flash Interface (ONFI) group.

What few consumers realize is that since their founding in late 2000 Phison has shipped over half a <i>billion</i> NAND flash controllers. Needless to say they may not have name brand recognition in the consumer marketplace, but they are a highly respected company behind the scenes.

Unlike competitors like Indilinx or Marvel, Phison's first generation SATA AHCI NAND flash controller was a touch underwhelming and never gained much traction outside the extreme value end of the spectrum. Considering the experience and patent portfolio of Phison such a misstep must have been very irksome to them. This is why they went back to the drawing board and started with a completely fresh approach to controller design. The end result of their hard work is the new PS3110 controller.

On the surface this controller may not appear to be all that unique as it is 55nm, AHCI, SATA 6Gb/s based design that offers seemingly middle of the road abilities. In fact 550/530 MBps read and write performance with 100K/90K random IOPS is fairly typical for modern controllers. Being an eight channel design is also fairly typical and nothing standout. Of course compared to their previous generation '108 these specifications are a major step up. However, simple specifications such as maximum sequential performance, and even random IOPs rarely tell the full story and the PS3110 is a perfect example of why consumers should never judge by such basic specifications.

Going hand in hand with the increased channels and IOPS is a different approach to load balancing. In most designs, all cores in the controller are continuously conducting a delicate balancing act between garbage collection, other low level tasks and real time I/O requests. This is a big portion of what firmware refinement boils down to: modifying the amount of cycles the controller dedicates towards specific tasks.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/ps3110.jpg" border="0" alt="" /></div>

Compare and contrast the standard approach with Phison's approach: rather than load balance cycles at the firmware level, Phison has dedicated a whole processor core for real time requests and then sets the other <i>three</i> cores for garbage collection, and other low level tasks. The only other company's controller which does something even remotely similar is OCZ's Indilinx Barefoot 3 which uses an additional core for specific low level tasks. However, the Phison and PS3110 are the first to take load balancing to such extremes.

On the surface such a move is highly controversial as common sense would lead you to believe that real time IO requests are more important than low level requests. The reality is not so cut and dry, and the opposite is actually closer to the typical scenarios most consumer grade controllers will encounter. The fact of the matter is home users rarely have deep queue depth IO requests and instead the majority of the time the controller has to deal with internal housekeeping measures to keep the data available, the free NAND in a virgin state, and even ensuring the ECC is not corrupted. As such having one core always on standby for user requests and having the other 3 cores being able to seamlessly do behind the scenes tasks is rather brilliant…or insane. In either case it is unique answer to the age old problem of proper load balancing.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/perf.jpg" border="0" alt="" /></div>

The most obvious benefits to such a configuration is that long term performance drop off should be drastically lower than most controllers. The drive's performance will not be noticeably impacted by ongoing behind the scenes garbage collection and even data reliability will also be greater.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/smartECC.jpg" border="0" alt="" /></div>

To help ensure all data is safe, secure, and not corrupted the PS3110 takes a page from SandForce's RAIN technology and treats all its internal ICs not as one large Logical Unit, rather it treats them as a logical array of independent NAND ICs. Such a configuration provides data redundancy and additional ECC via this 'RAID'. This multi-tiered approach to error correction is what Phison calls SmartECC. SmartECC however is not simply an internal RAID array and instead consists of multiple levels.

Let's break down what happens when the controller encounters corrupted data during a read IO request. When such an occurrence happens the PS3110 immediately does a read retry to ensure that the data is actually corrupted and unreadable. This allows a second chance at reading the data at full performance as ECC does significantly impact real time performance. If the data is still found to be corrupted, the controller then reads the 120bit/2KB BCH ECC that is stored in the spare area NAND - i.e. not in the same NAND ICs as the corrupted data. If the standard BCH ECC is not able to fully repair the corruption the controller then reads the RAID ECC parity to fill in the missing data. It is only if after doing all this that the data is still found to be corrupted that the drive will fail the data and notify the system of the issue.

