• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Kingston NV2 1 TB M.2 NVMe SSD

papayoyo said:
I have the 2TB NV2 bought in the UK, fitted internally in an Asus laptop... I can't get Flash ID to work no matter what I try... latest version v0.564a, run as administrator, tried in safe mode, with/without the "u" modifier, running the exe directly or through command line... always the error message "IOCtl: Identify failed 0x32!". Could it be a driver problem, or windows 11?

Benchmarks on the freshly formatted drive:


View attachment 275070View attachment 275071View attachment 275072
Click to expand...
I guess you have the TLC version based on your scores...
 
papayoyo said:
I have the 2TB NV2 bought in the UK, fitted internally in an Asus laptop... I can't get Flash ID to work no matter what I try... latest version v0.564a, run as administrator, tried in safe mode, with/without the "u" modifier, running the exe directly or through command line... always the error message "IOCtl: Identify failed 0x32!". Could it be a driver problem, or windows 11?

Benchmarks on the freshly formatted drive:


View attachment 275070View attachment 275071View attachment 275072
Click to expand...

i used http://vlo.name:3000/ssdtool/

try both phison and smi
 
Went do some Appliance shopping with the Wife and Mother in Law. After they were finished they wanted to go to the Dollar Store. Right beside the Dollar store was a Memory Express. While they were in there is was dispatched to find a Mouse and Keyboard set for my nephew. I picked up a headset for my niece as well and when I was waiting for the associate to get my headset I saw the NV2 on sale. I ended up getting a 1 TB ($72) and will be happy to test it against the 2 TB variant.
 

Attachments

  • IMG_20221219_162506649.jpg
    IMG_20221219_162506649.jpg
    3.4 MB · Views: 408
I worked out how to get Flash ID working - had to disable VMD in BIOS.

Results for my 2TB NV2 posted previously... It's QLC...

Drive : 1(NVME)
Scsi : 1
IOCtl: NVME_OF failed 0x1!
Driver : W10
Model : KINGSTON SNV2S2000G
Fw : SBM02103
HMB : 65536 - 65536 KB (Enabled, 64 M)
Size : 1907729 MB [2000.4 GB]
LBA Size : 512
Controller: [SM2267XT] bufferless
Bank00: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank01: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank02: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank03: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank04: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank05: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank06: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank07: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank08: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank09: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank10: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank11: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank12: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank13: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank14: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Bank15: 0x89,0xd3,0xac,0x32,0xc6,0x0,0x0,0x0 - Intel 144L QLC 1024Gb/CE 1024Gb/die
Channel: 4
CE : 4
Plane : 4
Die/Ce : 1
Ch map : 0x0F
CE map : 0x0F
Bit Per Cell : (SLC)
Inter. : 4
Pages/Block : 9216(2304)
First Fblock : 1
Total Fblock : 258
Total Hblock : 29809
Fblock Per Ce : 258
Fblock Per Die: 258
Original Spare Block Count : 21
Vendor Marked Bad Block : 0
Bad Block From Pretest : 6
IOCtl: NVME_W10_GetLog 0x45d!
 
Here are the sustained write & read speeds on my 2TB QLC NV2.

Compared to the reviewed drive and the results posted earlier by member cts, the QLC sustained writes in particular are pretty abysmal. It took 5 hours to fully write to the 2TB QLC drive, vs 34 minutes for the 1TB TLC drive.

QLC is around a quarter the speed of TLC in sustained writes (bottoming out at 80MB/s when cache is full - slower than a modern mechanical drive)
QLC is around two thirds the speed of TLC in sustained reads.

Also, speeds are strangely fluctuating up & down with this drive.

