Toshiba 19nm NAND SSDs coming in August

hhhd1

Senior member
Apr 8, 2012
667
3
71
we should be waiting for the same from crucial, the m4 was released over a year ago, and they are continuing to decrease the price, probably to clear out inventory. (just guessing, no real info)

I am interested in seeing the endurance of the new 19nm NANDS.
 

razel

Platinum Member
May 14, 2002
2,337
93
101
'Deterministic Zeroing TRIM' that's an 'exclusive with Windows 8' sounds different from 'Queued TRIM command' that's part of SATA 3.1. I'm glad they are making improvements. I'll have to research up on what DZ TRIM is.
 

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
'Deterministic Zeroing TRIM' that's an 'exclusive with Windows 8' sounds different from 'Queued TRIM command' that's part of SATA 3.1. I'm glad they are making improvements. I'll have to research up on what DZ TRIM is.
TRIM makes no guarantee that the data is actually cleared out. It's kind of a big deal for parity-based RAID. The next read could read zeros, all ones, the old data, or any old garbage. Not the brightest idea, really. If it's not supported in Windows 7 directly, I'd guess it's a new TRIM variation. But if you're going to support that, why not go ahead and make all TRIM immediately zero out?
 

taltamir

Lifer
Mar 21, 2004
13,576
6
76
TRIM makes no guarantee that the data is actually cleared out. It's kind of a big deal for parity-based RAID. The next read could read zeros, all ones, the old data, or any old garbage.

But there is absolutely no reason to make a new command for this NOR is there any need for the controller to actually erase the data.
All you need to do is modify (if it doesn't already do that) the drive's firmware return all 0s if an OS tries to read an address whose current status is trimmed. Then you put a sticker on the box that says "RAID Compatible TRIM". Presto, problem solved.
The spec should have been amended to require this years ago. (as in, from the first TRIM drive)

But instead they are trying to milk companies for the "feature" (same way WD limits TLER... TLER is merely disabling a feature called Deep Error Recovery because it freaks out badly designed raid controllers).

The advantage of having it as a unique command means that if you connect a hypothetical older drive that returns garbage data to a new controller that sends TRIM to parity capable RAID then it will cause corrption. If the new controller does NOT issue TRIM commands, instead only issuing dtrim, then it will simply do nothing on such an older drive. No data loss, but no trim either (which is a shame).
The downside is you now need to add a new type of command at every point of the chain and older drives that have TRIM and return all 0s when a trimmed address is queried will not work without a firmware upgrade adding dtrim support (which is highly unlikely to happen).

What irks me (beyond the fact that this was not included in the original trim spec) is that this issue was raised and work began on dtrim early enough that it was simple enough to amend the TRIM specification to include this and then have a VERY short list of drives that were made pre amendment (and whose manufacturers refused to provide a firmware update)... short enough that you could test them all/ask the manufacturers what behavior they have and compile a complete list of under a dozen specific drive models to exclude and put that exclusion list in RAID controllers.
 
Last edited:

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
But there is absolutely no reason to make a new command for this NOR is there any need for the controller to actually erase the data.
I agree. An updated spec with defined TRIM behavior would be preferred. Since the current behavior is undefined, any new behavior would be both forwards and backwards compatible.
 

Hellhammer

AnandTech Emeritus
Apr 25, 2011
701
4
81
AnandTech has something on this drive also - no mention of the new TRIM feature though..

http://www.anandtech.com/show/5912/toshiba-announces-thnsnf-series-ssds-19nm-nand-is-here

As far as new drives go - I found this interesting also..

http://thessdreview.com/latest-buzz...r-for-10-channel-6gbps-ssd-performance-first/

The US press release didn't say anything about TRIM at all, hence I left it out. I'm sure we'll take a look at it once we get a review sample.

Take care,
Kristian Vättö
SSD Reviewer for AnandTech