Any "Fit Size" USB Flash drives with static wear leveling?

[cbn]

Here are some that I am looking at:

https://www.samsung.com/us/computin...-0-flash-drive-fit-32gb-muf-32bb-muf-32bb-am/

https://www.sandisk.com/home/usb-flash/ultra-fit-usb

http://www.pny.com/elite-x-fit-usb-3-0-flash-drive

EDIT: http://www.lexar.com/products/usb-f...S45-USB-3-flash-drive.html#SKU=LJDS45-32GABNL

P.S. I'm looking for 32GB size (and usb 3.0).

 

[cbn]

I suppose an alternative to a 32GB Fit size USB with static wear leveling might be:

1. A 64GB Fit size USB Flash drive with dynamic wear leveling.

2. A 32GB Larger size USB Flash drive with static wear leveling. (I know the Sandisk Extreme usb 3.0 fits this category, but it is pretty large. A lot larger than I want.)

 

[cbn]

http://www.phison.com/English/ICSpeed.asp?SortID=60

^^^^ All these Phison USB 3.0 flash controllers support static (ie, Global) wear leveling.

But are any of these small enough to be used in the "Fit" form factor? (Also do all implementations of usb 3.0 flash drives with these controllers have the static wear leveling enabled?).

P.S. Here are the Silicon Motion USB Flash controllers. I didn't see wear leveling mentioned when I clicked on each controller's link. (It could be the controller still does have it....and that the website is simply not comprehensive enough. Need more info. EDIT: "Advanced wear-leveling to extend product lifecycle" mentioned on this page and this page.)

I couldn't find any USB Flash controllers for Marvell.

SIDE NOTE: Only controller from Phison and Silicon Motion Compatible with 3D NAND is Phison's PS2251-09. All current USB Flash controllers (from both Phison and Silicon Motion) are using BCH ECC.

 

[cbn]

Just got off Live chat (or phone call) with Samsung, Sandisk, PNY and Lexar regarding their FIT drives (listed in opening post).

The Samsung agent did write their drive has has wear leveling, but he couldn't confirm whether it is Static or Dynamic.

The Sandisk agent did not know if the drive had wear leveling.

The PNY agent did not know if the drive had wear leveling.

The Lexar agent said the drive has wear leveling, but he couldn't confirm whether it is Static or Dynamic.

Any software available that can help me with this determination?

 

[Glaring_Mistake]

Just got off Live chat (or phone call) with Samsung, Sandisk, PNY and Lexar regarding their FIT drives (listed in opening post).

The Samsung agent did write their drive has has wear leveling, but he couldn't confirm whether it is Static or Dynamic.

The Sandisk agent did not know if the drive had wear leveling.

The PNY agent did not know if the drive had wear leveling.

The Lexar agent said the drive has wear leveling, but he couldn't confirm whether it is Static or Dynamic.

Any software available that can help me with this determination?

Well, there is software that can help you determine what controller they use at least.
Like for example Flash Drive Information Extractor and ChipGenius: http://www.softpedia.com/get/System/System-Info/Flash-Drive-Information-Extractor.shtml
http://www.softpedia.com/get/System/System-Info/ChipGenius.shtml
But you'd still need to figure out if the controller uses Static or Dynamic wear levelling (or both).

Now the Samsung and the Lexar drive look to use the same controller (SM3267).
There is no mention of the type of wear levelling used for Silicon Motion's USB controllers on their page other than that they all use: "Advanced wear-leveling to extend product lifecycle".

The controller in the SanDisk Ultra Fit cannot be detected by either Flash Drive Information Extractor and ChipGenius so no idea what controller they use.

Seeing as I don't own the PNY drive I'm unable to check what controller it uses.


Now some general information on the drives:

The SanDisk Ultra Fit is not one I am inclined to recommend.
The reason for that being both because write speeds can be a bit erratic and even worse - that it has a tendency to disconnect from the computer, particularly if you're writing more than a few GB to it.

If I remember correctly the Lexar S45 is not prone to disconnections (or at least not as much as the SanDisk Ultra Fit) and performance is more stable.
Unlike the SanDisk Ultra Fit and the Samsung FIT that have a small light pulse during any disk activity however it has a pretty obnoxious light flash rapidly which hasn't exactly endeared it to me.

Samsung FIT tends to perform about as good as the Lexar S45 or possibly a bit better even and does not disconnect even under sustained workloads.
Nor does it flash constantly like the S45 so it is certainly my favorite of the three.