Screenshot 2022-12-22 234417.png
Screenshot 2022-12-23 003158.png
 
Last edited:
KINGSTON SNV2S500G, came today:

v0.27a
OS: 10.0 build 22621
Drive : 2(NVME)
Scsi : 1
Driver : W10
Model : KINGSTON SNV2S500G
Fw : SBI02102
HMB : 65536 - 65536 KB (Enabled, 64 M)
Size : 476940 MB [500.1 GB]
LBA Size : 512
AdminCmd : 0x00 0x01 0x02 0x04 0x05 0x06 0x08 0x09 0x0A 0x0C 0x10 0x11 0x14 0x80 0x84 0xC0 0xC1 0xC2 0xE0 0xE1
I/O Cmd : 0x00 0x01 0x02 0x04 0x05 0x08 0x09
Controller: [SM2267XT] bufferless
Bank00: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank01: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank02: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank03: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank08: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank09: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank10: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank11: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Channel: 4
CE : 2
Plane : 2
Die/Ce : 1
Ch map : 0x0F
CE map : 0x05
Bit Per Cell : (TLC)
Inter. : 2
Pages/Block : 1344(448)
First Fblock : 1
Total Fblock : 1662
Total Hblock : 29809
Fblock Per Ce : 1662
Fblock Per Die: 1662
Original Spare Block Count : 28
Vendor Marked Bad Block : 0
Bad Block From Pretest : 15
IOCtl: NVME_W10_GetLog 0x45d!

SM2263XT Telemetry data:
 
arandjele said:
KINGSTON SNV2S500G, came today:

v0.27a
OS: 10.0 build 22621
Drive : 2(NVME)
Scsi : 1
Driver : W10
Model : KINGSTON SNV2S500G
Fw : SBI02102
HMB : 65536 - 65536 KB (Enabled, 64 M)
Size : 476940 MB [500.1 GB]
LBA Size : 512
AdminCmd : 0x00 0x01 0x02 0x04 0x05 0x06 0x08 0x09 0x0A 0x0C 0x10 0x11 0x14 0x80 0x84 0xC0 0xC1 0xC2 0xE0 0xE1
I/O Cmd : 0x00 0x01 0x02 0x04 0x05 0x08 0x09
Controller: [SM2267XT] bufferless
Bank00: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank01: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank02: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank03: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank08: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank09: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank10: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Bank11: 0x98,0x3e,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 512Gb/CE 512Gb/die 2Plane/die
Channel: 4
CE : 2
Plane : 2
Die/Ce : 1
Ch map : 0x0F
CE map : 0x05
Bit Per Cell : (TLC)
Inter. : 2
Pages/Block : 1344(448)
First Fblock : 1
Total Fblock : 1662
Total Hblock : 29809
Fblock Per Ce : 1662
Fblock Per Die: 1662
Original Spare Block Count : 28
Vendor Marked Bad Block : 0
Bad Block From Pretest : 15
IOCtl: NVME_W10_GetLog 0x45d!

SM2263XT Telemetry data:
Click to expand...

i like how our 500GB models are both TLC :cool:
 
Summary of actual state:
  • reviews
    • most of the reviewers got test samples from Taiwan built before y22cw40 equipped with faster TLC memory
    • an exception is the computerbase.de test for example, they got an 2GB, Taiwan, 2243 with SMI & TLC with 002.A00G and Firmware SBK00104
  • speed
    • QLC memory random write speed is lower, example 1TB, SMI, CDM8.04 (1GB), RND4K Q32T1
      • QLC: 295, see #78
      • TLC: 490 , see #47
      • the values differ depending from drive size and general conditions (test programm, hardware) - example: computerbase.de test results on 2TB TLC: 344
    • from general specs the Phison controller should be a little bit faster than the SMI
  • used memory
    • option 1) install the ssd and use a software tool to see what memory is used, see #103 (you have to use the right programm for your controller otherwise you will get an error message and no result)
      • for SMI controller: "SMI NVME SSD flash id v0.27a"
      • for Phison controller: "Phison NVME SSD flash id" (working with E21 not confirmed)
    • option 2) when removing the sticker (warranty void) it is possible to distinguish the used memory, see #93
      • TLC drives has memory 'Toshiba 122 BiCS5' with the following terms: FxxxxxxUCT1
      • QLC drives has memory 'Intel 144L QLC' with the following terms: FxxxxxxUCN1
  • used controller
    • option 1) the PCBs for SMI and Phison can be visually distinguished:
      • SM2267XT controller has many semiconductor components below the memory chips, see #41
      • Phison E21T controller has only a few components below the memory chips, see techpowerup example
    • option 2) when removing the sticker (warranty void) it is possible to distinguish the used controller, see #93
 