In addition to this extremely aggressive and impressive data protection Phison's PS3110 also implements what they call SmartRefresh. The typical data retention abilities of consumer grade NAND varies from weeks (TLC) to years (SLC) with MLC being good for about a one year period. Put another way, data that is written and then is left static tends to fade and becomes harder to read the longer it is left alone. This is the nature of having four voltage settings for MLC NAND and <i>eight</i> for TLC NAND as the voltage states the smaller the drift has to be before corruption occurs. In some controllers these highly static NAND cells are mostly left alone and only tested once the data is needed. This laissez faire approach to data retention is one reason why Samsung's 840 Evo series had such wildly varying performance.

Phison on the other hand has taken a much more active approach. Since they have so many free cycles - via the three dedicated cores - for housecleaning and maintenance the PS 3110 will actively and routinely read NAND cells and test the results against its ECC - even when the data is not request by the host system. This way 'fading' data can be caught early, refreshed during low IO periods and thus PS3110 controller based drives will almost never suffer the same fate as Samsung 840 Evo drives.

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/smartFlush.jpg" border="0" alt="" /></div>

Also helping against data corruption, Phison's PS3110 also has two more features called SmartFlush and GuranteedFlush. Much like OCZ's Power Failure Management Plus (PFM+), Phison's SmartFlush minimizes the time that data stays in the ram buffer and is constantly flushing the data to minimize In-Flight data loss from a unexpected loss of power. This is not the same as true Enterprise grade Data Loss Protection but is still nevertheless a nice addition. GuranteedFlush simply means that the PS 3110 is Flush Cache (E7h) command enabled, and as such when the Operating System tells it to flush its buffers so that the system can enter a low power state it will indeed flush its buffer and only when completed will it signal the host controller that it is ready to go to sleep.

Further helping consistent long term performance, Phison has opted for a moderate ~7% of over-provisioning. In the 240GB model this is 16GB while the 480GB models have 22GB set aside. This spare area can be used for everything from bad block replacement to garbage collection. Most importantly, it ensures the controller will always have access to free blocks even when the drive is approaching full capacity. This multi-layer approach to data protection is overkill for the consumer marketplace, but does explain how Kingston is able to give a total bytes written specification of 306TB for their HyperX Savage 240GB SSD.

As with most high performance controllers PS3110 controller also meets TCG’s OPAL standards and has built in auto-encryption with 256-bit AES support. However, much like the Barefoot 3 controller, the PS3110 doesn’t have this feature enabled by default. Auto encryption is not required for the average home user or enthusiast and as such will not be missed; however once again thanks to Phison dedicated three cores to internal duties, implementing encryption on a PS3110 drive should have very little to no noticeable impact on performance.

When taken as a whole this new PS3110 seems to be a capable high performance controller that is also extremely adaptable to ever changing conditions found in the real world. Seeing normally enterprise capabilities rolled into a mass market controller is certainly interesting. More importantly these features should make for a great addition to any company’s current stable of high performance drives and we doubt Kingston will be the last to option the PS3110 in the coming months.

 

[SKYMTL]

Test System & Testing Methodology

Testing Methodology

Testing a drive is not as simple as putting together a bunch of files, dragging them onto folder on the drive in Windows and using a stopwatch to time how long the transfer takes. Rather, there are factors such as read / write speed and data burst speed to take into account. There is also the SATA controller on your motherboard and how well it works with SSDs & HDDs to think about as well. For best results you really need a dedicated hardware RAID controller w/ dedicated RAM for drives to shine. Unfortunately, most people do not have the time, inclination or monetary funds to do this. For this reason our test-bed will be a more standard motherboard with no mods or high end gear added to it. This is to help replicate what you the end user’s experience will be like.

Even when the hardware issues are taken care of the software itself will have a negative or positive impact on the results. As with the hardware end of things, to obtain the absolute best results you do need to tweak your OS setup; however, just like with the hardware solution most people are not going to do this. For this reason our standard OS setup is used. However, except for the Windows 7 load test times we have done our best to eliminate this issue by having the drive tested as a secondary drive. With the main drive being an Intel DC S3700 800GB Solid State Drive.

For synthetic tests we used a combination of the ATTO Disk Benchmark, HDTach, HD Tune, Crystal Disk Benchmark, IOMeter, AS-SSD, Anvil Storage Utilities and PCMark 7.

