Pending Read Speed reports faliure on a new Kingston A400 SA400S37/960G 960GB SSD

[talep]

I have decided to try ReadSpeed and since I am a linux user I dd the image into some USB drive.
I booted to the USB Drive and run the test.
Ande here are the results:

Ignoring now the results for the 2.0 TB HD & 512GB NVME which I had for over 8 years now, the last SSD drive by kingston was added about a month ago since I needed some space and looked for something cheap on one hand not to fast but not HDD.

It is important to mention that I live in Israel and bought it new from a reputable computer parts store called Ivory which is actually one of the 2 big chains that sales computer parts here.
So the probability of it being a fake is 0%.

I know that this drive works since before I rebooted my PC to ReadSpeed it worked....so I booted back to my linux and did some tests when the drive is mounted and it is at 25% capacity.

First test was a simple running dd creating 5GB file.

dd if=/dev/zero of=5G bs=1G count=5
5+0 records in
5+0 records out
5368709120 bytes (5.4 GB, 5.0 GiB) copied, 4.84961 s, 1.1 GB/s

This does not make sense as drive specs are talking about 450 MB/s , so it must be the linux buffers , I did things again but now using the odirect flag

dd oflag=direct if=/dev/zero of=5G bs=1G count=5
5+0 records in
5+0 records out
5368709120 bytes (5.4 GB, 5.0 GiB) copied, 20.9011 s, 257 MB/s

Well not exactly like the specs still it is working.

Last test was using dd again to read the file I have written

dd if=5G of=/dev/zero
10485760+0 records in
10485760+0 records out
5368709120 bytes (5.4 GB, 5.0 GiB) copied, 24.3732 s, 220 MB/s

Well the spec talks about 500MB/s and the speed might be related to block size ... one thin for sure it is working.

I wounder if you have encountered problems related using ReadSpeed with this specific drive ?

 

[PHolder]

We would need to know why the drive is failing. There are other utilities on the ReadSpeed drive I believe (it's been a while, I forget if they're in the image for sure or not.) If not you can download them from https://www.grc.com/dev/SpinRite/ . Download them and put them in the drive with ReadSpeed (or find them already there.) Run PCI.EXE and maybe DrivEnum.exe and send the results back here to us.

 

[Tazz]

Huh, I just tried ReadSpeed on my machine and it recognized that I had nvme drives and wouldn't display/test them. --no driver

 

[DanR]

I just tried ReadSpeed on my machine and it recognized that I had nvme drives and wouldn't display/test them. --no driver

Correct! ReadSpeed v1 only does internal AHCI/ATA/IDE drives. No NVMe or external USB drives.

Presumably someday there will be a newer version of ReadSpeed with additional capability.

 

[DanR]

the last SSD drive by kingston was added about a month ago since I needed some space and looked for something cheap on one hand not to fast but not HDD

You could try the following and report the results back here:

C:/>readspeed /ident >> rs.txt

The ident Command Line switch may ferret out a bit more info on your controller(s) and drives. The results will be in the text file rs.txt. You may replace the rs in rs.txt with whatever you wish.

 

[Tazz]

Correct! ReadSpeed v1 only does internal AHCI/ATA/IDE drives. No NVMe or external USB drives.

Presumably someday there will be a newer version of ReadSpeed with additional capability.

Yes, but look at the screen grab. It tested his. Those speeds are not SATA3.
This leads me to think that the drives are being mis-identified to RS, maybe?

 

[PHolder]

My understanding is that the code in ReadSpeed is a previous incarnation (the code used to build and debug) of the code in SpinRite. Accordingly, I believe if the drive is exposed by the BIOS to DOS, it should be able to be tested by ReadSpeed. I believe not all BIOSes will expose a NVMe drive if it wasn't booted from.

 

[DanR]

Yes, but look at the screen grab. It tested his. Those speeds are not SATA3.
This leads me to think that the drives are being mis-identified to RS, maybe?

Tazz, I think we may be talking apples and oranges here?

I read the OP as focusing on the Kingston SSD drive, drive 83, with all FAILURE results. ReadSpeed should not have an issue with that SSD drive, unless there is an issue with the controller for that drive? Hence my suggestion to run ReadSpeed with the /ident switch for possibly more hardware information.

Then there is the Samsung drive, drive 82. This is an NVMe drive. Normally ReadSpeed should not be able to access this drive., as ReadSpeed can only access drives via its native AHCI ATA, IDE drivers. That means no BIOS drives, be they be internal NVMe or external USB.

However, ReadSpeed did indeed test this internal NVMe drive, albeit at speeds well below where they should be. This suggests that the system BIOS is presenting this drive as a SATA drive, allowing ReadSpeed to use its AHCI driver. There have been occasional reports of this during SpinRite 6.1's Alpha development and testing. I am thinking therefore that this Samsung NVMe drive, drive 82, may be a candidate for SpinRite 6.1 level 3 (or 4 ) to refresh all the data and return its performance much closer to where it should be.

 

[DanR]

My understanding is that the code in ReadSpeed is a previous incarnation (the code used to build and debug) of the code in SpinRite.

Right!

ReadSpeed was the first product of the ATA/IDE (2013) and AHCI (2020) driver development work. This code was then the initial code for the Pre-Alpha development of SpinRite 6.1. That Pre-Alpha development did extensive work developing and improving drive detection and emulation code logic for SR 6.1. As a result, SpinRite 6.1 now has superior versions of these drivers, compared to the early versions in ReadSpeed.

Accordingly, I believe if the drive is exposed by the BIOS to DOS, it should be able to be tested by ReadSpeed. I believe not all BIOSes will expose a NVMe drive if it wasn't booted from.

Right again!

In the OP's case, the BIOS must be presenting the NVMe drive as a SATA drive, which allows ReadSpeed to test it. Otherwise ReadSpeed would see it as a BIOS drive but would not be able to test it.

 

[AlanD]

From my research, Drive 82 is not NVMe, it is an m.2 SATA drive. That would be why is it being seen by ReadSpeed.

"Description: Samsung SM951 M.2 format 512GB SSD with the AHCI interface, MZHPV512HDGL."

Why it does not see Drive 83 is a different question

 

[Scott]

@talep can you swap the Kingston with the Seagate and see if it’s accessible from that SATA port?

 

[talep]

Sure I will try it later today

 

[talep]

Sorry for the late response , I swapped the cables and it worked now read speed could read the KINGSTON drive, but than when i wanted to boot to linux it would not boot.
Returning back the cables did not help.....to make a long story short something went wrong with the GRUB , I suspect since it was configured with dual boot (the 2TB drive is a windows)....I fixed the grub and there is no real need for Dual boot if your other operating system is on a different drive.
Also I replaced one SATA cable which was not holding so well and now it all works 10x for the help.

 

[PHolder]

Also I replaced one SATA cable which was not holding so well

So many issues can be traced down to marginal cables and connections. Glad you you got it sorted!