SSD Write Endurance 25nm Vs 34nm

[Christopher]

Yeah, so the Octane? Already broken. I was just putting some static data on it, and it locked up. I was able to reflash the FW, but to no avail.

I'm more than a little irate. It had like 20GB of writes on it.

It began BSoDing my system with it connected, and cannot be accessed through disk management. In addition, the OCZ Toolbox crashes on drive detection now.

I'll try it in some other systems, but... damn.

UPDATE://
I was able to get the drive working on the SATA to IDE bridge in a second HDD bay in my laptop of all things. But I'm not sticking something sketchy in the e-rig, not if it's gonna crash the whole system. The Vertex Turbo and the 830 are magnificent, and it would be a shame to put something in that will jeopardize that.

[bulanula]

Yeah, so the Octane? Already broken. I was just putting some static data on it, and it locked up. I was able to reflash the FW, but to no avail.

I'm more than a little irate. It had like 20GB of writes on it.

It began BSoDing my system with it connected, and cannot be accessed through disk management. In addition, the OCZ Toolbox crashes on drive detection now.

I'll try it in some other systems, but... damn.

UPDATE://
I was able to get the drive working on the SATA to IDE bridge in a second HDD bay in my laptop of all things. But I'm not sticking something sketchy in the e-rig, not if it's gonna crash the whole system. The Vertex Turbo and the 830 are magnificent, and it would be a shame to put something in that will jeopardize that.

Exactly the thing I meant. I personally never had a HDD just disappear out of the blue like that.

SSD controllers have a long way to go till they mature like HDDs did. Early HDDs were just as bad as these early SSDs.

But IMHO there will be a new tech around the corner because it seems that NAND will not scale very well with decreasing litography sizes etc.

[Christopher]

I'm doing more testing, but there could a problem with my motherboard. It's gonna need more testing, but it is possible that the Octane might be a victim and not the offender. I do have it working in my laptop, and I'm going to do a port-by-port test of my mainboard.

UPDATE
Yup. Somehow -- and I don't know exactly what happened, but it looks like the SATA cable may have been bad on the Octane. Once it stopped working, it could get detected but not function on other ports and/or other cables.

So the Octane is fine, but I don't know what happened for sure, other than some of the files on the system drive became corrupted too.

Bottom line -- the Octane is off the hook. One bad SATA cable seemed to screw everything up.

I'll start the Octane immediately.


UPDATE://
Everything appears to be fine now, but the cause of the issue was not actually a bad SATA cable, but rather that the system drive had become corrupted. I believe I have everything fixed now, and the Octane soldiers on, and is now prepped to enter the endurance rig.

[DragoonXX]

Exactly the thing I meant. I personally never had a HDD just disappear out of the blue like that.

SSD controllers have a long way to go till they mature like HDDs did. Early HDDs were just as bad as these early SSDs.

I've had HDDs entirely DROP their controller out of the blue (hitachi 500GB drive- MAYBE had 1TB worth of writes to it, <500 power on hours- this drive was bought in ~2008)- drive controller failed, was simply recognized as an 8mb drive in windows, would hang a server os, and would corrupt data in a linux environment.
there are random bad drives (and as christopher mentioned) and random bad cables/data corruption (i've had all kinds of fun tracking those down too before). I won't disagree that some controllers need a lot of work, but it mostly breaks down to what components the manufacturer decides to use. I've personally been very impressed with my SSDs (a pair of x-25V drives...which, from my usage patterns, and if mine only last to 1/10 of the kingston here is...will very likely outlive me), and I will continue to use SSDs in my future computers (just...bigger drives...)

[Christopher]

Okay,

A mini update
OCZ Vertex Turbo 64GB
179,517GiB (610 Hours)
84.08 Avg MB/s

Samsung 830 256GB
108,057GiB (In 106 Hours)
295.63 Avg MB/s

OCZ Octane 128GB
7,411GiB (13 Hours)
162.80 Avg MB/s

--------------------------------
The Samsung gets faster, the Turbo still has no reallocations, and the Octane is fine after 13 hours.