For real world benchmarks we timed how long a single 10GB rar file took to copy to and then from the devices. We also used 10gb of small files (from 100kb to 200MB) with a total 12,000 files in 400 subfolders.

For all testing a Asus Sabretooth TUF X99 LGA 2011-v3 motherboard was used, running Windows 7 64bit Ultimate edition. All drives were tested using either AHCI mode using Intel RST 10 drivers, or NVMHCI using Intel NVMe drivers.

All tests were run 4 times and average results are represented.

In between each test suite runs (with the exception being IOMeter which was done after every run) the drives are cleaned with either HDDerase, SaniErase or a manufactures 'Toolbox' and then quick formatted to make sure that they were in optimum condition for the next test suite.

Processor: Core i7 5930K
Motherboard: Asus Sabretooth TUF X99
Memory: 16GB G.Skill Ripjaws DDR4
Graphics card: NVIDIA GeForce GTX 780
Hard Drive: Intel DC S3700 800GB, Intel P3700 800GB
Power Supply: XFX 850

SSD FIRMWARE (unless otherwise noted):

OCZ Vertex 2 100GB: 1.33
Vertex 460 240GB: 1.0
Intel 7230 240GB: L2010400
Samsung 840 Pro 256GBXM06B0Q
Plextor M6e 256GB: 1.03
AMD R7 240GB: 1.0
Crucial MX200: MU01
G.Skill Phoenix 480GB: 2.71
Intel 750: 8EV10135
Kingston HyperX Predator 480GB: 0C34L5TA
Kingston HyperX Savage 240GB: SAFM00.r

Samsung MDX controller:
Samsung 840 Pro 256GB- Custom firmware w/ 21nm Toggle Mode NAND

SandForce SF1200 controller:
OCZ Vertex 2 - ONFi 2 NAND

SandForce SF2281 controller:
G.Skill Phoenix 480GB - Custom firmware w/ 128Gbit ONFi 3 NAND

Marvell 9183 controller:
Plextor M6e 256GB- Custom firmware w/ 21nm Toggle Mode NAND

Marvell 9189 controller:
Crucial MX200 - Custom firmware w/ 128Gbit ONFi 3 NAND

Marvell 9293 controller:
Kingston HyperX Predator - Custom firmware w/ 19nm Toggle Mode NAND

Barefoot 3 controller:
AMD R7 (M00) - 19nm Toggle Mode NAND w/ custom firmware
OCZ Arc 100 (M10) - 19nm Toggle Mode NAND

Intel X25 G3 controller:
Intel 730 - Custom firmware w/ ONFi 2 NAND

Intel NVMe G1 Controller:
Intel 750 - Customer firmware w/ MLC 20nm NAND

Phison PS3110 Controller:
Kingston HyperX Savage 240GB - 19nm Toggle Mode NAND

 

[SKYMTL]

Read Bandwidth / Write Performance

Read Bandwidth

<i>For this benchmark, HDTach was used. It shows the potential read speed which you are likely to experience with these hard drives. The long test was run to give a slightly more accurate picture. We don’t put much stock in Burst speed readings and thus we no longer included it. The most important number is the Average Speed number. This number will tell you what to expect from a given drive in normal, day to day operations. The higher the average the faster your entire system will seem.</i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/read.jpg" border="0" alt="" /></div>

Write Performance

<i>For this benchmark HD Tune Pro was used. To run the write benchmark on a drive, you must first remove all partitions from that drive and then and only then will it allow you to run this test. Unlike some other benchmarking utilities the HD Tune Pro writes across the full area of the drive, thus it easily shows any weakness a drive may have.</i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/write.jpg" border="0" alt="" /></div>

While certainly not the fastest performer we have recently seen, the Kingston Savage posted very decent sequential performance numbers. More importantly the amount of variance from maximum to minimum in both tests was impressively small and does point towards Phison being on the right track with their unique load balancing scheme.

