SSD Write Endurance 25nm Vs 34nm

Second update:


OCZ Agility 3 AGT3-25SAT3-120G - Both drives are throwing failures but drives are still functional
Corsair Force CSSD-F120GB2-BRKT - a couple minor sector failures on one of the drives.
OCZ Vertex 2 OCZSSD2-2VTXE60G - Nothing yet to report
Corsair Performance 3 Series CSSD-P3128GB2-BRKT - One drive failed completely July 26 @ 11:43am and second drive is throwing failures but has yet to fail.
Crucial RealSSD C300 CTFDDAC064MAG-1G1 - a couple minor sector failures on one of the drives.
SAMSUNG 470 Series MZ-5PA128/US - a couple minor sector failures on one of the drives.
Intel 510 Series (Elm Crest) SSDSC2MH120A2K5 - One drive failed completely July 24 @ 9:14am and second drive failed July 26 @ 3:37pm
Intel X25-M SSDSA2MH160G2K5 - a couple minor sector failures on one of the drives.
Kingston SSDNow V+ Series SNVP325-S2B/128GB - One drive failed completely July 26 @ 2:14pm and second drive is throwing failures but has yet to fail.


-------- End of line ----------

Quote:

---

Originally Posted by johnw

How can you not realize that there are people reading your posts who have actually written linux kernel code and kernel modules? It is quite obvious that you have no clue how to write a kernel module.

---

thank you for your words, be what they may.

From what I've seen, it seems that the rest of the drives will all die this month.

Quote:

---

From what I've seen, it seems that the rest of the drives will all die this month.

---

trying to stay out of this, but cant.

Your "testing" follows no standard that has been followed in this thread. You arent participating in any form with the testing in this thread.

this is a distraction from what is being done here. Your results arent matching the carefully observed and recorded results of others in this thread.
The fact that there is *supposedly* no throttling on your SF devices points to something being wrong/misrepresented with your results.

You are muddying the waters for those reading, with little to NO hard data.

Quote:

---

2) Intel drives received 3% less sunlight than the other drives

---

statements such as this are inane. i think it is in jest? dont know your motivation there. I dont feel it is being said in a cooperative manner.

I encourage you to continue your testing, but I also feel that you should start your own thread, and i also *feel* that other active members of this subsection feel the same.

Quote:

---

Originally Posted by Computurd

trying to stay out of this, but cant.

Your "testing" follows no standard that has been followed in this thread. You arent participating in any form with the testing in this thread.

this is a distraction from what is being done here. Your results arent matching the carefully observed and recorded results of others in this thread.
The fact that there is *supposedly* no throttling on your SF devices points to something being wrong/misrepresented with your results.

You are muddying the waters for those reading, with little to NO hard data.



statements such as this are inane. i think it is in jest? dont know your motivation there. I dont feel it is being said in a cooperative manner.

I encourage you to continue your testing, but I also feel that you should start your own thread, and i also *feel* that other active members of this subsection feel the same.

---

Thank you for your calm perspective and I agree, this thread and my experiment despite similar goals are not related enough to share the same thread.

Quote:

---

Originally Posted by Vapor

Prototype chart, thoughts?

---

Nice to see this, especially as writes are far exceeding MWI. Would something like this not be more easy to visualise?

Attachment 118245

Quote:

---

Originally Posted by bluestang

@Anvil...How's the app looking for release?

---

It's still in beta but can be downloaded from the following link (TheSSDReview)

--

190.33TB Host writes
Reallocated sectors 6

MD5 OK, no change in avg speed.

M4 update:

162.3427 TiB
544 hours
Avg speed 88.85 MiB/s.
AD gone from 10 to 6.
P/E 2833.
MD5 OK.

Attachment 118261

Quote:

---

Originally Posted by Computurd

I encourage you to continue your testing, but I also feel that you should start your own thread, and i also *feel* that other active members of this subsection feel the same.

---

If he is actually doing any testing. There are at least three anomalies in what he has posted that suggest he is making it up:

1) He has not posted any SMART attributes despite repeated requests to do so

2) He claims to be writing continuously to the SSDs at 50MB/s, despite the fact that we know that Sandforce drives throttle much slower than that, and he seems unaware of Sandforce warranty throttling.

