Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB • Hacker News Board

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#31

Post by jotux » Tue, Apr 24, 2018, 6:49 PM UTC

nine_k wrote: post #29

marshray wrote: post #26

Earlier quoted context omitted.

A common minimum voltage requirement for the nominal +5V supply is 4.5 V. Let's assume a voltage below that won't damage the drive, but it won't operate. The diode is going to drop 0.6V, leaving you with only 4.4 V max for the drive. Then the capacitor is going to immediately start dropping in voltage as soon as you start pulling the stored energy out of it. So, in practice, to do a quality job of this it will probab…

This makes sense! OTOH low-drop diodes do exist [1], and the drive likely needs just a (small) fraction of a second to flush the caches; it's SSD, not need to wait for the disk to rotate to the right position or move heads. I suppose the power loss signal from the controller still propagates, even if the computer's PSU plug is just yanked out of socket. [1]: https://electronics.stackexchange.com/a/282471/9909

I implemented a power-loss circuit on a custom USB memory device (battery-backed LPDDR disks) and used something similar: the MAX40200[1].

[1]https://www.maximintegrated.com/en/products/analog/amplifier...

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#32

Post by MrMember » Tue, Apr 24, 2018, 6:50 PM UTC

gautamb0 wrote: post #19

Hoping this reduces used 960 EVO/PRO prices. Not a whopping improvement.

I bought a 500GB 960 EVO on sale for $200 last week. Now I know why it was on sale.

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#33

Post by stefan_ » Tue, Apr 24, 2018, 6:51 PM UTC

whitepoplar wrote: post #17

Do these drives feature power-loss capacitors? I'm still wondering why most manufacturers think it's okay to offer SSDs that lose data on power loss (!!)

So what happens when you have "power-loss capacitors"? Since these things are limited the drive races to empty its buffers onto the SSD and then shuts down. It does not ask the host "are you done writing", so now instead you lose whatever the host had in its buffers. The host is gone anyway when the drive loses power.

No. Data loss protection needs to be built into the software itself, fixing that on the SSD level doesn't make a lot of sense. If anything, you install a UPS, and then of course you don't need to pay extra for SSD capacitors because the whole system keeps running through power loss and shuts down gracefully.

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#34

Post by pkaye » Tue, Apr 24, 2018, 6:54 PM UTC

whitepoplar wrote: post #17

Do these drives feature power-loss capacitors? I'm still wondering why most manufacturers think it's okay to offer SSDs that lose data on power loss (!!)

I've worked on PLP implementations in the past. Mostly it was the added cost of capacitors. Also things like laptops don't really need them since they have batteries and will shut down gracefully when battery is running low.

The capacitors needs to be reliable, small form factor and hold enough charge to keep the controller, DRAM and NAND powered for a minimum of 15-30ms. 10 years ago what was available was not so great but now the options are much better.

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#35

Post by whitepoplar » Tue, Apr 24, 2018, 6:57 PM UTC

stefan_ wrote: post #33

whitepoplar wrote: post #17

Do these drives feature power-loss capacitors? I'm still wondering why most manufacturers think it's okay to offer SSDs that lose data on power loss (!!)

So what happens when you have "power-loss capacitors"? Since these things are limited the drive races to empty its buffers onto the SSD and then shuts down. It does not ask the host "are you done writing", so now instead you lose whatever the host had in its buffers. The host is gone anyway when the drive loses power. No. Data loss protection needs to be built into the software itself, fixing that on the SSD level do…

Power-loss capacitors are needed because without them, flush commands issued by the OS return bogus values. Write confirmations are meaningless if the data isn't actually written to disk.

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#36

Post by whitepoplar » Tue, Apr 24, 2018, 6:59 PM UTC

pkaye wrote: post #34

whitepoplar wrote: post #17

Do these drives feature power-loss capacitors? I'm still wondering why most manufacturers think it's okay to offer SSDs that lose data on power loss (!!)