I swear -- I have never had a system drive become corrupted before, but that is precisely what happened. Except just about everything was fine. I don't know what the source of the corruption was, but I left for a few minutes, and when I came back the system was off. But when I turned the system back on, Windows didn't about about having been shutdown improperly. I was using the Plextor M3P as the boot drive on my system, and while I've never had this happen before, this could be one of those times when PLP would have saved the day.

[devsk]

Hey Chris, do you wanna give Arowana firmware on the vertex a try? It has much better WA and 4k random speeds.

[Anvil]

Kingston SSDNow 40GB (X25-V)

816.12TB Host writes
Reallocated sectors : 05 32 // 2 up
Available Reserved Space : E8 99
POH 7330
MD5 OK

32.38MiB/s on avg (~308 hours)

--

@canthearu

You need to enable the "keep running totals" option in order to get it refreshed, it does work, I have checked. (and it worked for the ones with issues)

@christopher

Awesome avg speeds on those drives.
The C drive corruption was unexpected and rare, I haven't experienced anything like it except for during the occasional bad OC. (very rare)

[Christopher]

Hey Chris, do you wanna give Arowana firmware on the vertex a try? It has much better WA and 4k random speeds.

Sure, I'd love to try it. But --

A) It's not been released, even after almost a year
B) It will only work on non-Samsung NAND equipped drives (or so they've said)

Meaning, I couldn't use it on the Turbo even if it existed.

Why? Do you know where I can get it?
--------------------------------------------------------------------

Just to be clear, it was not the endurance rig that suffered the issue, but rather my main system. The Endurance Rig has shown itself to be quite stable.

@Anvil -- It took a while for me to discover the root of the issue, and I suspect I'll never know the true cause. But needless to say, I've never seen anything like it.

Those average speeds are marvelous, aren't they? Ironically, the 64GB 830 basically puts the rest of the field to shame -- even the 256GB 830 -- when it comes to avg speed vs. capacity. The Octane is basically 2x the Turbo, so that's a good representation. Having a 64, 128, and 256 in the test is pretty nice.

[Anvil]

Endurance_cr_20120324.png

I have removed Vapor's drives from the chart. (can be easily restored)
(no sign from B.A.T, I expect he's travelling/working)

[Christopher]

The funny thing is, I have almost no doubts about the pedigree of the controller inside the Octane, but it acts almost EXACTLY like the Turbo, just about 2x the speed.

@Ao1 -- I am waiting with bated breath for the results of the "hair dryer test". Please, don't keep me waiting

[Anvil]

@christopher

Could you check and post info on the Octane (static data etc), did you have a look at the internals?

[Christopher]

Intel 29F64G08ACME2 x 16 25nm Intel Synchronous NAND

I'm using 37.9GB Static data.

Test officially started 23-March-2012 at 23:14:25 EST, which is UTC -5.

It has 512MB of cache, but I don't know whether it's DDR2 or DDR3

Overall, I think it's a pretty decent drive. I wish I had another one to play with, but I'm not spending the money. They're not really competitive on price.

[B.A.T]

Todays update:
And I'm back from another worktrip. Sorry for the delay but with only edge mobile coverage it was impossible to download an update with teamviewer.

Intel X25-M G1 80GB
431,9031 TiB
21820 hours
Reallocated sectors : 02 to 04
MWI=91 to 82
MD5 =OK
45.40 MiB/s on avg


m4
571.3704 TiB
2485 hours
Avg speed 74.07 MiB/s.
AD gone from 56 to 24.
P/E 9971.
MD5 OK.
Reallocated sectors : 00


The rig has restartet due to windowsupdate during my trip but no it's running again.

[Christopher]

Thanks for gracing us with your presence, B.A.T.

[Anvil]

Endurance_cr_20120324.png

[B.A.T]

Thanks for gracing us with your presence, B.A.T.

You welcome
I see you been using some more money lately

[Christopher]

You welcome
I see you been using some more money lately

Why would I waste money on food when there are drives in need of testing?

[Christopher]

OCZ Octane 128 Day 1

15543 GiB
163.28 MB/s Average
27 hours

Samsung 830 256 Day 5

121986 GiB
291.91 MB/s Average
120 hours

551 Wear Leveling Count
MWI 84

0/0 Erase/Program Fail

OCZ Vertex Turbo 64 Day 25

183786 GiB
84.14 MB/s Avg
624 hours

3294 Avg PE count
35 MWI

0/0/0 Read/Program/Erase Fail

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

Here is a shot of the Octane, hard at work trying to make me happy.