 

[SKYMTL]

ATTO Disk Benchmark

ATTO Disk Benchmark

<i>The ATTO disk benchmark tests the drives read and write speeds using gradually larger size files. For these tests, the ATTO program was set to run from its smallest to largest value (.5KB to 8192KB) and the total length was set to 256MB. The test program then spits out an extrapolated performance figure in megabytes per second. </i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/atto_r.jpg" border="0" alt="" />
<img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/atto_w.jpg" border="0" alt="" /></div>

While the small file read performance was a touch lower than we like to see in a mainstream drive these results are still very, very decent. More to the point they are eons better than what any PS3108 controller based drive could ever hope to do.

 

[SKYMTL]

Crystal DiskMark / PCMark 7

Crystal DiskMark

<i>Crystal DiskMark is designed to quickly test the performance of your drives. Currently, the program allows to measure sequential and random read/write speeds; and allows you to set the number of tests iterations to run. We left the number of tests at 5 and size at 100MB. </i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/cdm_r.jpg" border="0" alt="" />
<img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/cdm_w.jpg" border="0" alt="" /></div>

PCMark 7

<i>While there are numerous suites of tests that make up PCMark 7, only one is pertinent: the HDD Suite. The HDD Suite consists of numerous tests that try and replicate real world drive usage. Everything from how long a simulated virus scan takes to complete, to MS Vista start up time to game load time is tested in these core tests; however we do not consider this anything other than just another suite of synthetic tests. For this reason, while each test is scored individually we have opted to include only the overall score.</i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/pcm7.jpg" border="0" alt="" /></div>

The reason the HyperX Savage does not post higher in the Crystal DiskMark and PCMark7 results is fairly simple to explain, and yet does highlight a weakness of Phison PS3110's unique load balancing. By only dedicating one core for real time IO requests this controller quickly gets saturated and has to rely more upon its cache buffer than we would like. You can easily see this happening with deeper queue depths and explains why even though its single queue depth results are good, the deep queue depth results are at best mediocre.

This is the downside to dedicating 75% of a controllers power to long term stability: the controller can quickly find itself in over its head and have no way to quickly dig itself out of a hole of its own making. Phison and Kingston may want to revisit their 1 in 4 core idea and possibly tweak the firmware so that a second core can be diverted in an 'as needed' basis.

 

[SKYMTL]

AS-SSD / Anvil Storage Utilities Pro

AS-SSD

<i>AS-SSD is designed to quickly test the performance of your drives. Currently, the program allows to measure sequential and small 4K read/write speeds as well as 4K file speed at a queue depth of 6. While its primary goal is to accurately test Solid State Drives, it does equally well on all storage mediums it just takes longer to run each test as each test reads or writes 1GB of data.</i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/asd_r.jpg" border="0" alt="" />
<img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/asd_w.jpg" border="0" alt="" /></div>

Anvil Storage Utilities Pro

<i>Much like AS-SSD, Anvil Pro was created to quickly and easily – yet accurately – test your drives. While it is still in the Beta stages it is a versatile and powerful little program. Currently it can test numerous read / write scenarios but two in particular stand out for us: 4K queue depth of 4 and 4K queue depth of 16. A queue depth of four along with 4K sectors can be equated to what most users will experience in an OS scenario while 16 depth will be encountered only by power users and the like. We have also included the 4k queue depth 1 results to help put these two other numbers in their proper perspective. All settings were left in their default states and the test size was set to 1GB.</i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/anvil_r.jpg" border="0" alt="" />
<img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/anvil_w.jpg" border="0" alt="" /></div>

Once again this new Phison controller has <i>tons</i> of potential and yet is being held back by its own firmware. As you can see at low queue depths it is smooth sailing, but once things get deep it simply flounders faster than a brand new <i>quad core</i> controller should. On the positive side this PS3110 controller can hold its own at even 16 queue depths and this should be more than enough for home user scenarios.

