Good, bad or fake Samsung card?

[Derekmillion]

I bought this 512gb Samsung Evo Select U3 card some time ago from Amazon.
It's the older style green card before they changed to blue recently.
I've had a few dash cams over the years, some of them take it, some of them don't, but I never ran any tests on it before.
Can anyone tell me if these results are normal for this card?
Do the speeds represent what a Samsung card would put out or is it a fake?

The card was formatted using GUI Format to FAT32.
Tests performed are with Crystal Disk Mark 8, H2testw and SpeedOut.

Thanks

 

[kamkar]

I think that is fine ( not familiar with 512 Gb cards or that particular model )
I have the new blue and white cards, and the white dont really want to work in any of my dash cameras, but it appear to work fine in the action camera that record 4K/60 FPS, though i have of course not recorded the full 256GB in that camera.

A fake card can still be fast OK, but the fake memory size programs would not find, so they would just test on the actual memory size. but the 80 / 90 MB/s is also what i get testing U3 / V30 cards,

 

[Kremmen]

That write speed is dreadful in my opinion.

Dashcams these days ideally need to be over 50Mbps to avoid a bottleneck.

Here is a link to my Samsung Evo card I posted ages ago : https://dashcamtalk.com/forum/threa...ut-unplayable-on-computers.47389/#post-578470

 

[kamkar]

Well i clearly did not handle as little sleep as i got yesterday well i was looking at the read speeds only
I have seen the same a few times now with 256 GB cards, try and run it over with SD Formantter, that restored the lagging speed for me.
Dunno why it is like that with brand new cards, but i have tried it 2 times lately

 

[DT MI]

Agree the write speed is on the slow side.

Some time back I had a somewhat similar issue with a new Samsung 512 card that was resolved by doing a low-level format with SDFormatter (https://www.sdcard.org/downloads/formatter/) and the card has been fine ever since. Might be worth a try.

 

[kamkar]

The cards i experienced slow out of blister pack speeds with was the Adata high endurance, and then the white Samsung ( EVO pro ) card too
Sdformatter restored speed just fine, but both cards will not work in viofo cameras, but will work in vantrue E2 and my DJI Osmo Action camera.

I have never experienced this with smaller memory cards, only 256GB ones.

 

[Derekmillion]

Dashcams these days ideally need to be over 50Mbps to avoid a bottleneck.

I have seen the same a few times now with 256 GB cards, try and run it over with SD Formantter, that restored the lagging speed for me.

Some time back I had a somewhat similar issue with a new Samsung 512 card that was resolved by doing a low-level format with SDFormatter

I have taken everyones feedback and performed a full overwrite format using sd formatter and ran the tests again (minus H2testw as the first test shows no errors)
The write speeds seem vastly improved! So what does this mean exactly? Does the file system make a difference in speeds fat32 vs exfat?
Anyways I'm happy the card is now performing faster.

 

[DT MI]

...So what does this mean exactly? Does the file system make a difference in speeds fat32 vs exfat?...

From the SDCard.org page:

It is strongly recommended to use the SD Memory Card Formatter to format SD/SDHC/SDXC Cards rather than using formatting tools provided with individual operating systems. In general, formatting tools provided with operating systems can format various storage media including SD/SDHC/SDXC Cards, but it may not be optimized for SD/SDHC/SDXC Cards and it may result in lower performance.

It could be that formatting with GUIFormat adversely affected the performance of the card. In the case I mentioned in my post I suspect that's what happened to me - a 'corrupted' low-level format caused by my Win10 system rebooting to apply an update when I was doing the format overnight.

 

[Derekmillion]

It could be that formatting with GUIFormat adversely affected the performance of the card.

My next question is, what would you do if the dash cam you're using required the card to be formatted as FAT32?

Scenario:
You purchased a new dash cam and a new sd card.
The sd card comes pre-formatted as Exfat, the best idea is to format the card using the dash cam.
But let's say the dash cam won't recognize the card until it's formatted as FAT32.
From what you've said here, GUIFormat will reduce the performance of the card and possibly cause an issue for the dash cam recording.

Are there any other programs to convert to FAT32?

 

[DT MI]

My next question is, what would you do if the dash cam you're using required the card to be formatted as FAT32?

Scenario:
You purchased a new dash cam and a new sd card.
The sd card comes pre-formatted as Exfat, the best idea is to format the card using the dash cam.
But let's say the dash cam won't recognize the card until it's formatted as FAT32.
...

Can't help you there - all my cameras recognized any SD card I've used.

...
From what you've said here, GUIFormat will reduce the performance of the card and possibly cause an issue for the dash cam recording.

No, I said GUIFormat could have adversely affected the card performance - and that was based on the cautionary statement on the SDCard.org page.

...
Are there any other programs to convert to FAT32?