3) He has posted code that he claims is everything for a linux kernel module that obviously cannot be a kernel module without additional code

Quote:

---

2) Intel drives received 3% less sunlight than the other drives

---

That really made my day and the fact that he never encountered the SF throttling and the code is incomplete ( who writes a kernel module for this kind of stuff ??? ) just proves that he either is taking us as fools or really suffers from a condition that makes him overcomplicate stuff.

Besides, you never post any SMART data and just say "drive is near failing or has failed" which is obviously not good enough in this situation.

Please take this farce of a test somewhere else or consider solving some of the issues above and posting all of your data before saying "drive has failed" or similar statements.

This really looks like a company sponsored "independent" testing that makes their product look best. They never show facts they just tell us "our product is gazillion times better" and no proof / data whatsoever. Thanks but no thanks.

Should have used UV block.

Suffer the consequences without proper protection! :rofl:

Quote:

---

Originally Posted by Vapor

It burned through its MWI SMART attribute very quickly through an apparent write amplification of ~5x (which is why it looks 'bad' in many charts), but it's still going strong having written its formatted size (64GB) over 4000 times (enough for writing 64GB/59.6GiB per day for ~11 years) :up:

---

Thats good to know. I wonder, is it possible the MWI attribute isnt calibrated/programed properly or something like that for the Samsung drive? Which is why the number ran out so fast, yet the NAND appears to be perfectly fine and going strong?

Quote:

---

Originally Posted by Ao1

Nice to see this, especially as writes are far exceeding MWI. Would something like this not be more easy to visualise?

---

I've been playing with it awhile, and this is as good as I can make it and I don't personally like it....but if others do I'll start including it :)

Attachment 118273

Vapor

It is easy to read the results. +1 :up:

Quote:

---

Originally Posted by Vapor

I've been playing with it awhile, and this is as good as I can make it and I don't personally like it....but if others do I'll start including it :)

---

I like it better than the line chart of the same data.

Alright, new chart is in then :) I'll figure out a way to fix the parts I don't like :p:

Yeah Vapor I like the new chart...even I can read it and understand it!

I too like the new chart as it is much easier to read. The other chart made the Samsung look like a POS, in this new chart it shows that the MWI does indeed wear out faster on the top drives, yet they are still able to perform successful writes even after the SMART attribute is done.

Something interesting I noticed on that Samsung on top. You are testing a 60GB drive, and the MWI lasted pretty much right at 60GB. Coincidence?


According to that chart the drives look like they last significantly past the MWI value. I havent read every single post in this thread but I have been skimming it since the beginning so IDK if this has been asked yet or not. But with your program, are you actually checking to make sure the write that is done is able to be read back successfully? Or is it just writing data without verifying all the data is written properly?

Yes its being taken care of by Anvils app.

As the chart is above....it only replaces "Host Writes So Far" because expected MWI Exhaustion is not included. New chart replaces both "MWI Exhaustion" bar graph and "Host Writes So Far" bar graph :up: (and it has the spacing fixed, which was part of what was bugging me before)

Attachment 118283

While this is more complicated (3 sets of data per bar), I think it's still readable as there's never more than 2 active sets of data per SSD.

Great chart Vapor :up:

I wondered what was bugging you with the previous chart, but this one is just great!

C300 Update

112.775TiB, 1905 raw wear indicator, 62 MWI, 62MiB/sec, MD5 OK, reallocated still at 1 event/2048 sectors.

Sweet chart Vapor, huge improvement over the other graph. Very easy to read.

i liked the other one:mad:

LOL
both are fine :)

Quote:

---

Originally Posted by johnw

247.432 TiB, 678 hours, sa177: 1/1/20345

---

257.598 TiB, 705 hours, sa177: 1/1/21150

Average speed reported by Anvil's app has been steady at about 112MB/s.

The other two unknown SMART attributes, 178 and 235, are still at 72/72/276 and 99/99/2, just as they were when the SSD was fresh out of the box.

64GB Samsung 470.

Thanks for the graph updates Vapor !

The new charts looks great Vapor :up:

M4 update. Almost there now. I think we will hit 1% during the day.

