• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Here's the skinny on Samsung USB Flash Drives WRITE SPEEDS

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

c627627

c(n*199780) Senior Member
Joined
Feb 18, 2002
So here's the skinny on USB Flash Drives.
SanDisk overheats and cannot love you for a long time, whereas Samsung does not overheat when making gigantic tens of GB transfers.

Possibly/perhaps newer SanDisk models are different but been burned too many times by SanDisk to even think of giving them another chance.

Read speeds do not mean jack **** - many USB 3 flash drives read fast enough...
ONLY write speeds matter and truly distinguish USB Flash Drives, but guess what, no one "ever" posts real life write speeds anywhere, they are extremely hard to find.

Here they are as tested personally by me:
Real Life Write Speeds when transferring large amounts of data:

24MB/s Samsung BAR 32GB
29MB/s Samsung DUO 128GB
37MB/s Samsung DUO 64GB
40MB/s Samsung DUO PLUS 128GB
60MB/s Samsung BAR PLUS 128GB
96MB/s Samsung DUO PLUS 256GB

We are told that 128GB and 256GB are in a different [faster] speed classes than their same model counterparts in 32GB and 64GB varieties, across the board. Although luck on which batch you get plays a far bigger role, as you can see.
I would still buy at least a 128GB model, but you may get lucky when buying a 256GB model, to get unbelievable 100MB/s WRITE speeds.
 

trents

Senior Member
Joined
Dec 27, 2008
Same could be said for Samsung SSDs. They generally have faster write speeds than their competitors.

I think it comes down to use scenario, however. If I'm going to be using a flash drive frequently for read and write data transfers then the write speed will matter more than if I just use one to store a utility or other data that stays the same or that I only copy over to other storage devices. In that latter case, any old cheap USB will do.
 

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
Thanks for the info.

I wanted to add a couple of things for those who may swing by. :)

Transfer rates vary by file size. For example it will take a lot longer to transfer 32GB worth of 4k average file sizes than 4MB. If you run a benchmark called ATTO you can see this in action (or try it real world). This may be part of the reason you don't see many reviews cover it. You'll have MFG specs, and the specs from each review (they can be different depending on the size/# of files in their 'real world' test). No idea. That said, it does show a relative performance difference... but people may wonder, how did review A get xxx rate and review b got xxx rates?

As far as reads not being important and most good enough.. I can agree with the latter, certainly not the former. As trents said, this clearly depends on how you use the USB stick. Me for example, I have a 128GB Sandisk Extreme Pro and there is around 60GB of data on it that I need to be portable. Because of this, I send all the data (spreadsheets, databases, pictures, OS images, .exe for applications) to the SanDisk. When I need to leave to work, I pull the drive and go. So, reads for me are quite important and part of the reason I bought such a bangin fast drive in the first place.

That said, I'm curious to know a couple of things about your testing. Knowing this information can help passers by understand a bit better. How big, exactly, and how many files did your 10s of GB transfer? Where did you get that result...watching the transfer window? Is that value peak, sustained, or average? Were the drives empty and formatted before the tests? I assume it was teh same exact data transferred to each? Anyway, for giggles, I posted a pic of my Sandisk transferring 32GB of files in ~2 minutes along with ATTO in the background at the bottom. Love that 300MB+ writes!!!

Last, I am not sure what you mean by 'luck' and 'batch' with these. If you had a bucket full of the same USB sticks and pulled out a handful, they should all perform very close to each other. There aren't any 'factory freaks' that are faster or slower than the others. The reason why larger drives tend to perform faster than slower of the same ilk is due to NAND chip count on the higher capacity chips. In essence, the controller can access the NAND with more channels in parallel. Now, this isn't The Gospel, there are exceptions, but this is due to hardware configurations and not luck or getting a good batch.

Anyway, thanks again for the information! It's good to see some real results! I wanted to supplement the information and fill in some gaps. :)

4.jpg
3.jpg
 
Last edited:
OP
c627627

c627627

c(n*199780) Senior Member
Joined
Feb 18, 2002
So first of all to reiterate that it is nothing short of criminal, that writing speeds close to real-life performance are not mentioned by many sites... While I agree that professionals like yourselves EarthDog and ATM do have a greater need for reading speeds, it is the writing speed that every day person needs more - so they can quickly backup XYZ.

