Feedback

Samsung EVO Plus (2017 Model) 128GB UHS-I U3 microSD Card Quick Look

Subject: Storage
Manufacturer: Samsung

Today Samsung released an update to their EVO+ microSD card line. The new model is the 'EVO Plus'. Yes, I know, it's confusing to me as well, especially when trying to research the new vs. old iterations for this mini-review. Here's a few quick visual comparisons between both models:

View Full Size

On the left, we have the 'older' version of the Plus (I mean the '+'), while on the right we have the new plus, designated as a '2017 model' on the Samsung site. Note the rating differences between the two. The '+' on the left is rated at UHS-I U1 (10 MB/s minimum write speed), while the newer 'Plus' version is rated at UHS-I U3 (30 MB/s minimum write speed). I also ran across what looked like the older version packaging.

View Full Size

The packaging on the right is what we had in hand for this review. The image on the left was found at the Samsung website, and confuses things even further, as the 'Plus' on the package does not match the markings on the card itself ('+'). It looks as if Samsung may have silently updated the specs of the 256GB '+' model at some point in the recent past, as that model claims significantly faster write speeds (90 MB/s) than the older/other '+' models previously claimed (~20 MB/s). With that confusion out of the way, let's dig into the specs of this newest EVO Plus:

View Full Size

For clarification on the Speed Class and Grade, I direct you to our previous article covering those aspects in detail. For here I'll briefly state that the interface can handle 104 MB/s while the media itself is required to sustain a minimum of 30 MB/s of typical streaming recorded content. The specs go on to claim 100MB/s reads and 90 MB/s writes (60 MB/s for the 64GB model). Doing some quick checks, here's what I saw with some simple file copies to and from a 128GB EVO Plus:

View Full Size

Our figures didn't exceed the specified performance, but they came close, which more than satisfies their 'up to' claim, with over 80 MB/s writes and 93 MB/s reads. I was able to separately confirm 85-89 MB/s writes and 99 MB/s reads with Iometer accessing with 128KB sequential transfers.

Pricing (MSRP)

Pricing seems to be running a bit high on these, with pricing running close to double of the previous version of this very same part (the EVO+ 128GB can be found for $50 at the time of this writing). Sure you are getting a U3 rated card with over four times the achievable write speed, but the reads are very similar, and if your camera only requires U1 speeds, the price premium does not seem to be worthwhile. It is also worth noting that even faster UHS-II spec cards that transfer at 150 MB/s can be had and even come with a reader at a lower cost.

In summary, the Samsung EVO Plus microSD cards look to be decent performers, but the pricing needs to come down some to be truly competitive in this space. I'd also like to see the product labeling and marketing a bit more clear between the '+' and the 'Plus' models, as they can easily confuse those not so familiar with SD card classes and grades. It also makes searching for them rather difficult, as most search engines parse 'Plus' interchangeably with '+', adding to the potential confusion.

View Full Size

*edit* a reader requested a CrystalDiskMark run. Here it is:

View Full Size


March 28, 2017 | 04:19 PM - Posted by PCPerFan (not verified)

Lexar makes even faster ones than you listed. These 1800x ones have over 250MB/s reads: https://www.amazon.com/Lexar-Professional-microSDXC-UHS-II-Reader/dp/B01...

March 28, 2017 | 04:34 PM - Posted by Allyn Malventano

Agreed but those are even more expensive than the EVO Plus, which I already find pricier than they should be compared to the 1000x (UHS-I) Lexar offerings I mentioned near the end.

March 28, 2017 | 06:56 PM - Posted by sircod (not verified)

No application-class certification? Does it not have the IOPS to meet that?

March 28, 2017 | 11:54 PM - Posted by jayden2002 (not verified)

What is the reader? I find it important to know that on memory cards testing.

March 29, 2017 | 05:25 PM - Posted by Allyn Malventano

I tried several UHS capable readers (from Kingston, Lexar, etc). They all return nearly identical results. I even got results within 2 MB/s with the reader in an Anker USB hub that was connected to the USB hub of an Acer LCD, and then finally connected to the PC.

March 29, 2017 | 02:29 AM - Posted by Edwin Hoeks (not verified)

It might seem like a strange request, but would it be possible to add the 'High Speed' (non UHS) read and write performance? Or do these cards simply 99% saturate the available bandwidth when using the spec 2.00 bus. As an embedded user of Secure Digital a nice metric would be the 4kB random write IOPS @ 'High Speed'

March 29, 2017 | 05:28 PM - Posted by Allyn Malventano

Yeah, the cards that can go faster will generally just peg the bus when connected to a slower link.

March 29, 2017 | 10:10 AM - Posted by Jann5s

Allyn, do you have any comments on latencies. I know they are less important for portable storage, but opening the gallery app on my phone can be a pain on some SD cards, regardless of their advertised throughput.

Perhaps also interesting for those using SDcards for their Raspberry Pis or similar.

March 29, 2017 | 05:38 PM - Posted by Allyn Malventano

I've found that a lot of the latency you find with SD comes from overhead from the protocol / interface. Cards that can do 100+ MB/s can only accomplish that if their own latency is relatively low. It's just not a nimble protocol and that makes things worse for cards that can't brute force their way through with sheer throughput.

March 29, 2017 | 11:35 AM - Posted by JohnGR

I would love to see CrystalDiskMark numbers if not in this case, at least in future quick looks. Or any other program that can measure 4K performance. I have stopped buying SD cards that suck at 4K, long ago.