There are a number but everything I've found does it by reformatting, same as GUIFormat, so the same potential risk exists. It would be nice if SDFormatter provided it as an option but that doesn't appear to be the case.


As a bit of a 'test' I took a 256GB Samsung card, low-level formatted with SDFormatter then checked read/write speeds with Crystal Disk Mark. I then did a 'quick format' using both GUIFormat to FAT32 and Windows format to exFAT and ran Crystal Disk Mark after each format. There were no significant differences in performance in any of the tests. Not sure how valid that is because it was only 1 test using only 1 card so not at all scientifically rigorous - just anecdotal.

It may show that low-level formatting with SDFormatter followed by a quick format using GUIFormat might be a method of 'converting' to FAT32 safely (as in not affecting performance) but that is nothing more than speculation on my part with virtually zero data to confirm it.


After SDFormatter -

------------------------------------------------------------------------------
CrystalDiskMark 8.0.4 x64 (C) 2007-2021 hiyohiyo
Crystal Dew World: https://crystalmark.info/
------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

[Read]
SEQ 1MiB (Q= 8, T= 1): 92.033 MB/s [ 87.8 IOPS] < 89814.88 us>
SEQ 256KiB (Q= 1, T= 1): 88.604 MB/s [ 338.0 IOPS] < 2940.83 us>
RND 4KiB (Q= 32, T= 1): 11.146 MB/s [ 2721.2 IOPS] < 11738.74 us>
RND 4KiB (Q= 1, T= 1): 10.581 MB/s [ 2583.3 IOPS] < 385.93 us>

[Write]
SEQ 1MiB (Q= 8, T= 1): 80.107 MB/s [ 76.4 IOPS] <102659.92 us>
SEQ 256KiB (Q= 1, T= 1): 76.620 MB/s [ 292.3 IOPS] < 3408.91 us>
RND 4KiB (Q= 32, T= 1): 3.616 MB/s [ 882.8 IOPS] < 36012.98 us>
RND 4KiB (Q= 1, T= 1): 3.529 MB/s [ 861.6 IOPS] < 1156.41 us>

Profile: Default
Test: 1 GiB (x3) [D: 0% (0/239GiB)]
Mode: [Admin]



After GUIFormat -

------------------------------------------------------------------------------
CrystalDiskMark 8.0.4 x64 (C) 2007-2021 hiyohiyo
Crystal Dew World: https://crystalmark.info/
------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

[Read]
SEQ 1MiB (Q= 8, T= 1): 92.311 MB/s [ 88.0 IOPS] < 89987.00 us>
SEQ 256KiB (Q= 1, T= 1): 87.718 MB/s [ 334.6 IOPS] < 2986.64 us>
RND 4KiB (Q= 32, T= 1): 11.257 MB/s [ 2748.3 IOPS] < 11601.93 us>
RND 4KiB (Q= 1, T= 1): 10.120 MB/s [ 2470.7 IOPS] < 402.16 us>

[Write]
SEQ 1MiB (Q= 8, T= 1): 80.672 MB/s [ 76.9 IOPS] <102376.98 us>
SEQ 256KiB (Q= 1, T= 1): 76.062 MB/s [ 290.2 IOPS] < 3434.97 us>
RND 4KiB (Q= 32, T= 1): 2.382 MB/s [ 581.5 IOPS] < 54570.80 us>
RND 4KiB (Q= 1, T= 1): 2.366 MB/s [ 577.6 IOPS] < 1717.77 us>

Profile: Default
Test: 1 GiB (x3) [D: 0% (0/239GiB)]
Mode: [Admin]



After Windows format -

------------------------------------------------------------------------------
CrystalDiskMark 8.0.4 x64 (C) 2007-2021 hiyohiyo
Crystal Dew World: https://crystalmark.info/
------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes

[Read]
SEQ 1MiB (Q= 8, T= 1): 92.367 MB/s [ 88.1 IOPS] < 89906.63 us>
SEQ 256KiB (Q= 1, T= 1): 87.528 MB/s [ 333.9 IOPS] < 2992.25 us>
RND 4KiB (Q= 32, T= 1): 11.215 MB/s [ 2738.0 IOPS] < 11668.68 us>
RND 4KiB (Q= 1, T= 1): 10.703 MB/s [ 2613.0 IOPS] < 381.70 us>

[Write]
SEQ 1MiB (Q= 8, T= 1): 80.388 MB/s [ 76.7 IOPS] <102853.05 us>
SEQ 256KiB (Q= 1, T= 1): 76.243 MB/s [ 290.8 IOPS] < 3426.48 us>
RND 4KiB (Q= 32, T= 1): 3.623 MB/s [ 884.5 IOPS] < 35966.08 us>
RND 4KiB (Q= 1, T= 1): 3.494 MB/s [ 853.0 IOPS] < 1167.93 us>

Profile: Default
Test: 1 GiB (x3) [D: 0% (0/239GiB)]
Mode: [Admin]