Samsung’s Portable SSD X5 reviewed

We may have published a review of an external SSD just a few short days ago, but when it rains, it pour-tables. Today, Samsung is unveiling a brand-new external SSD. Feast your eyes on the pinnacle of lightning-fast compact storage: the Samsung Portable SSD X5.

There are no existing Samsung X drives, so why is a completely new product called the X5? Simply to bring the numbering in line with Samsung’s existing “T” line of portable SSDs, the latest of which earned our hearty endorsement when we reviewed it last year. The T5 already matched the performance of mainstream internal SATA drives, so the next summit for Samsung to crest was to build an external drive that rivaled internal PCIe storage. Of course, such a feat would be impossible under the 10-Gbps constraint imposed by USB 3.1 Gen 2, so Samsung has upped the ante by outfitting the X5 with a Thunderbolt 3 interface.

We couldn’t pass up testing the Portable SSD X5 1 TB ahead of its launch, but it presented a bit of an awkward situation. My test motherboard has more M.2 slots, SATA ports, and USB 3.1 Gen 2 capabilities than you can shake a stick at, but there’s not a single TB3 controller to be found on it. With little time to spare before the X5’s announcement, we had Samsung send the drive directly to boss man Jeff. With his inexhaustible supply of hardware, he was able to patch together a test setup identical to mine with the single substitution of a TB3-equipped motherboard. Once I’d sent him my test scripts and he’d put on his storage hat, he was easily able to benchmark the drive and relay the results to me.

At 4.7″ x 2.5″ x 0.8″ and 5.3 ounces, there’s more meat to the X5 than the other portables we’ve reviewed. Some of the heft can be attributed to the drive’s magnesium body. Samsung claims the X5 can withstand drops of up to two meters. Samsung also says the drive’s design is “inspired by a supercar,” and that might not all be hyperbole. Jeff reports that the drive’s sparkly gray finish and metal-clad exterior has an almost automotive-paint feel to it, reinforcing a perception of stiffness and durability.

Image: Samsung

We didn’t break this drive apart, but nonetheless we know exactly what’s inside. The NAND is Samsung’s 64-layer TLC V-NAND. The SSD controller is Samsung’s Phoenix. The X5, then, is simply a 970 EVO hiding in a metal shell. This is the first external drive we’ve reviewed with an NVMe-over-PCIe stick inside. But then again, it would make no sense to pair a lesser drive with Thunderbolt 3.

Speaking of which, the TB3 controller in the X5 is Intel’s Alpine Ridge DSL6340. Crucially, that means that this TB3 drive is only a TB3 drive. If Samsung had coughed up for Intel’s newer Titan Ridge family of Thunderbolt controllers, the drive could have been made to work in USB 3.1 fallback modes, but older controllers don’t have that capability enabled for sink devices. Samsung also cautions that Thunderbolt 2-equipped devices won’t work with this baby, either.

We repeat: don’t be fooled by the drive’s USB Type-C connector. If you don’t have bona fide Thunderbolt 3 ports to work with, the X5 becomes an expensive paperweight. Jeff (madman that he is) forged ahead and plugged it into a non-TB3 Type-C port just to see what happens, and the answer is a whole lotta nothing. The drive’s LED emits nary a flicker of acknowledgement.

A heatsink that spans the entire length of the drive undoubtedly contributes a lot to the drive’s weight. Hopefully it allows the drive to operate at full-bore for longer before Samsung’s Dynamic Thermal Guard (read: throttling) kicks in. Regardless of DTG, Jeff says the X5 emits substantial heat in operation—enough that he was careful not the leave anything atop the drive.

Fancy paint and Alpine Ridge don’t come cheap. Samsung set the X5 1 TB’s launch price at a whopping $700. That’s about two-and-a-half times the current street prices for top-tier drives like the T5 and SanDisk’s Extreme Portable. A 2-TB version will be available for $1400, and a 500-GB version will be available for $400. Unfortunately, the warranty doesn’t scale with the asking price. The X5 is backed by the same three-year warranty that Samsung extends to the T5. Samsung’s Portable SSD Software is also along for the ride, giving access to the X5’s AES 256-bit hardware encryption. It looks to be unchanged since we played with it for the T5.