March 29, 2017 | 05:20 PM - Posted by Allyn Malventano

The problem with testing SD cards is that Windows will buffer the USB link comms in most cases, so all you get is buffered results that don't match reality. The other complication with random writes on SD cards is that performance varies wildly based on the amount of random / conditioning you hit the card with. Either way, if you're using these for bulk storage applications, the card will never see 4k. Even small images these days are pushing 1MB or higher. 4k is really more of a desktop file system worst-case granularity thing, or more of an enterprise workload.

All of that considered, I do want to do random write testing of these, but want to do it correctly, and there just wasn't time for this particular piece. I did a quick CDM run for you and added it to the piece.

March 30, 2017 | 03:31 AM - Posted by JohnGR

microSD cards do see some limited at least desktop usage today, with all those smartphones, tablets and netbooks with limited internal storage. If someone adds a microSD card on a netbook and chooses to install Chrome for example on the microSD, is Chrome going to use the C partition for caching or the microSD card where it was installed? If it uses the microSD, then I guess 4K speeds are important.

Now I do trust that the numbers from CrystalDiskMark can show what a microSD card can do, but I am no expert in storage testing.

In any case thank you very much for adding those results. The 4K scores look the same with an EVO, non Plus, card I own. The sequential speeds on the other hand are more than double for the Plus.

April 4, 2017 | 09:07 AM - Posted by Photonboy (not verified)

I would assume the SD card for any device would simply be assigned its own drive letter, thus would act in that capacity like adding a secondary HDD.

If you install Chrome to the SD card (or any other program) then where it buffers data depends on how the program was written. Many programs have hidden folders, or even viewable ones that by default stay on the OS partition.

Some programs also allow you to point save folders, or working cache folders to your own folder if need be.

(because you want to avoid constant writing to flash memory... that would be more important than 4K performance)

How to TRUST benchmark results?

It depends on the program, but as said 4K (four kilobyte data size) isn't that common, but again you want to avoid a situation where you constantly write to flash so investigate carefully how that works.

*Write endurance also varies between flash drives, though obviously there's a cost difference. It might be worth just EXPERIMENTING with a cheap-ass card and see how that works for you. Aside from possibly doing most write on the main OS partition, programs also tend to load into the main system memory (on desktop that is DDR3/DDR4).

BTW, I ran Linux on an 8GB USB thumb drive for a year with about 100 hours total usage. It obviously is wearing out but still worked fine.

Lastly, I would investigate if any of these modern flash devices have any WEAR LEVELLING hardware.

March 30, 2017 | 02:41 PM - Posted by ehoeks

I'm wondering if a 'High Speed' (non UHS) random 4 kB very controlled (pseudo) random (read-/)write- Test would be useful. I'm building a codebase for Atmel E70/V70/V71 controller SD-Card High Speed tightly controlled raw video (low resolution) playback for outdoor LED signage.

We could design a simple benchmark that we can control tightly. And it can run on a $40 evaluation board like the ATMEL ATSAME70-XPLD Evaluation Board. (like: http://www.newark.com/atmel/atsame70-xpld/evaluation-board-atsame70-32bi... )

Please let me know if you would be interested in helping to specify and test a benchmark like that.

April 4, 2017 | 09:11 AM - Posted by Photonboy (not verified)

I don't think flash memory is suitable for video usage, or is this just to hold an OS (Linux?) and stream the signal to another device with an HDD?

So if you want to be constantly writing I don't think flash is a good idea.

Conversely, if it's just the OS, and application for such a simple task do you need a high-speed flash drive?

March 30, 2017 | 12:24 PM - Posted by Chris C (not verified)

I can see why they kept the same name, confusion is good for marketing.
I own the older variant of the evo+ and only purchased it last week which makes me annoyed considering the timing of this release, the pre 2017 version has much worse write speeds. If samsung gave it a new name then word would spread and sales of the older evo+ cards would tank. Now people may read these reviews, think wow, buy a evo+ and may not even be aware its the older version, either way the older evo+ cards will keep their sales.
The reason the older evo+ 256 gig card was much faster than the older 128gig card I believe is due to using 3d nand, which I expect is what these new 2017 cards all use as well.

March 30, 2017 | 02:20 PM - Posted by Anonymous (not verified)

Are you sure a 256GB EVO+ U1 card ever existed? Ever since the 256GB EVO+ was released last summer I've only ever seen the version with speeds like the pro-line (B&W card) which didn't have a 256GB option at all. The card should have been in the pro-line from the beginning because of the similar performance. Not in the evo+ line with its slower siblings.

April 2, 2017 | 02:58 PM - Posted by Anonymous (not verified)

Does 10000 mating cycles actually mean the same thing as write cycles?

That would suggest these are as robust as MLC and useable in applications like dashcams and video surveillance at a much lower cost than competing endurance flash cards that use MLC and SLC...

April 4, 2017 | 09:22 AM - Posted by Photonboy (not verified)

It appears that it's the number of writes per cell.

TLC is as low as 3,000 per cell so ONLY looking at this metric suggests its better.

However there's more to it such as how wear leveling works. Do you have a lot of SPACE left over?

Frankly, even if wear leveling didn't even exist I'm not sure how easy it would be to wear this thing out because a TLC SSD in a desktop computer that gets used everyday can easily last over 10 years (depending on free space etc).

April 4, 2017 | 09:25 AM - Posted by Photonboy (not verified)

Also, my old 8GB USB drive is rated for 500 writes. Seems low of course so I'm surprised I ran Linux.

Finally, don't forget if using as a partition that some flash media may or may not work with the RECYCLE BIN (so deleted files may disappear without easy recovery).

April 12, 2017 | 02:28 AM - Posted by dafubo

I have been using this 128GB one for 5 months. The speed of data transferring is really good. 128 GB is really cool because I can install N number of applications. Before that, I could not install apps that I want due to insufficient space in my device. Games like Fifa and NBA from tutu helper takes a lot of space. This is the best.