The Samsung is still super fast, the Vertex is still free of bad blocks, and the Octane is acting... what's the word.....?


Marvilinxidelic?
I'm getting a tattoo of that word I just made up to decribe the Octane... somewhere.

[Christopher]

Look it up. I swear it's a real word.

I will swear up and down on a whole stack of Bibles that the Octane is basically like a M4 on 0001FW. Or, a whole hell of a lot like the Intel 510. Additionally, before the 1.13FW update, it would have had almost identical higher-QD random scores. So if it walks like a duck and quacks like a duck... you'd really have to prove that it isn't a duck. It looks, sounds, and tastes a lot like a duck.

[canthearu]

The octane is doing pretty well. Hopefully ndurance doesn't mean lifetime throttling, otherwise it really will be a pitty.

[canthearu]

Patriot Torqx-2 64GB - Day 38

Drive hours: 910
GiB written: 57,968.00 GiB (56.61 TiB)
Avg MB/s: 20.39 MB/s

Bad blocks: 2/0/83
Wear cycle counter: 0/3623/5515

Intel 520 60GB - Day 29

Drive hours: 711
Avg MB/s: 91.56 MB/s

Host GB written (F1): 217,184.53 GiB (212.09 TiB, 6949905 raw)
NAND writes (F9): 153,796 GiB (150.19 TiB)

Reallocated sectors (05): 0
Failure count (AB, AC): 0 program, 0 erase
Media Wearout Indicator (E9): 52

[bulanula]

Thank you for the chart updates Anvil.

Any guesses as when the X25-V will RIP ?

Seems it likes to go slow and steady to its demise.

[devsk]

Sure, I'd love to try it. But --

A) It's not been released, even after almost a year
B) It will only work on non-Samsung NAND equipped drives (or so they've said)

Meaning, I couldn't use it on the Turbo even if it existed.

Why? Do you know where I can get it?
--------------------------------------------------------------------

Just to be clear, it was not the endurance rig that suffered the issue, but rather my main system. The Endurance Rig has shown itself to be quite stable.

@Anvil -- It took a while for me to discover the root of the issue, and I suspect I'll never know the true cause. But needless to say, I've never seen anything like it.

Those average speeds are marvelous, aren't they? Ironically, the 64GB 830 basically puts the rest of the field to shame -- even the 256GB 830 -- when it comes to avg speed vs. capacity. The Octane is basically 2x the Turbo, so that's a good representation. Having a 64, 128, and 256 in the test is pretty nice.

Its up and I just upgraded my vertex. The seq numbers took a fall but random 4k access numbers are up.

http://www.ocztechnologyforum.com/fo...-Solid2-drives

[Christopher]

Its up and I just upgraded my vertex. The seq numbers took a fall but random 4k access numbers are up.

http://www.ocztechnologyforum.com/fo...-Solid2-drives

So I looked over on the OCZ forums...

It looks like when you move to Arrowana you lose all the tasty Barefoot SMART attributes. What do you get?

Octane's SMART attributes.



One possible explaination: OCZ is not really keen on users having access to SMART attributes anymore, because I find it strange that they'd refuse to expose those attributes through FW on the Octanes and Arrowanas. Perhaps they didn't have a choice with SF FW? Who knows? But this is not a good trend.

[devsk]

So I looked over on the OCZ forums...

It looks like when you move to Arrowana you lose all the tasty Barefoot SMART attributes. What do you get?

Octane's SMART attributes.



One possible explaination: OCZ is not really keen on users having access to SMART attributes anymore, because I find it strange that they'd refuse to expose those attributes through FW on the Octanes and Arrowanas. Perhaps they didn't have a choice with SF FW? Who knows? But this is not a good trend.

Yeah. I don't like that part. Some of OCZ practices are in bad taste.

The only thing they expose is remaining %age life. Which starts at 100 after this flash.... You may have written the drive hard in last 2 years but it starts at 100.