Now let’s examine the fruits of Jeff’s quest to become a storage reviewer.


IOMeter — Sequential and random performance

IOMeter fuels much of our latest storage test suite, including our sequential and random I/O tests. These tests are run across the full capacity of the drive at two queue depths. The QD1 tests simulate a single thread, while the QD4 results emulate a more demanding desktop workload. For perspective, 87% of the requests in our old DriveBench 2.0 trace of real-world desktop activity have a queue depth of four or less. Clicking the buttons below the graphs switches between results charted at the different queue depths. Our sequential tests use a relatively large 128-KB block size.

The X5’s sequential read speeds are much faster than anything else we have to test with, but its write speeds are spectacularly unspectacular. This result clearly is not indicative of the drive’s true performance, but is likely rather some of sort of issue between the drive and IOMeter. Samsung goes as far as to say in the drive’s literature that “some performance benchmark tools may not work properly.” Perhaps Alpine Ridge is to blame. Let’s see if random response times are affected. 

The X5’s read and write response times are snappy as heck.  But sequential writes are still a mystery, so let’s turn to RoboBench to clear up the confusion.


TR RoboBench — Real-world transfers

RoboBench trades synthetic tests with random data for real-world transfers with a range of file types. Developed by our in-house coder, Bruno “morphine” Ferreira, this benchmark relies on the multi-threaded robocopy command build into Windows. We copy files to and from a wicked-fast RAM disk to measure read and write performance. We also cut the RAM disk out of the loop for a copy test that transfers the files to a different location on the SSD.

Robocopy uses eight threads by default, and we’ve also run it with a single thread. Our results are split between two file sets, whose vital statistics are detailed below. The compressibility percentage is based on the size of the file set after it’s been crunched by 7-Zip.

  Number of files Average file size Total size Compressibility
Media 459 21.4 MB 9.58 GB 0.8%
Work 84,652 48.0 KB 3.87 GB 59%

RoboBench’s write and copy tests run after the drives have been put into a simulated used state with 30 minutes of 4KB random writes. The pre-conditioning process is scripted, as is the rest of the test, ensuring that drives have the same amount of time to recover.

The media set is made up of large movie files, high-bitrate MP3s, and 18-megapixel RAW and JPG images. There are only a few hundred files in total, and the data set isn’t amenable to compression. The work set comprises loads of TR files, including documents, spreadsheets, and web-optimized images. It also includes a stack of programming-related files associated with our old Mozilla compiling test and the Visual Studio test we run in other SSD reviews. The average file size is measured in kilobytes rather than megabytes, and the files are mostly compressible.

Let’s take a look at the media set first. The buttons switch between read, write, and copy results.

The X5 brought a gun to a knife fight. It outpaces the rest of the field by such lengths that it’s almost comical. At eight threads, the drive’s read speeds are four times faster and its write speeds three times faster than any other drive in our test suite. The IOMeter problem clearly didn’t extend to RoboBench.

The work set, on the other hand, should be more difficult for the X5 to run right over.

Indeed, the margin of the X5’s victory shrinks a lot. But still, the Thunderbolt 3 drive’s performance is simply on another level.

Turns out putting an NVMe drive inside an external enclosure gives it an NVMe-like advantage over more pedestrian portables. To further explore the X5’s incredible speeds, Jeff performed some informal testing of his own devising.


Getting creative

In benchmarking the X5, we found ourselves wondering who might even need such a product. Everybody likes speed and portability, but this is a crazy fast and expensive piece of hardware. Here’s Samsung’s vision of the drive: “The X5 is designed for content creators and high-tech professionals who need fast and secure access to data in a lightweight and compact storage solution.”