WRITING SPEED. That's what it's *all* about. That is where they all differ, significantly.
The fact that writing speeds are not even mentioned so often... I have no explanation for, other than widespread computer hardware illiteracy / ignorance.
Reading speeds are so much higher universally - it's not that they are irrelevant - it's just that they are higher, sometimes by several factors over the one true measurement of performance for an average person using a flash drive - the WRITING speed.

I also do not see the importance of writing speed accentuated by computer hardware reviewers, to teach the illiterate. They just mention what it is and leave it at that. But it is *difficult* to find out the writing speeds just by doing a search. No doubt.


I have just had a chance to review a brand new 32GB BAR PLUS for myself yesterday and am surprised to see it match the low/abysmal performance of an old 32GB BAR, measuring only 24MB/s, both of them, both 32GB BAR and 32GB BAR PLUS vs. 60MB/s of writing speed that a BAR PLUS 128GB has, a truly different product. So 24MB/s for 32GB vs. 60MB/s for 128GB version.

I now agree that "batches" are not a thing. There could be faulty drives but as a rule, buying a 128GB BAR PLUS instead of a 64GB/32GB will get you faster writing speeds, sometimes 300% faster speeds, that's each time you backup something, three times faster - truly worth the extra one time cost to buy a 128GB BAR PLUS vs. 64GB/32GB versions.

So I noticed the difference in speeds in every day life usage, differences I could feel but decided to use an identical huge file to test these.
Yes, I've always used ATTO and AS SSD Benchmark software and USBFLASHSPEED.com which I freaking KNEW I never should have used because of their Russia-based servers and their site is totally down now and has been for a long while, and I used it to generate a screenshot of performance years ago. So now I can't use them to generate the same final screen shot, but that's another story. LESSON learned about reliability of Russian-based servers.

But I digress, I decided to use a huge Blu-tay file after I discovered how much a 4K DVD player upscales Blu-rays. So I would take a humongous 30GB Blu-ray movie file and stop-watch measure copy-paste of this identical gigantic file onto a freshly formatted USB Flash drive - in the process forcing the possible overheating of the Flash Drive due to the enormous single file size. But this is unnecessary for Samsung drives, which do not overheat as much, whereas SanDisk of course craps out due to overheating and stops file transfer until it cools down. Making SanDisk nothing but USELESS for large file transfers. Another thing I have not seen reviewers accentuate - to warn people not to buy SanDisk USB Flash Drives for this reason.


Real Life Write Speeds when transferring large amounts of data:
=================

24MB/s Samsung BAR 32GB
24MB/s Samsung BAR PLUS 32GB
33MB/s Samsung DUO 128GB
37MB/s Samsung DUO 64GB
40MB/s Samsung DUO PLUS 128GB
60MB/s Samsung BAR PLUS 128GB
96MB/s Samsung DUO PLUS 256GB
 
Last edited:

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
That's Me, ATM, and trents who said it depends on use. If you use a USB stick for backups, I agree you want writes to be fast.. But, I use HDD's and the cloud for backups. To reiterate depends on who is using it, what it is being used for, and how large the files are that matter.

I think you are being overly dramatic calling reviews/specs it "nothing short of criminal". As was explained earlier, and if you look at any reviews that do real-world testing (we started a couple of reviews ago, note... thanks David!) you can see this. A bit of common knowledge also goes a long way in that larger file sizes will transfer over faster. It has worked this way since HDDs... so it isn't something new to most users. If that is 'criminal' so is just about every specification a mfg puts out.

Maybe its just me, but I feel a more typical use scenario is for transporting files from different PCs or locations to use, not for backups. But who knows, really. You're sharing how you use it, and others use it differently... which brings things back full circle - it depends on the use case as to what metric is the most important.

Many benchmarks are 'theoretical'. This holds true for SSDs and HDDs as well as USB sticks. YMMV should always be in a reader's head. Especially one who's been in computing for years. As I said, the problem with reviewers doing real world is that every review would be different depending on the size of data they are transferring and the file size. So if someone puts 32 1GB files and runs that as real world, you get different results than copying over your Windows folder at 32GB as there are thousands of files versus 32 massive ones. Which one is more accurate? The former is more accurate to me for my uses... the latter, you likely. Are either wrong?

To that end, I feel it is beneficial in reviews to see both theoretical results and real world. The reader still has to know that results will vary depending on file size no matter what.