I've worked on PLP implementations in the past. Mostly it was the added cost of capacitors. Also things like laptops don't really need them since they have batteries and will shut down gracefully when battery is running low. The capacitors needs to be reliable, small form factor and hold enough charge to keep the controller, DRAM and NAND powered for a minimum of 15-30ms. 10 years ago what was available was not so gr…

Given your work with PLP, which family of SSDs would you recommend most?

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#37

Post by jerf » Tue, Apr 24, 2018, 7:01 PM UTC

On an only sort-of related note, has anyone found a USB (or anything else external) adapter that can read these yet?

I know why it's hard, and I know about the ones that read that form factor but not NVMe drives, and why that is. Just curious if my searches are coming up dry because they don't exist or because they're getting flooded by the results that match the search but don't work.

I enjoy the performance of the drive I have but it makes me nervous that the only thing in my house that can read it is the laptop it is in. If push came to shove I could create some options, but the best would be a stand-alone reader, like you can get for basically every other kind of storage tech.

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#38

Post by stefan_ » Tue, Apr 24, 2018, 7:06 PM UTC

whitepoplar wrote: post #35

stefan_ wrote: post #33

Earlier quoted context omitted.

So what happens when you have "power-loss capacitors"? Since these things are limited the drive races to empty its buffers onto the SSD and then shuts down. It does not ask the host "are you done writing", so now instead you lose whatever the host had in its buffers. The host is gone anyway when the drive loses power. No. Data loss protection needs to be built into the software itself, fixing that on the SSD level do…

Power-loss capacitors are needed because without them, flush commands issued by the OS return bogus values. Write confirmations are meaningless if the data isn't actually written to disk.

Doesn't software like databases that care about this already assume the OS flush return values are bogus?

And it still doesn't solve the problem of the host buffers disappearing. What's the scenario, the host experiences a power-loss, but was somehow anticipating it and managed to stop writing to the SSD?

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#39

Post by _wmd » Tue, Apr 24, 2018, 7:13 PM UTC

stefan_ wrote: post #33

whitepoplar wrote: post #17

Do these drives feature power-loss capacitors? I'm still wondering why most manufacturers think it's okay to offer SSDs that lose data on power loss (!!)

So what happens when you have "power-loss capacitors"? Since these things are limited the drive races to empty its buffers onto the SSD and then shuts down. It does not ask the host "are you done writing", so now instead you lose whatever the host had in its buffers. The host is gone anyway when the drive loses power. No. Data loss protection needs to be built into the software itself, fixing that on the SSD level do…

The buffers maintained by the OS and drive are totally different, the entire point of having huge buffers in the device is so that it can agglomerate and reorder otherwise totally unrelated writes into flash banks it is presently convenient to write to.

When the OS flushes stuff to the device, it is done such that metadata writes reach the disk in an order that ensures a recoverable filesystem image on power failure. If the device does not preserve that ordering all bets are off. The mapping that goes on inside a SSD is of sufficient complexity that it's much easier to simply guarantee the device has enough independent power to ensure all acknowledged writes are persisted after acknowledgement, and that means capacitors, or best case, data (e.g. committed, acked DB transactions) that magically disappear during a power loss event.

You cannot have a device that temporarily writes to RAM and tells the host computer "OK that's on disk now" (i.e. every SSD in existence with a throughput exceeding a few MB/sec) be reliable without an independent source of power to ensure that RAM reaches flash in every situation

Re: Samsung's 970 EVO SSDs Reviewed – 500GB and 1TB

#40

Post by 21 » Tue, Apr 24, 2018, 7:14 PM UTC

jerf wrote: post #37

On an only sort-of related note, has anyone found a USB (or anything else external) adapter that can read these yet? I know why it's hard, and I know about the ones that read that form factor but not NVMe drives, and why that is. Just curious if my searches are coming up dry because they don't exist or because they're getting flooded by the results that match the search but don't work. I enjoy the performance of the…

If you have a desktop you can buy a $10 PCI express adaptor for it.