I like to think of myself as a “high-tech professional,” but I just don’t need “fast and secure access to data” in a portable form factor as I go about my week. But it stands to reason that creative professionals might take work home with them, and they likely have TB3-equipped computers (such as MacBook Pros) at either end of their commutes.

Jeff decided to tackle the content creator side of the equation by testing transfer speeds of a truly enormous 114 GB Final Cut project bundle. This project contained all of the source clips, music, and miscellaneous files from our Ryzen Threadripper video-editing build earlier this year, so its contents include a bunch of relatively short 4K movie files and some lengthier 4K source material. It should represent a good mix of what the average video editor might want to move between PCs.

Jeff dumped the bundle on an Intel 750 Series 400-GB SSD connected to a Gigabyte X299 Designare EX motherboard and timed how long it took to move the bundle over to three different target disks using the board’s Thunderbolt 3 interface or USB 3.0 ports, as required. He used a single-threaded instance of Robocopy to perform the move, since that’s likely representative of real-world use.

The first target drive was an ancient Seagate Backup Plus 8 TB external hard drive with a USB 3.0 interface and a shingled-magnetic-recording drive inside. That disk represents the baseline of what most folks might use to move massive files around. The second drive we used was a Corsair Neutron XT 480 GB SSD with a SATA 6 Gbps interface. We connected this drive directly to a SATA 6 Gbps port on the motherboard, since USB 3.1 Gen 2 externals tend to perform about as well as directly-connected storage. The final target was, of course, the X5 1 TB.

The X5 is the clear winner here, hitting fairly close to double the write rate of the Neutron XT. The X5 took only 131 seconds to transfer over the 114 GB bundle, compared to 244 seconds for the Corsair. The Seagate external took an incredible 40 minutes. Using an SMR HDD for a heavy write load is a reminder of why we love solid-state storage so much.

With the cost of the Samsung drive being what it is, we also wondered whether some of Thunderbolt 3’s alternate modes might obviate the need for a Portable SSD X5 or similar products entirely. For example, one can directly connect two Thunderbolt 3-equipped computers and form an ad-hoc, high-speed network between them. We tried this approach with Jeff’s Alienware 13 R3 notebook and our X299 Designare EX motherboard and found that Windows recognized the network formed this way as a 20 Gbps link. Jeff’s Alienware 13 has a Samsung PM961 NVMe SSD in it, or roughly the equivalent of a 960 EVO, as well, so copying between the machines should have been limited by link speed alone.

Despite that impressive-sounding speed, however, our file transfer ultimately ended up a bit slower than even our transfer to the Corsair SATA SSD above. We’re not sure why this is, but it just goes to show that there’s more to Thunderbolt 3 performance than raw link bandwidth. The Portable SSD X5 definitely earns a place for itself when there’s a need to load up an external storage device quickly.

That’s it for testing. On the next page is a breakdown of our test methods. The conclusion follows on the subsequent page.


Test notes and methods

Here are the essential details for all the drives we tested:

  Interface Flash controller NAND
Adata SE730H External SSD 512GB USB 3.1 Gen 2 Silicon Motion SM2258 Micron 3D TLC
Intel X25-M G2 160GB SATA 3Gbps Intel PC29AS21BA0 34-nm Intel MLC
Samsung Portable SSD T5 1TB USB 3.1 Gen 2 Samsung MGX 64-layer Samsung TLC
Samsung Portable SSD X5 1TB Thunderbolt 3 Samsung Phoenix 64-layer Samsung TLC
SanDisk Extreme Portable SSD 1TB USB 3.1 Gen 2 Marvell 88SS1074 64-layer SanDisk TLC

The SATA SSDs were connected to the motherboard’s Z270 chipset. The portable SSDs were connected via the motherboard’s USB 3.1 Gen 2 Type-C port.