As far as Sandisk... you saw my results on that 2017 drive (and why I asked how you define GBs of data..... the ~2 min 32GB transfer (32 files) was consistent throughout and clearly did not overheat in a similar testing scenario. Clearly the one I have works and is not useless for transferring large files as shown above. I'd also imagine that a bunch of small files would be more difficult to manage than fewer larger ones (but I haven't tested this). My thinking is the controller has to process more files thus working harder and why the lower the file size, the less the transfer speed is (see ATTO).
 
Last edited:
OP
c627627

c627627

c(n*199780) Senior Member
Joined
Feb 18, 2002
So Google
Write Speed of Samsung 64GB Duo
After all these years, it's the second half of 2020, you can't get the idea of its write speed, unless you own it.
Generally sites do not mention write speeds. Criminal is an exaggeration to accentuate negligence of sites selling/reviewing these.


I tested it using a USB 3 port of course, consistent factors were same port/same file, a 30GB Blu-ray file, a single file.
I ran Benchmarks on all of them too but I found the comparison *between them* under same circumstances relevant.


So as long time ago, for my car I bought a bunch of 32GB SanDisks, these:

SanDisk.jpg

I noticed that when I filled these with music files all the way up, that the file transfers would go for a while THEN stop / pause pause pause pause pause pause / continue and so on.
IT WAS COMMON KNOWLEDGE on Hot Deals forums that SanDisk overheat like this and to stay away from them, common knowledge. Yet not so common knowledge when you read reviews.


So in short. all reviewers tell you that it's complicated. It depends. And I don't think it is that complicated.
I could throw 30GB worth of 10,000 files on these and re-do the test, are you saying the results would be vastly different?
 
OP
c627627

c627627

c(n*199780) Senior Member
Joined
Feb 18, 2002
Additionally: Using a cloud for business - yes.
Using a cloud for personal? Insane for average Joe.

To depend on the existence of power/internet service/cloud service/third parties for your own personal data? Not smart.
128GB/256GB USB Flash drives allow you to backup everything in the palm of your hand and have it not be dependent on power/internet service/cloud service/third parties. Smart.


Here's my crown argument: I just transferred the same 30GB file in couple of minutes back to the computer from a 128GB DUO PLUS.

Measured it to be just under 250MB/s. That's why the 'who cares' attitude about read speeds.
The write speed is 40MB/s. So it takes FAR longer to write, the read/write difference is a couple of minutes vs. eleven minutes for a 30GB file. It is therefore far, far more important to know the general write speed of a USB Flash Drive for an average Joe.
 

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
So Google Write Speed of Samsung 64GB Duo. After all these years, it's the second half of 2020, you can't get the idea of its write speed, unless you own it.
If I google "Samsung 64GB duo" I don't even run across any reviews of the drive in the first place. USB sticks aren't commonly tested as much as SSDs/HDDs, etc. There are literally hundreds of different brands and models which makes it difficult to find regardless.
https://www.google.com/search?rlz=1...hUKEwi105zyt4TrAhXMZc0KHcukDzQQ4dUDCAw&uact=5

Generally sites do not mention write speeds. Criminal is an exaggeration to accentuate negligence of sites selling/reviewing these.
The generally mention the write speeds... as spec'd out by the MFG or against the benchmark they run. Real-world results are not common, however...indeed. Negligence is also being overly dramatic.

I noticed that when I filled these with music files all the way up, that the file transfers would go for a while THEN stop / pause pause pause pause pause pause / continue and so on.
IT WAS COMMON KNOWLEDGE on Hot Deals forums that SanDisk overheat like this and to stay away from them, common knowledge. Yet not so common knowledge when you read reviews.
No idea what model that is... but it speaks to what I am saying, no? Something like that (overheating) varies by product. Again, mine doesn't in my test of 32, 1GB files, but maybe it does transferring smaller and hundreds more music files in the same amount of time? Either way, my device does not heat up for either situation.

I could throw 30GB worth of 10,000 files on these and re-do the test, are you saying the results would be vastly different?
Yes. That is why I posted the ATTO results was to show you that different file sizes allow for different R/W speeds.

Here is a shot of my Windows folder being copied to the same Sandisk drive. According to properties, this folder is 19.4GB and contains 137,394 files in 36,894 folders...peak was ~32MB/s...It took 2/3 the data about twice as long to write to this drive (don't mind the time on the image, as you know it varies depending on speed and what is left. I saw it peak at 8 minutes at one time.

7.jpg
8.jpg

EDIT: I take that back... it is still going on these last files... LOL. The image looks the same as the bottom one (data values are different of course), but its still spinning through files...the drive isn't hot to the touch ftr. HWinfo is reading the temps on this USB wrong... its below ambient... but the drive is not even hot to the touch right now, several minutes in.