 

[SKYMTL]

IOMETER

IOMETER

<i>IOMeter is heavily weighted towards the server end of things, and since we here at HWC are more End User centric we will be setting and judging the results of IOMeter a little bit differently than most. To test each drive we ran 5 test runs per HDD (1,4,16,64,128 queue depth) each test having 8 parts, each part lasting 10 min w/ an additional 20 second ramp up. The 8 subparts were set to run 100% random, 80% read 20% write; testing 512b, 1k, 2k,4k,8k,16k,32k,64k size chunks of data. When each test is finished IOMeter spits out a report, in that reports each of the 8 subtests are given a score in I/Os per second. We then take these 8 numbers add them together and divide by 8. This gives us an average score for that particular queue depth that is heavily weighted for single user environments.</i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/iom.jpg" border="0" alt="" /></div>

This perfectly sums up why the Phison PS3110 in its current form will never be seen in the business environment, it just simply does not have the deep queue depth performance to handle such tasks. By the same token the Kingston HyperX Savage is not designed nor marketed towards anyone besides the home user and it is hard to find fault with shallow IOMeter results.

 

[SKYMTL]

Windows / Adobe CS5 Load Time

Windows 8.1 Start Up with Boot Time A/V Scan Performance

<i>When it comes to hard drive performance there is one area that even the most oblivious user notices: how long it takes to load the Operating System. We have chosen Windows 8.1 64bit Pro as our Operating System with all 'fast boot' options disabled in the BIOS. In previous load time tests we would use the Anti-Virus splash screen as our finish line; this however is no longer the case. We have not only added in a secondary Anti-Virus to load on startup, but also an anti-malware program. We have set Super Anti-Spyware to initiate a quick scan on Windows start-up and the completion of the quick scan will be our new finish line. </i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/boot.jpg" border="0" alt="" />
</div>

Adobe CS5 Load Time

<i>Photoshop is a notoriously slow loading program under the best of circumstances, and while the latest version is actually pretty decent, when you add in a bunch of extra brushes and the such you get a really great torture test which can bring even the best of the best to their knees. Let’s see how our review unit fared in the newly updated Adobe crucible! </i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/adobe.jpg" border="0" alt="" />
</div>

Unlike the synthetic test results, the real world numbers are much more favorable to the Kingston HyperX Savage and its controller's unique way of load balancing. Of course these tests consist of mainly shallow queue depth IO requests and as such do not stress this controller nearly as much as some real world tasks will. By the same token, most home users do not stress their storage subsystem all that much so an argument could be made that Phison is on the right track.

 

[SKYMTL]

Firefox Performance / Real World Data Transfers

Firefox Portable Offline Performance

<i>Firefox is notorious for being slow on loading tabs in offline mode once the number of pages to be opened grows larger than a dozen or so. We can think of fewer worse case scenarios than having 100 tabs set to reload in offline mode upon Firefox startup, but this is exactly what we have done here.

By having 100 pages open in Firefox portable, setting Firefox to reload the last session upon next session start and then setting it to offline mode, we are able to easily recreate a worst case scenario. Since we are using Firefox portable all files are easily positioned in one location, making it simple to repeat the test as necessary. In order to ensure repetition, before touching the Firefox portable files, we have backed them up into a .rar file and only extracted a copy of it to the test device.</i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/ff.jpg" border="0" alt="" />
</div>

Real World Data Transfers

<i>No matter how good a synthetic benchmark like IOMeter or PCMark is, it cannot really tell you how your hard drive will perform in “real world” situations. All of us here at Hardware Canucks strive to give you the best, most complete picture of a review item’s true capabilities and to this end we will be running timed data transfers to give you a general idea of how its performance relates to real life use. To help replicate worse case scenarios we will transfer a 10.00GB contiguous file and a folder containing 400 subfolders with a total 12,000 files varying in length from 200mb to 100kb (10.00 GB total).

Testing will include transfer to and transferring from the devices, using MS RichCopy and logging the performance of the drive. Here is what we found. </i>

<div align="center"><img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/copy_sm.jpg" border="0" alt="" />
<img src="http://images.hardwarecanucks.com/image/akg/Storage/HyperX_Savage/copy_lg.jpg" border="0" alt="" /></div>

The combination of less than optimal deep queue depth performance and lower than usual maximum sequential performance does result in mediocre results in both Firefox and our copy benchmark tests. By the same token these results are still decent and if this drive had been released early last year we would have easily classified it as an enthusiast grade drive.