SSD Write Endurance 25nm Vs 34nm

[johnw]

Also, Intrepid 3 is a fantastic name IMHO.

OCZ's sole area of excellence, marketing and hype!

[Anvil]

Endurance_cr_20120425.png

[Anvil]

Samsung 830 256 Day 36
--------------------------------------------------\\
The Samsung 830 finally hit MWI 1 at 1/1/3600 WLC 177.
I'll look through the logs to try and estimate when it hit MWI 1 more precisely.
--------------------------------------------------\\
I'm about to perform some testing on the Octane to see if I can resurect it.

Wish me luck.

Did you find out when it changed or should I just use the value you reported?

.

THE OCTANE LIVES!!

------------------------------------

Okay. The Octane is really, really dead now. It dropped out for the last time, and is now a lifeless hulk.

Looks to behave just like a few of the other drives that "just died", a few disconnects and then game-over in a mattter of days.

[B.A.T]

Todays update:

Intel X25-M G1 80GB
544.50 TiB
22575hours
Reallocated sectors : 123 to 127
Available Reserved space: 92
MWI=163 to 156
MD5 =OK
42.97 MiB/s on avg


m4
759.8775 TiB
3242 hours
Avg speed 73.09 MiB/s.
AD gone from 178 to 175.
P/E 13133.
MD5 OK.
Reallocated sectors : 00


@Anvil
Can you update the chart with my last numbers?

[Anvil]

@B.A.T
Sure

Endurance_cr_20120425_ii.png

--

And so the X25-V finally got past 900TiB

Kingston SSDNow 40GB (X25-V)

900.83TB Host writes
Reallocated sectors : 05 41
Available Reserved Space : E8 99
POH 8085
MD5 OK

33.17MiB/s on avg (~356 hours)

I'm thinking of doing another data retention test, I might wait until it's past 1 year of testing though.

[Christopher]

Anvil,

The 830 exhaused MWI at 3600 with 831,233GiB

---------------------
UPDATE
That's not actually true whatsoever.

It actually hit MWI 1 at what is most likely 3582.

Last log at MWI2: 2/2/3578 at 827,351 GiB
First log at MWI1: 1/1/3582 at 828,317 GiB

So it's probably 1/1/3580 and 827,911 GiB

That's as close as I can get. Somewhere between 3579 and 3582 Wear Leveling Count, MWI hit 1.

The 830 is using 230GiB per WLC, so an increase in WLC of 3 is almost exactly 690GiB. I believe 3580 is actually the value when MWI hit 1, but as I said, I can only narrow it down to a PE cycle count range of 3 (or 690GiB).

[Ao1]

A couple of hours early, but the 830 passed its 3 month data retention test with flying colours. Hmm, what to do now. Leave if for a year?

[Ao1]

--

And so the X25-V finally got past 900TiB

I'm thinking of doing another data retention test, I might wait until it's past 1 year of testing though.

Wait for the petabyte

[sergiu]

A couple of hours early, but the 830 passed its 3 month data retention test with flying colours. Hmm, what to do now. Leave if for a year?

How about another kind of stress test? A continuous read of the files until it dies and watch for the evolution of bad sectors evolution. Now, the cells should be weaker and read disturb effect might have a bigger impact.

[Christopher]

A couple of hours early, but the 830 passed its 3 month data retention test with flying colours. Hmm, what to do now. Leave if for a year?

I actually think you should continue testing the drive as you were before. Do it now!!! And post some recent smart data, too. Letting it sit in a drawer for 365 days isn't going to do the people of Earth much good. For all you know, running that in the dirt will cure cancer. So that's the course I'd advise.

[johnw]

A couple of hours early, but the 830 passed its 3 month data retention test with flying colours. Hmm, what to do now. Leave if for a year?

My vote goes for writing the same amount again, and then leaving it for 3 more months for another retention test.

[Computurd]

dudes check this out
this is from a SSD that the Editor just posted our review for. We had some fun with for the last few weeks. I will say this...man i would LOVE To endurance test this monster, but i dont have the next ten +++ years!
i try not to post links, but this is cool stuff guys
http://thessdreview.com/our-reviews/...lc-ssd-review/

[Christopher]

dudes check this out
this is from a SSD that the Editor just posted our review for. We had some fun with for the last few weeks. I will say this...man i would LOVE To endurance test this monster, but i dont have the next ten +++ years!
i try not to post links, but this is cool stuff guys
http://thessdreview.com/our-reviews/...lc-ssd-review/

That's straight out of Ball-So-Hard University. I heart SLC.