As stated earlier I can't comment on the PNY drive however.

 

[cbn]

There is no mention of the type of wear levelling used for Silicon Motion's USB controllers on their page other than that they all use: "Advanced wear-leveling to extend product lifecycle".

Thanks. I fixed post #3.

 

[cbn]

Here is a review I found of the Samsung FIT USB 3.0 32GB.

And here is the CrystalDiskMark:

http://www.cdrlabs.com/reviews/samsung-fit-usb-flash-drive/performance-and-final-thoughts.html

^^^^ That is a pretty good score compared to the 32GB eMMC I reported in the following post:

https://forums.anandtech.com/thread...-128gb-ssd-drops.2467389/page-3#post-38295302

 

[Glaring_Mistake]

Here is a review I found of the Samsung FIT USB 3.0 32GB.

And here is the CrystalDiskMark:

http://www.cdrlabs.com/reviews/samsung-fit-usb-flash-drive/performance-and-final-thoughts.html

I'm a bit skeptical actually.
You see, there's this issue I've observed with CDM when testing SSDs where write speeds (at least with sequential write speeds which is one of the easier to compare) can be much higher than they should be.
Noticed that sometimes even when I tested with 16-32GiB that write speeds were not much different from when I tested with 1GiB despite that 16-32GiB should have been more than sufficent to overwhelm the SLC-cache and cause write speeds to drop.
Like, for example with one drive that I tested with both AS-SSD and CDM it had a sequential write speed of 357MB/s with 1GiB and 354MB/s with 16GiB when using CDM.
Testing it with AS-SSD however it had a write speed of 340MB/s at 1GB and slightly below 70MB/s at 10GB.
4K writes differed a bit too for that drive (though that is not always the case): Around 100MB/s with CDM and 60MB/s with AS-SSD (16GiB and 10GB).

Also I tested my Samsung FIT (though it's a different capacity and has seen some use already so it's not a perfect comparison).
Anyway, it did not score as well as the one tested by CDRlabs but was not too far behind.
Still a bit too high though: Sequential write speeds were at 68MB/s while in actual use 35-40MB/s is pretty much the max (except for like at the very beginning) and testing with just writing a few files to it resulted in like 20-25MB/s speeds.
AS-SSD disagrees about the 4K random writes too: At best it has it pegged at one third of what CDM reports (1.5 for CDM and 0.3-0.5 with AS-SSD).

So in this case I think that CDM may have been a bit generous.

In conclusion, I think it's decent for what it is (considering the options) but doubt it is quite as fast as these results indicate.

 

[cbn]

I'm a bit skeptical actually.
You see, there's this issue I've observed with CDM when testing SSDs where write speeds (at least with sequential write speeds which is one of the easier to compare) can be much higher than they should be.
Noticed that sometimes even when I tested with 16-32GiB that write speeds were not much different from when I tested with 1GiB despite that 16-32GiB should have been more than sufficent to overwhelm the SLC-cache and cause write speeds to drop.
Like, for example with one drive that I tested with both AS-SSD and CDM it had a sequential write speed of 357MB/s with 1GiB and 354MB/s with 16GiB when using CDM.
Testing it with AS-SSD however it had a write speed of 340MB/s at 1GB and slightly below 70MB/s at 10GB.
4K writes differed a bit too for that drive (though that is not always the case): Around 100MB/s with CDM and 60MB/s with AS-SSD (16GiB and 10GB).

Also I tested my Samsung FIT (though it's a different capacity and has seen some use already so it's not a perfect comparison).
Anyway, it did not score as well as the one tested by CDRlabs but was not too far behind.
Still a bit too high though: Sequential write speeds were at 68MB/s while in actual use 35-40MB/s is pretty much the max (except for like at the very beginning) and testing with just writing a few files to it resulted in like 20-25MB/s speeds.
AS-SSD disagrees about the 4K random writes too: At best it has it pegged at one third of what CDM reports (1.5 for CDM and 0.3-0.5 with AS-SSD).

So in this case I think that CDM may have been a bit generous.

In conclusion, I think it's decent for what it is (considering the options) but doubt it is quite as fast as these results indicate.