We used the following system for testing:

Processor Intel Core i7-6700K
Motherboard Gigabyte Aorus Z270X-Gaming 5
Firmware F10B
Memory size 16 GB (2 DIMMs)
Memory type Corsair Vengeance LPX DDR4 at 2133 MT/s
Memory timings 15-17-17-35
System drive Corsair Force LS 240GB with S8FM07.9 firmware
Power supply Rosewill Fortress 550 W
Operating system Windows 10 x64 1803

Thanks to Gigabyte for providing the system’s motherboard, to Intel for the CPU, to Corsair for the memory and system drive, and to Rosewill for the PSU. And thanks to the drive makers for supplying the rest of the SSDs.

We used the following versions of our test applications:

  • IOMeter 1.1.0 x64
  • TR RoboBench 0.2a

Some further notes on our test methods:

  • To ensure consistent and repeatable results, the SSDs were secure-erased before every component of our test suite. For the IOMeter database, RoboBench write, and RoboBench copy tests, the drives were put in a simulated used state that better exposes long-term performance characteristics. Those tests are all scripted, ensuring an even playing field that gives the drives the same amount of time to recover from the initial used state.

  • We run virtually all our tests three times and report the median of the results. Our sustained IOMeter test is run a second time to verify the results of the first test and additional times only if necessary. The sustained test runs for 30 minutes continuously, so it already samples performance over a long period.

  • Steps have been taken to ensure the CPU’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 frequency at 4.0 GHz. Transitioning between power states can affect the performance of storage benchmarks, especially when dealing with short burst transfers.

The test systems’ Windows desktop was set at 1920×1200 at 60 Hz. 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.



The X5 blew all of our other contenders clean out of the water. There is no doubt that it will land far ahead of the other portables in our rankings.  We distill an overall performance rating using an older SATA SSD as a baseline. To compare each drive, we then take the geometric mean of a basket of results from our test suite.

The difference between the X5 and mere USB drives is about as stark as that between NVMe sticks and ordinary SATA drives. That makes perfect sense, since the X5 essentially has a 970 EVO pushing bits around under the hood. If your only concern is pure speed, look no further. But since most of us don’t have unlimited budgets to spend on hardware, we must consider the price-to-performance proposition. In the graph below, the most compelling position is toward the upper left corner, where the price per gigabyte is low and performance is high.

Samsung’s launch price of $700 for the Portable SSD X5 1 TB comes out to 70 cents per gigabyte. That’s a tough pill to swallow when the 970 EVO 1 TB is available for $350 at Newegg or $340 at Amazon. You’re basically paying double the price of the bare drive for Alpine Ridge and a heatsink. If we only consider portable drives, however, Samsung may be justified in thinking it can charge this much. Samsung’s Portable SSD T5 and SanDisk’s Extreme Portable may be a whole lot cheaper, but our testing definitively proves that they can’t hold a candle to the X5’s speeds. If your time is money, then the Portable SSD X5 won’t leave you tapping your foot when you need to move big files around.

We have a couple big problems with the Portable SSD X5 despite its blistering performance. First, there’s the issue with its strict requirement for Thunderbolt 3. There are a lot of machines out in the wild that don’t have TB3 ports. It’s entirely reasonable for a prospective buyer to want to use the X5 to transfer data from TB3-equipped computers to older machines, but the X5 can’t talk to them by design. Heck, even video cameras are recording to external hard drives these days over USB Type-C cables, but the X5 can’t accept that input. Had Samsung opted to let the X5 operate in a USB 3.1 Gen 2 fallback mode, it could have been the perfect high-performance portable. Alas.

Second, the drive is almost too fast for its own good. It’s all well and good that the X5 can operate at NVMe-like speeds, but those speeds could be bottlenecked by host machines’ internal storage a lot of the time. You can’t write to the X5 faster than you can read from your PC’s data drives and vice versa. Therefore, the X5 can only be used to its fullest effect by users who are toting data in between multiple machines with TB3 ports and high-speed NVMe drives inside.

