9

I recently bought a SanDisk Ultra 3d SSD and upon testing it never reaches the advertised speed of 560mb/s read, 530mb/s write. Sometimes the read speed is even lower than 400 mb/s, also the write speed is always higher than the read speed. My old 120gb Kingston uv400 reaches 500+mb/s read speed which is the advertised speed.

enter image description here

  • 11
    I've found that having the OS you're booted into installed on the disk you're testing always affects the performance. – At0mic Nov 17 at 17:47
37

SEQ1M maxes at 527MB/s which is pretty close.

Unless the disk is completely empty and not used at all by your system then this is pretty much what you'd expect to see. 550MB/s is the theoretical maximum and assumes that the testing program is the only thing running and that the link is otherwise "perfect".

There is almost always some overhead, either in the device chain between your CPU and the SSD, or in something else using the SSD.

You are within about 5% of the advertised drive speed. There is nothing wrong with the drive. Advertised speeds are "up to" the given speed, not "will always".

Sandisk themselves state the following in the disclosures section of their site

Based on internal testing; performance may be lower depending on drive capacity, host device, OS and application

Unless you can replicate their exact hardware, test procedure, OS and so on then "up to" is about as good as you're going to get.

6

You will need to install it on a Linux system as a second drive, make sure it is totally empty and formatted as EXT4, then re-run your test using the "dd" command in Linux.

Under those conditions you should find the drive is actually slightly faster than the stated performance numbers from Sandisk, which get rounded down to the nearest 10MB.

The Windows NTFS filesystem adds a little overhead, as does using the drive as a boot drive, which it looks like is what you're doing. Windows will constantly read and write tiny amounts from the drive it boots from while you are using it. Both of these factors will reduce the numbers in your test slightly.

  • 8
    I don't see the need for ext4 - if anything, it adds overhead. Just use dd on the disk as a raw block device. – Radovan Garabík Nov 18 at 14:22
  • 1
    Sure, if you want. It makes things a lot simpler if you do have a filesystem on the drive, and the results are about the same. – Geoff Griswald Nov 18 at 14:24
4

The advertised speed is displayed as up to 500mb/s. There could be multiple reasons why it's slower then the advertised speed like

  • need to update the firmware
  • AHCI mode is disabled
  • wrong SATA port, you need to use the SATA 3 port(if it's below 300mb/s)
  • optimize the drivers
  • confirm "trim" is running

You can find other things you can try here

If this doesn't bump the speed up, you can contact the manufacturer but they probably won't replace it due to this "up to" rule.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.