167.3673 TiB
561 hours
Avg speed 88.77 MiB/s.
AD gone from 6 to 3.
P/E 2922.
MD5 OK.

Attachment 118291

I really have no business poking my nose in here, but I've been lurking on this thread since the start, and I must say it's fascinating to follow. I hope all you guys get the recognition you deserve for your efforts. :clap:

Question: What's the relationship between MWI exhaustion and PE cycles? When a drive's MWI is exhausted, does that mean the specified number of PE cycles for that type of NAND have been used? Is this a certainty or an assumption?

Cheers! :)

If you look at the charts you should be able to spot the relationship between MWI and P/E count, it's pretty close to the typical 3K or 5K cycles for 25nm and 34nm NAND.
MWI looks to be set as a "worst case" scenario and the NAND might stay alive for a long time having exhausted MWI.

--

192.99TB Host writes
Reallocated sector count : 6

MD5 OK, 33.77MiB/s on avg (34.5 hours)

The really interesting bit will come when one of these drives fails and reports things like MD5 errors etc.

Quote:

---

Originally Posted by johnw

257.598 TiB, 705 hours, sa177: 1/1/21150

---

265.419 TiB, 726 hours, sa177: 1/1/21827

Average speed reported by Anvil's app has been steady at about 112MB/s.

The other two unknown SMART attributes, 178 and 235, are still at 72/72/276 and 99/99/2, just as they were when the SSD was fresh out of the box.

64GB Samsung 470.

WOW it really is incredible seeing the Samsung go on for so long !

Yeah, the Samsung is a real surprise.

So, just think about what the 128GB drive could be capable of doing.
I've been using my 128GB lately for benchmarks (and OS as well) and there are no signs of degrading performance.

And finaly are we there: 1% is no passed and the fun can begin :D

172.3671 TiB
577 hours
Avg speed 88.76 MiB/s.
AD gone from 3 to 1.
P/E 3010.
MD5 OK.

Attachment 118308

Attachment 118309

C300 Update

118.378TiB, 60 MWI, 2000 raw wear, MD5 OK, 61.95MiB/sec, reallocation still at 2048/1.

Now I'm the only one with drives that haven't exhausted their MWIs :lol:

Cool :) and B.A.T is still at 0 reallocated sectors.

Quote:

---

Originally Posted by bluestang

Any interest in testing a modified Crucial M225 64GB???
I have one that had about 3 months of use before I pulled it from my home system (was with another one in RAID0). It had a couple hundred GB of use on it I believe (definitely under 1TB and also was at 98-99% life) when pulled. I then flashed it to a Vertex Turbo. Why?, cuz I wanted to know if it could be done and why not? The flash was destructive and reset everything and also no longer say Crucial M225.

It's been sitting around for about 6-7 weeks since the OCZ flash. I'm not as savy as you guys so I'm not sure it has enough smart data. Plus it might burn up too soon since the controller and NAND are running at higher than rated speed. But, I'd be willing to toss it on my work PC as a spare drive and let it run. The problem I see having is that I'm running XP, so NO TRIM, only GC to rely on or the Wiper Tool.

Don't forget that even though it says OCZ Vertex Turbo it is a Crucial M225. Let me know what you guys think.

EDIT: Controller is Indilinx Barefoot IDX110M00-LC, Cache is Elpida 64MG SDRAM S51321CBH-6DTT-F and NAND is Samsung K9LBG08U0M-PCB0.

---

Started running Avil's app today @ 12:00 GMT-5 on this drive.
You guys want me to post daily results in here?

Any results would be welcome! I am eager to find out how an older generation SSD stands against the new ones.

Sure, just post your results here!

Will do tomorrow!

Got some baseline info so I can things started in the spreadsheet? :)

Quote:

---

Originally Posted by Vapor

Got some baseline info so I can things started in the spreadsheet? :)

---

Sure...

http://img198.imageshack.us/img198/4...texturbo07.jpg

http://img818.imageshack.us/img818/7...zvertextur.png

Alright, I'll track: C7, Host writes; D1, Remaining Drive Life (equiv. MWI); D0, Average Erase Count (equiv. raw wear indicator); CF, Maximum Erase Count; C4, Erase Failure Block Count (seems to be indicative of reallocated sectors on my drive, hopefully it is with yours). :)

I'll call it "M225->Vertex Turbo 64GB" and will need the average speed reported from Anvil's App after a few hours and then you should be all set....for future reports you can either do a screenshot or post the values that are tracked (probably best to do a screenshot initially in case I'm wrong about C4 and if the other values move in interesting ways :)).