Here are some other results I found on the Samsung FIT USB 3.0 32GB drive (and yes these do show much lower Sequential writes, one even shows a much lower 4K write):

https://techicize.com/samsung-usb-3-0-fit-review/#prettyPhoto

http://www.thessdreview.com/featured/samsungs-newest-flash-drives-reviewed/2/

 

[Glaring_Mistake]

Here are some other results I found on the Samsung FIT USB 3.0 32GB drive (and yes these do show much lower Sequential writes, one even shows a much lower 4K write):

https://techicize.com/samsung-usb-3-0-fit-review/#prettyPhoto

http://www.thessdreview.com/featured/samsungs-newest-flash-drives-reviewed/2/

Now those numbers are a lot more believable.
Think it may be that it became an issue with CDM 5.0 since I have no recollection of it having such an issue previously.
Can't say for sure however.

Something that I might mention about Samsung FIT is that while they've mostly had a low but fairly stable sequential write for some time a few of them have had their average write speeds drop a bit and become a bit more erratic or "jerkier".
Don't know what is behind it however since drives with a lot of free space may suffer from such issues while those with little free space may be free of them.

 

[cbn]

Review of the PNY Elite-X Fit 32GB (Controller is Phison PS2251-07):

From the information received by utility ChipGenius it is possible to note the following:
• Manufacturer of Phison controller
• PS2251-07 controller model (PS2307)
• Firmware F / W 08.03.5D from 2017-05-05
• TLC memory type, manufactured by Toshiba

And here are the Benchmark Scores:

CrystalDiskMark 5.2.2

 

[cbn]

I am assuming the PNY Elite-X Fit has static (ie, Global) Wear leveling because it has the Phison Controller

Too bad the 4K write is so low.

 

[cbn]

Here is a new fit size usb ( Sandisk Ultra Fit usb 3.1):

https://www.sandisk.com/home/usb-flash/ultra-fit

 

[Glaring_Mistake]

Here is a new fit size usb ( Sandisk Ultra Fit usb 3.1):

https://www.sandisk.com/home/usb-flash/ultra-fit

Haven't tested it myself but according to reviews on Amazon it doesn't overheat like the earlier versions used to which makes it significantly more useful.

 

[cbn]

Haven't tested it myself but according to reviews on Amazon it doesn't overheat like the earlier versions used to which makes it significantly more useful.

Thanks for that info.

I tried looking for the controller used but I haven't been able to find that yet.

I do wonder though if it could be Phison PS2251-09 (if it does use Phison) because the 256GB max capacity seems high to me for planar NAND and PS2251-09 is the only Phison USB controller to support 3D NAND.

 

[Glaring_Mistake]

Thanks for that info.

I tried looking for the controller used but I haven't been able to find that yet.

I do wonder though if it could be Phison PS2251-09 (if it does use Phison) because the 256GB max capacity seems high to me for planar NAND and PS2251-09 is the only Phison USB controller to support 3D NAND.

Depends, I think SanDisk tend to use their own controllers and they don't have a nice list of them available like Phison does that allows you to read their specifications.
And controllers do not necessarily need to support a type of NAND for them to actually be used with that NAND - case in point being SanDisk Ultra which was released despite the controller lacking support for TLC NAND (but with a lot of OP and with a function called Virtual Parity Recovery).

It is possible however that since this drive is nowhere near their top end, that they thought they could use an existing third party controller that already has confirmed support for 3D NAND rather than develop their own.
Both Lexar and Samsung have used controllers from Silicon Motion in some of their drives after all, and that's despite them using good (?) old 2D NAND in those as far as I know.

Anyway, I don't think you'll find out what controller they're using without plugging it in to see if some software can identify the controller used.
My guess? The controller is identified as "SanDisk".

 

[cbn]

Anyway, I don't think you'll find out what controller they're using without plugging it in to see if some software can identify the controller used.
My guess? The controller is identified as "SanDisk".

My Sandisk Ultra Fit USB 3.1 32GB finally came and the controller does indeed say "Sandisk" when I use that tool you provided in post #5.

Here are my CrystalDiskMark results (using CDM 3.0, CDM 5.5, CDM 6.0):

(This done via the USB 3.0 port of a HP Z420 Workstation)

 

[Glaring_Mistake]

Results for my Sandisk Ultra Fit USB 3.1 128GB:

A bit better than the 32GB it seems.

 

[Glaring_Mistake]

An update on the Sandisk Ultra Fit USB 3.1 128GB after some real life use: Sequential writes are fairly good at around 60MB/s... for like 1GB and then they go down and become pretty volatile with an (estimated) average of 15-20MB/s.
Still, that's pretty much what you would expect.