Last edited:
In december I bought at Alternate (EU) a Kingston NV2, 2000GB / 2TB SSD and placed it in a Deskmini A300 as a secondary SSD without cooling.
Therefore speeds may not be optimal. Also I could not get the SMI info working. Synthetic CDM seems very good. Aida64 write is much lower.
I cannot explain the 1000 MB/s difference in maximum write speed between Crystaldiskmark and Aida64.

2022-12-27 10_02_29-Window.png


2022-12-27 16_38_22-Window.png
 
I also recently purchased Kingston NV2 M.2 NVMe SSD. It is a beast. Made comparison tests and review with SanDisk Ultra M.2 NVMe 3D. But unfortunately not allowed to post links here.

TL; DR a very worthy SSD for its price
 
MrTPU said:
@Relaxaflexer Do you have a picture of the SSD sticker/ do know the production date/firmware version?
Click to expand...
The Firmware number is, SBM02103.
Data from CD info.

Edit:
Something strange happened today. I had to format the Kingston 2TB NV2 for the second time. Windows 10 would not recognize it anymore as a working disk.
If this is normal after a write run of Aida64 then pls. let me know.
 
Last edited:
I will provide another data point. I have purchased a 2TB version from a Nordic reseller. It's TLC with SM2267XT controller.

Code: 
Drive     : 0(NVME)
Scsi      : 1
IOCtl: NVME_OF failed 0x1!
Driver    : W10
Model     : KINGSTON SNV2S2000G
Fw        : SBK00104
HMB       : 65536 - 65536 KB (Enabled, 64 M)
Size      : 1907729 MB [2000.4 GB]
LBA Size  : 512
Controller:  [SM2267XT] bufferless
Bank00: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank01: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank02: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank03: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank04: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank05: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank06: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank07: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank08: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank09: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank10: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank11: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank12: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank13: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank14: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Bank15: 0x98,0x48,0x98,0x3,0x76,0xe4,0x8,0x16 - Toshiba 112L BiCS5 TLC 16k 1024Gb/CE 1024Gb/die 2Plane/die
Channel: 4
CE     : 4
Plane  : 2
Die/Ce : 1
Ch map : 0x0F
CE map : 0x0F
Bit Per Cell  : (TLC)
Inter.        : 4
Pages/Block   : 1344(448)
First Fblock  :    1
Total Fblock  : 3222
Total Hblock  : 29809
Fblock Per Ce : 3222
Fblock Per Die: 3222
Original Spare Block Count :    28
Vendor Marked Bad Block    :     0
Bad Block From Pretest     :    46
IOCtl: NVME_W10_GetLog 0x45d!
 

Attachments

  • 4.png
    4.png
    34.9 KB · Views: 178
Last edited by a moderator:
Relaxaflexer said:
The Firmware number is, SBM02103.
Data from CD info.

Edit:
Something strange happened today. I had to format the Kingston 2TB NV2 for the second time. Windows 10 would not recognize it anymore as a working disk.
If this is normal after a write run of Aida64 then pls. let me know.
Click to expand...
Could you post a screenshot with the error that windows show ?

CatalinT said:
Could you post a screenshot with the error that windows show ?
Click to expand...
You tried the windows disk management utility ?
 
CatalinT said:
Could you post a screenshot with the error that windows show ?


You tried the windows disk management utility ?
Click to expand...
Sorry, I don't have screenshots of these events.

What happened was:

Installed the SSD. It popped up in windows that it had to be formatted. I did format it (quick NTFS) under disk management. Did some speedtests with CDM and Aida64.
The next day to my surprise the disk did not appear in my computer: disks. So I looked in disk management and there the disk was seen as not formatted or initialized. :confused:
Formatted it again and now the format seems to be "holding". Never had this B4 in 35 years of harddrives and SSD management. So I figured maybe Aida64 causes this?