You see it really is a balancing act. The old SSDs had old algorithms and old 50nm NAND with loads of cycles to burn through while the new SSDs have better algorithms and less endurant 25nm NAND. One would expect to see both these types last the same because of the compromises made in each case etc. Imagine a SSD with 50nm NAND and a new modern controller !

M4 update:

175.3810 TiB
587 hours
Avg speed 88.74 MiB/s.
AD gone from 0 to 254.
P/E 3063.
MD5 OK.

Strange development really. AD as gone from 0 to 254. CA has gone from 1 to 254 and CDI reports good health status. Mayby the M4 is hidding something from me.

Attachment 118331

BAT, unexpected indeed.

Could you change back to Hex and post a screenshot, it might just be an issue with CDI.

CA looks correct though, 102% indicates that you are on overtime, showing "Good" is a bug though :)

Quote:

---

Originally Posted by B.A.T

Strange development really. AD as gone from 0 to 254. CA has gone from 1 to 254 and CDI reports good health status. Mayby the M4 is hidding something from me.

---

Well, you behaved like a nice customer with alot of needs from writing perspective, so probably it decided to give you credit for a few more months to make you happy :)

Ontopic: If there is no displaying issue, then is probably just a nice bug regarded to the way this is decremented (a fanatic OOP programmer who just hate "if" statements).

196.04TB Host writes
Reallocated sector count : 6

MD5 OK, 34.92MiB/s on avg (25.25 hours)

Quote:

---

Originally Posted by Anvil

BAT, unexpected indeed.

Could you change back to Hex and post a screenshot, it might just be an issue with CDI.

CA looks correct though, 102% indicates that you are on overtime, showing "Good" is a bug though :)

---

Here is a screenshot with hex
Attachment 118338

Quote:

---

Originally Posted by sergiu

Well, you behaved like a nice customer with alot of needs from writing perspective, so probably it decided to give you credit for a few more months to make you happy :)

Ontopic: If there is no displaying issue, then is probably just a nice bug regarded to the way this is decremented (a fanatic OOP programmer who just hate "if" statements).

---

Always nice to get more for free :p:

There has been no other indications of something wrong with the m4 you guys are probably right.

253.5TB. 33 reallocated sectors. That value which I believe to be the drive wear is 126.

C300 Update

121.79TiB, 59 MWI, 2058 raw wear, 2048/1 reallocation, 61.95MiB/sec (116hrs average), MD5 OK.

M225->Vertex Turbo 64GB Update:

Sorry it took so long to update, I fighting a redirect/advert trojan I can't seem to get rid of :mad:
Had to stop Anvil's app for about 4 hours this morning, Restarted and running now.

2.81TiB, MWI 98, 111 Raw Wear, 37.99MiB/s

I feel lack of TRIM is killing it.

Attachment 118349

Looks like I would need to stop every 5 loops to run wiper tool to get some good avg speeds, but wouldn't that hurt the drive wear as well???

Quote:

---

Originally Posted by johnw

Nice! Thanks for posting the pictures.

So, 40GB, 80GB, and 120GB Intel 320 models all have 8GiB of flash for parity. The 160GB model has 16GiB, 300GB has 20GiB, 600GB has 40GiB for parity.

It is also interesting that the only 320 models that have flash on the back of the circuit board are the 160GB and 600GB models.

---

Hello tech guys. I am happy to find so many knowledgeable people in the same place.

I have two queries. it would be nice to have your replies on these matters.
1. Query of purpose: I have an Asus x51H notebook with 80GB SATA HDD from Hitachi. I do not know what kind of SATA port is that; I rpesume SATA 2. Now, I want to replace that with an Intel SSD 120GB or 160GB (depends on price lowering in the next two months). As this model of llaptop is very rarely covered by online resources, i am having a hard time figuring out what i can add and what i can't. i really need to know if these SSDs are compatible with my notebook.
** I use Win XP only. i won't switch to Win 7 even if someone pays me 100 bucks.