And indestructible to boot. And it has better SMART data (than most drives). But it's probably not cheap whatsoever. Which is unfortunate. One total failure of the 1xxx series is the 64GB increments though. It's 1000x better than nothing, but still nowhere as useful as it's modern siblings's 1GB steps.

I wonder how this drive would fare were it using a 2G SF instead of the 1G -- I bet it would be pretty damn awesome too.

[Computurd]

virtually indestructible lol. i think the case could probably take a pretty decent caliber and deflect, or absorb it. this thing is thick like a nickel as they say

really one of the coolest pieces of tech that Ive been lucky enough to touch, i just cant get over it. i almost had to pry it out with a crowbar that PCB is so secure......im hitting it with a hammer as we speak....


but wait until you see what we post tomorrow....

[canthearu]

im hitting it with a hammer as we speak....

Needs moar hammar :P

But yes, awesome drive. I'd have one in a heartbeat .... screw SATA III.

[Christopher]

SLC + indeeeestructable = win

Imagine a hardened Intel X25-E -- a drive which can survive a direct hit from most small arms fire and has 50nm SLC. I think I should retrofit mine with some kind of metallurgical ninja magic.

It's too bad there isn't more information on the physical limits of "non-ruggedized" SSDs. I have a couple which are built like tanks -- the 3.5" MTRON comes to mind -- but I want one I can run over with a tank. Or install in a tank.

I'd call my own ruggedized SSD The Intrepid or perhaps, the U.S.S. Intrepid.

[Hopalong X]

I'd call my own ruggedized SSD The Intrepid or perhaps, the U.S.S. Intrepid.

If they were OCZ they could name them "Titanic", "Bismark" and "Kamikaze".

Titanic is unsinkable and indestructable as long as it never makes a maiden voyage.
Bismark indestructable and faster than a Vertex Turbo 4 or even an Intel 520. At some point will start unending loops at boot and sink.
Kamikaze does 1500Mb/s. Once.

[Christopher]

You know, the 830 is up to 298MB/s Avg. That's pretty nutty.

So is the fact that in 904 hours of testing it's put down over 910,000GiB. So it can sorta do a PB every 40 days. That's moderately biblical.

[johnw]

If they were OCZ they could name them "Titanic", "Bismark" and "Kamikaze".

[B.A.T]

Todays update:

Intel X25-M G1 80GB
548.51 TiB
22603 hours
Reallocated sectors : 127 to 128
Available Reserved space: 92
MWI=156 to 150
MD5 =OK
43.49 MiB/s on avg


m4
766.6989 TiB
3269 hours
Avg speed 73.10 MiB/s.
AD gone from 175 to 171.
P/E 13247.
MD5 OK.
Reallocated sectors : 00


Only 2 GiB again.....

[Christopher]

Only 2 GiB again.....

Considering it's a refurbished drive, that's pretty damn excellent.

[B.A.T]

I know
I just hope it will last to 1 PiB...

[Christopher]

I know
I just hope it will last to 1 PiB...

Once you escape the surly bonds of MWI, the sky is the limit.

I wouldn't retention test that thing until you get to 2PB, but that's just me

[B.A.T]

I don't dare to try if I want to see those number. Better to leave it alone.

[Christopher]

I don't dare to try if I want to see those number. Better to leave it alone.

Think positive
//--------------------------------------------------\\

Samsung 830 256 Day 39

917464.46 GiB
896.96 TiB

295.38 MB/s Average
911 Hours

3996 Wear Leveling Count
MWI 1

0/0 Erase/Program Fail

OCZ Vertex Turbo 64 Day 57

416,685.08 GiB
406.91 GiB
83.71 MB/s Avg
1408 hours

7451 Avg PE count
-100 MWI

0/0/0 Read/Program/Erase Fail

//--------------------------------------------------\\

The Samsung seems to be doing even better than the 1000GiB/hour rate it was doing for some time. In 24 hours it was able to write 25,080GiB.