At the end of the day, the X5 is a tiny technological marvel. Its speeds are nothing short of ludicrous. The mission of external portable storage is to allow for quick and painless file transfers, and the X5 is absurdly good at that. We do have concerns about its price and practicality, but there’s still an audience for a drive like this.

If all the computers you work with are modern machines with Thunderbolt 3 and NVMe internal storage, if your work requires you to spend a lot of time transferring massive files between those computers, and if the minutes you save not waiting for transfers to complete are worth its high price, then the X5 is the drive for you. It’s an excellent portable SSD that we don’t hesitate to call TR Recommended. But for everybody else, slower and more broadly compatible storage like the Portable SSD T5 will do the job for far less money.

Comments closed
    • Waco
    • 4 years ago

    Complaining about this drive throttling in an abusive use case is like complaining that shingled drives are bad at random 4K writes. For [i<]what it was designed for[/i<] it works very well. That said...$.70 per GB? No thanks!

    • smilingcrow
    • 4 years ago

    I agree that the Anandtech test is not real world but at the point at which the performance dropped off was just beyond where your test ended it left an answered question:
    How does the drive throttle in real world usage using very long write situations.
    So I have looked around for a long real world test and it seems fine:
    [url<][/url<] I figured the Anandtech drive was faulty or something was up as it shouldn't throttle that badly. Thanks for pointing out how pointless their test is without qualifying it. Plain silly. Still curious how the other TB3 drive could be 10 times faster than this Samsung under extreme conditions. For the tiny niche that want to push one of these much harder than normal that is worth investigating as at what usage level does this drive fall to pieces as in the Anandtech.

    • Jeff Kampman
    • 4 years ago

    We can continue to dig into this edge case, but I will confidently state that there is nothing wrong with the drive for the types of workloads it will be subjected to from the types of applications it was designed for.

    Client workloads simply do not present SSDs with the level of parallelism that the test we’re discussing displayed for the lengths of time necessary to induce that kind of bottomed-out throttling, and that’s the long and short of it.

    • derFunkenstein
    • 4 years ago

    More importantly, who is going to use external storage for QD32 workloads? Portable SQL servers?

    • smilingcrow
    • 4 years ago

    Thanks, I missed that one.
    The Anandtech review shows a similar data rate as yours for roughly the time taken for your test but soon after the performance drops off dramatically.
    But with only 2 data points it’s not possible to say exactly what is going on here.
    But given that Anandtech have run their test on another TB3 SSD and it was 10x faster for longer runs it seems fair to assume that something is not right with the Samsung.
    It would be interesting to see how your real-world test performs with a larger data set as even if only 2 or 3 times larger it would push it into the area where it nosedived on the Anandtech review.
    I guess I will need to read a 3rd review to find out.

    • Jeff Kampman
    • 4 years ago

    I’m referring to our transfer of a 114-GB Final Cut Pro project bundle here: [url<][/url<] Even that transfer takes just about two minutes on this drive. Unless you're filling up the X5 from an NVMe drive of similar size with a single enormous file (implying that your expensive NVMe SSD isn't a system drive to begin with, which would be odd), you will likely a) find that your job completes before the X5 verges into thermal-throttling territory or b) that you are limited by interface transfer rates from the host system to begin with for drives where capacities exceed 1 TB. As I said before, I think people panicking about the thermal characteristics of this drive need to do a gut check on the length of time it will actually take them to move a given file to it and/or recall the nature of client workloads. 32 FIO workers clobbering the drive with continuous sequential writes that never complete or allow for any idle time is a synthetic workload, not a real-world test. It is not worth fixating on when client systems rarely put storage work on more than one thread. You also can't buy the TEKQ Rapide that other sites have tested with any longer so it's an academic point of comparison, not a practical one.

    • smilingcrow
    • 4 years ago

    30 minutes does seem excessive to me also but how long exactly are TR running the RoboBench tests?

    • Usacomp2k3
    • 4 years ago

    At that speed, you’re not going to likely have any application will run run flat-out for 30 minutes. That’s just not a real-world use case. The faster drives finish the tests faster. Who knew!