Have to say that this SSD is not installed very good. Could not find a screw, so its held in place with tape and it has no cooling at all, not even airflow.
 
Relaxaflexer said:
Sorry, I don't have screenshots of these events.

What happened was:

Installed the SSD. It popped up in windows that it had to be formatted. I did format it (quick NTFS) under disk management. Did some speedtests with CDM and Aida64.
The next day to my surprise the disk did not appear in my computer: disks. So I looked in disk management and there the disk was seen as not formatted or initialized. :confused:
Formatted it again and now the format seems to be "holding". Never had this B4 in 35 years of harddrives and SSD management. So I figured maybe Aida64 causes this?

Have to say that this SSD is not installed very good. Could not find a screw, so its held in place with tape and it has no cooling at all, not even airflow.
Click to expand...

This is normal with Aida64. It warns "write tests will destory all data on this drive". Just intialise and format the drive again.
 
Thanks for the info on Aida64 @papayoyo

Finally I managed to find out my 2TB Kingston NV2 has intel QLC 144L chips with the SM2267XT controller.
Could not copy the text or make a screenshot though.
Also I did some copy tests from the boot drive SS 970 Evo, with Windows explorer original and with Extreme Copy pro.
Both show write speeds up to 2800 MB/s. So that is a lot faster then Aida64 was showing as a max speed: 1880 MB/s.
I cannot explain the speed difference between Aida64 and Real copy tests. Also I cannot explain why this NV2 is twice as fast as @papayoyo 's NV2?
 
papayoyo said:
Yes, this is strange, for two apparently identical 2TB QLC drives . Could be driver related? I'm using Intel RST VMD driver. Check yours in device manager > storage controllers
Click to expand...
I looked in Device manager. It only says Microsoft for NVM Express storage controller and Standard NVM express controller.
It's on a AR deskmini AMD AM4 A300 board. AM4 is S.O.C. based so almost everything is on the CPU and not on the board. It seems there are no AMD NVME controllers in W10.

I am not saying you should or if its possible (or risky?) but maybe you can get better results with the standard W10 Microsoft NVME controller?
Then there is also the M.2 slot that the SSD is placed in. Sometimes secondary M.2 slots have less bandwidth or have combined (key B, M) features for both Sata and NVMe?
 
papayoyo said:
Yes, this is strange, for two apparently identical 2TB QLC drives . Could be driver related? I'm using Intel RST VMD driver. Check yours in device manager > storage controllers
Click to expand...
Do you remember how full your NV2 was before you ran the Aida64 write benchmark? Also, same question to @Relaxaflexer .

Here's my assumption. The more data the drive contains before the start of the test, the more time it will take during the test to erase that data. Erasing is slow and it's inevitable before writing, whether it's files within a file system within a partition or just test data that ignores any file system and partitions.

If you ever decide to burn some precious TBW and run the benchmark again, make sure the drive is (mostly) empty beforehand and give it enough time to digest the TRIM commands and (hopefully) erase all the stuff. A few hours would seem right because that's as much as your last Aida64 run took.
 
Wirko said:
Do you remember how full your NV2 was before you ran the Aida64 write benchmark? Also, same question to @Relaxaflexer .

Here's my assumption. The more data the drive contains before the start of the test, the more time it will take during the test to erase that data. Erasing is slow and it's inevitable before writing, whether it's files within a file system within a partition or just test data that ignores any file system and partitions.

If you ever decide to burn some precious TBW and run the benchmark again, make sure the drive is (mostly) empty beforehand and give it enough time to digest the TRIM commands and (hopefully) erase all the stuff. A few hours would seem right because that's as much as your last Aida64 run took.
Click to expand...

It was freshly formatted and empty, just before the Aida64 write test. Previously there was about 20GB written to it. Not sure if this matters? Might try again with a 24-hour gap to let TRIM do it's thing.
 
My NV2 was New, Formatted and never had any data on it before the Aida64 test.
After Aida64 and a second format the NV2 was 1000 MB/s faster in real life write tests. :confused:
 
You must log in or register to reply here.
Back
Top