Additionally: Using a cloud for business - yes.
Using a cloud for personal? Insane for average Joe.

To depend on the existence of power/internet service/cloud service/third parties for your own personal data? Not smart.
128GB/256GB USB Flash drives allow you to backup everything in the palm of your hand and have it not be dependent on power/internet service/cloud service/third parties. Smart.


Here's my crown argument: I just transferred the same 30GB file in couple of minutes back to the computer from a 128GB DUO PLUS.

Measured it to be just under 250MB/s. That's why the 'who cares' attitude about read speeds.
The write speed is 40MB/s. So it takes FAR longer to write, the read/write difference is a couple of minutes vs. eleven minutes for a 30GB file. It is therefore far, far more important to know the general write speed of a USB Flash Drive for an average Joe.
To your double post...

Using the cloud is insane for backups the average Joe? We'll agree to disagree on that point. Clearly, if it is warm storage and accessed frequently, chances are you would want it on a HDD/SSD. Backups, to me, are cold storage and not frequently accessed where a cloud is perfectly reasonable for an average joe to use. It's part of the reason they exist. A cloud isn't (shouldn't be) a PRIMARY backup location. It should back up a HDD/SSD that is local. I wish my 'personal files' (images over the past 15 years, documents, movies/shows etc) fit on a 256GB drive... so I have no choice except to put it on a HDD/SSD. I would also guess that HDDs/SSDs are more reliable than some of these USB sticks... but that is just a guess.

We know, you don't care about read speeds (though transferring that data to something is read dependent). Others do. How c6 uses XX isn't how me, trents, name here, name here, name here, may use their device. Every use model is different. Both reads and writes are important. One can have more importance over the other but, again, it varies by the user and their needs. Are reads, generally fast enough? Sure... I concede that as I did earlier), but make no mistake about it, the importance of it varies by users and use scenario.
 
Last edited:

Robert17

Premium Member
Joined
Feb 1, 2011
In the real world, when I am backing up some data to a USB stick I set the transfer to 'start', go to the restroom, refill my drink, play with the puppers for a few minutes and come back to a completed task. So it's not that I don't care what the R/W speeds are, I've simply adjusted my use of the downtime to allow for whatever the disk takes. YPMV (Your Puppers May Vary).
 
OP
c627627

c627627

c(n*199780) Senior Member
Joined
Feb 18, 2002
Rather than one file, let's take a 20.6GB folder with about sixty one thousand files inside.

The stop-watch test revealed write speeds of
• 20MB/s Samsung BAR 32GB
• 27MB/s Samsung DUO PLUS 128GB

Absolutely different from a single 27929 MB Blu-ray file write speed of
• 24MB/s Samsung BAR 32GB
• 40MB/s Samsung DUO PLUS 128GB


Significant percentage difference for both cases.
Write speed differences are far more important because they take so much longer, many hundreds of percent of time longer and each time you do a file transfer. Write speed takes ages more, so it's very important to know how much faster a USB Flash drive writes vs. another drive. That is the true test because they all write so much slower than read, so it takes so much longer when writing vs. reading.
 
Last edited:

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
So, to me, outside of the valuable real-world write data you provided on a couple of drives, I think the moral of the story is to know what you are using your device for (the type of data that is being transferred too) and what is most important to you. If you are the type of user who consistently writes a shed load of data to a drive, then getting one with faster writes would certainly be beneficial. Like Robert, if I do anything on a drive like that, I walk away or do something else while its working. Watched pots never boil... or something. :p

If write speeds are REALLY a concern, get a 128/256GB NVMe based M.2 drive and an external enclosure with USB 3.2 Gen 2 Type-C. Sure, the price will likely be higher (I can get in the ballpark and be faster), but if time is money or this is really an issue for a use case, there are better faster options that are actually tested as you want.

Thank you for your detailed information on your drives. :thup:
 
Last edited:

trents

Senior Member
Joined
Dec 27, 2008
So, to me, outside of the valuable real-world write data you provided on a couple of drives, I think the moral of the story is to know what you are using your device for (the type of data that is being transferred too) and what is most important to you. If you are the type of user who consistently writes a shed load of data to a drive, then getting one with faster writes would certainly be beneficial. Like Robert, if I do anything on a drive like that, I walk away or do something else while its working. Watched pots never boil... or something. :p

If write speeds are REALLY a concern, get a 128/256GB NVMe based M.2 drive and an external enclosure with USB 3.2 Gen 2 Type-C. Sure, the price will likely be higher (I can get in the ballpark and be faster), but if time is money or this is really an issue for a use case, there are better faster options that are actually tested as you want.