    • smilingcrow
    • 4 years ago

    Likewise for the RoboBench tests, how long are they run for in terms of repetitions or time?

    • Lianna
    • 4 years ago

    I read “Notes from TR’s next-gen storage testing”, but could not find details about IOMeter testing, especially as it relates to external drives. I guess it’s not “30-minute test” from “Notes…”.

    Could the author please tell us how long is the IOMeter testing for sequential128k/random(4k?) read/write QD1/QD4 results presented in this review (in minutes or GB transferred)?

    • Ultracer
    • 4 years ago

    But why such a horrendous design of the casing???

    • smilingcrow
    • 4 years ago

    I looked at your RoboBench write tests and the large files working set is 9.58GB so is that test which takes 6 to 13 seconds to complete run for a sustained 30 minutes?
    I looked at the notes and it wasn’t immediately clear which tests are classed as sustained so run for 30 minutes!

    “On the other hand, our DIY configuration with the SanDisk Extreme Pro 1TB SSD in the TEKQ Rapide aluminum Thunderbolt 3 enclosure managed to pass the same test without any issues in less than 1/10th of the time taken by the X5 for the same amount of data.”

    From here:
    [url<][/url<] So clearly the test isn't so demanding that other drives albeit a DIY set-up can't trounce it whilst costing over 20% less.

    • Jeff Kampman
    • 4 years ago

    Hammering a drive with continuous sequential writes via 32 FIO workers is like using Prime95 and Unigine Heaven to evaluate notebook thermals. It’s a worst-case scenario, not a typical workload. I would strongly urge you to consider the large file transfer test we performed for a better example of this product’s real-world performance.

    • smilingcrow
    • 4 years ago

    According to the Anandtech review this throttles like crazy after sustained writes down to 60MBs.

    • hubick
    • 4 years ago

    WTF Samsung, Alpine Ridge?

    This is a premium device, and these things are stupid expensive, and Titan Ridge is already out there!

    I’ve been waiting for a TB3 upgrade to my 1TB Samsung T1 drive, which, among other things, I use to backup photo/video from my action cam and phone while travelling. It would sure be nice to use the TB3 on my PC (Hades Canyon NUC) to retrieve all that when back home, but without USB-C compat, it’s basically impossible to get anything onto the drive out in the field, making this offering useless 🙁

    • derFunkenstein
    • 4 years ago

    Well, I’m asking about this thing specifically. Not that I want to drop $700 on a 1TB SSD just for the fun of it, but it’d give it more utility. Windows does not natively install to a USB target, and you have to do [url=<]some finagling[/url<]. Not that it's painfully difficult, but having something that could be "blessed" is nice.

    • Bauxite
    • 4 years ago

    I have some T5s and they are great, but they [b<]really[/b<] should have waited for the usb 3 fallback option. More speed is great but just being able to get to the data is #1 with a portable. I'd also argue few need real high speeds on a portable and the T5s are already pretty damn fast (400MB/s bulk transfers) so losing general access is too much.

    • DancinJack
    • 4 years ago

    Depends on the portable device’s hardware. You can boot via TB3/USB but it depends on the external device you’re using.

    • derFunkenstein
    • 4 years ago

    [quote<]We may have published a review of an external SSD just a few short days ago, but when it rains, it pour-tables.[/quote<] Thank you for that. Stupid question time: can you install Windows 10 to this thing and boot via TB3? Then you'd have a portable boot drive. That would have limited utility, I'm sure, but inquiring minds want to know!

    • FireGryphon
    • 4 years ago

    Nice review. I like the portable sdd reviews. I’m curious how the transfer on the X5 was slower overall even with such fast transfers. That seems like a pretty big deal. Also, the lack of connectivity and in fact the very limited connectivity of this item should have kept it from being recommended. I mean, not being able to use it on such a large number of systems is a big failing.

Pin It on Pinterest

Share This

Share this post with your friends!