Thank you for your detailed information on your drives. :thup:

Hee, Hee! Good one.
 

dejo

Senior Moment Senior Member
Joined
Nov 11, 2001
I have had an old Sandisk Extreme 32GB USB3.0 drive for 7 years or so and it is still as good at writing files as any flash drive I have seen. It also still works great to this day.
 
OP
c627627

c627627

c(n*199780) Senior Member
Joined
Feb 18, 2002
So what you do is, if for example you have a 32GB SandDisk, then dump 32GB worth of files onto it. After a while you will notice that it stops after a while to cool down. Perhaps only some models were affected.
This was constantly being mentioned on forums that talk about things going on sale, hot deal forums. And I personally own more then a dozen SanDisks with this problem when I tried to paste 32GB worth of uncompressed flac music files and mp3 files onto them, all at once.

It was also common knowledge that Samsungs do not overheat, in contrast to SanDisk flash drives.

Why this is being heard for the first time on overclockers, one of the oldest computer hardware web sites in the world, is definitely strange, it's 2020, this like a thing, SanDisks overheat... I can't trust them to buy another one to see if it still is a thing today.
 

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
Outside of being a great song in the late 80s, once bitten twice shy is a good mantra to have. :)

I can only speak to my experience with Sandisk and, no overheating on this drive. Perhaps it's more prevalent on the cheap older smaller drives compared to the more premium model I purchased a couple of years back? Who knows. I also wonder if all the issues are truly overheating issues as well. Some may be windows just getting in the way chugging through files or off the cache and onto something slower. I've seen that before as well.

Thanks for bringing attention to it at OCF...buyer beware on old small inexpensive Sandisks!
 
OP
c627627

c627627

c(n*199780) Senior Member
Joined
Feb 18, 2002
The moral of the story for the record, is that many users are not actually really, completely and fully aware that write speeds can be six hundred percent or more time slower than the drive's read speeds, making the write speed way, way more important as far as the number of seconds/minutes of your time being wasted...

When it comes to writes, Samsung Bar Plus 128GB is fifty percent faster than its 32GB equivalent and I would venture a guess it is faster than most of its competition.
256GB version's writes are more than a hundred percent faster. Significant time saver short of building an NVMe based M.2 drive and an external enclosure with USB 3.2 Gen 2 Type-C - which sadly won't fit in my little jeans pocket - but maybe if I bulk up? :D
 

EarthDog

Gulper Nozzle Co-Owner
Joined
Dec 15, 2008
Location
Buckeyes!
For the most part I agree. But you can use your PC and do other things while it is transferring. It's not like it seizes your PC and glues your *** to the chair. Iif writes are important for a user, as you said the larger the drive, generally, the faster they are... go larger and weigh its importance more.

I have to say I feel that many already know the larger the drive the faster as it has been this way with SSDs too for over a decade.


Oh, and stop wearing skinny jeans you hipster. :p :D
 

wagex

Chapstick Eating Premium Member
Joined
Jan 14, 2011
Additionally: Using a cloud for business - yes.
Using a cloud for personal? Insane for average Joe.
.

just wanted to jump in here... i back my stuff up to the cloud (personal) as well.... its not uncommon at all and far from insane, its so cheap, it's cheaper than a good thumb drive.

Lastly before you buy you can easily look up reviews of the drive online and generally people will post atto or crystal disk bench marks in the reviews on amazon or what ever other store and will complain the drive are slow. Lack of due diligence is the only excuse for not knowing a drives real world speeds unless its a brand new item to market.
 

Janus67

Benching Team Leader
Joined
May 29, 2005
another +1 here for someone that backs up to the cloud. My pictures from my phone automatically back up to my home NAS as well as to google photos (Pixel phone). My NAS backs up automatically to Crashplan which has worked fine for me as well. I currently have ~18TB of data backed up w/o any issues for around $12/mo or something.
 

dejo

Senior Moment Senior Member
Joined
Nov 11, 2001
I have had several PNY flashdrives fail with no warning as well as a single Samsung. I do have several sandisk drives and all but the Extreme or Extreme Pro have been noting special- and would mostly agree with your assessment