2. Query of curiosity: since i was a kid, whenever i see some empty place for a microchip on a circuit board, i always wonder if i solder one of those chips in the empty spaces, will the capacity go up? like in the case of the 80GB SSD board. just curious.

Quote:

---

Originally Posted by bluestang

M225->Vertex Turbo 64GB Update:

Sorry it took so long to update, I fighting a redirect/advert trojan I can't seem to get rid of :mad:
Had to stop Anvil's app for about 4 hours this morning, Restarted and running now.

2.81TiB, MWI 98, 111 Raw Wear, 37.99MiB/s

I feel lack of TRIM is killing it.

Attachment 118349

Looks like I would need to stop every 5 loops to run wiper tool to get some good avg speeds, but wouldn't that hurt the drive wear as well???

---

I'd err on the side of not doing anything, but I guess do what you would do with it in normal use, if you can.

I don't know how wiper works...is it TRIM retrofitted for Indilinx?

Forgot to ask this earlier: what do you have in terms of static data? johnw, B.A.T, and I all have ~39GiB of static data on our 64GB drives and leave 12GiB free via Anvil's app. You can use Anvil's app to make the static data by killing it from Task Manager when the counter gets up to 39GiB, rename the TEST folder, and then restart the app like nothing ever happened.

32.8GB static, 26.7GB free, and 12GB left free via Anvi's app.

Wiper is an Indilinx Tool for user initiated Garbage Collection.

IMHO for this test it is best to use Windows 7 or Linux as they support TRIM and the 1916 or similar latest Indilinx firmware otherwise the test would be pretty much unrealistic and meaningless to most people. I expect the drive to last less than 10 TB because the Indilinx drives have an older algo which has high write amplification.

Quote:

---

Originally Posted by bluestang

32.8GB static, 26.7GB free, and 12GB left free via Anvi's app.

Wiper is an Indilinx Tool for user initiated Garbage Collection.

---

Looked into Wiper and it is just a retrofitted and manually activated TRIM. If you feel you would run it every TB or so in real life, then do that :)

32.8GB static is close enough to what the rest of us with 64GB are doing, and a far cry from an empty drive.

Quote:

---

Originally Posted by bulanula

IMHO for this test it is best to use Windows 7 or Linux as they support TRIM and the 1916 or similar latest Indilinx firmware otherwise the test would be pretty much unrealistic and meaningless to most people. I expect the drive to last less than 10 TB because the Indilinx drives have an older algo which has high write amplification.

---

WA to date is only ~2.4x on the Indilinx, but it is very early to declare it so low (at an even later point in the test, the Samsung's was 'only' 3.5x). At the current rate, the M225->Vertex Turbo is slated to last 144TiB. While a 50x WA would be interesting to see, doesn't look like we will with the M225->Vertex Turbo.

Charts next post....new color scheme out of necessity, charts look very different now (sort of ran out of colors).

Charts Updated :)

Like I said, new color scheme. Intels are blue, Crucials are red, SandForces are green; newer/bigger within each color gets darker shades. Purple/aqua/orange are for Samsung 470, M225->Vertex Turbo, and whatever gets added later; each color can also be split for multiple shades later if needed.

Host Writes So Far

Attachment 118365

Attachment 118366


Normalized Writes So Far
The SSDs are not all the same size, these charts normalize for available NAND capacity.

Attachment 118367

Attachment 118368


Write Days So Far
Not all SSDs write at the same speed, these charts factor out write speeds and look at endurance as a function of time.

Attachment 118369

Attachment 118370


Host Writes vs. NAND Writes and Write Amplication
Based on reported or calculated NAND cycles from wear SMART values divided by total writes.

Attachment 118371

Attachment 118372

Attachment 118373

M4 update:

181.5208 TiB
607 hours
Avg speed 88.74 MiB/s.
AD gone from 254 to 251.
P/E 3171.
MD5 OK.
Still no reallocated sectors

Attachment 118374

Double C300 Update

Last night just before I went to bed:
124.23TiB, 58 MWI, 2100 raw wear, 2048/1 reallocation, 61.9MiB/sec, MD5 OK

Just now:
127.213TiB, 57 MWI, 2150 raw wear, 2048/1 reallocation, 61.95MiB/sec, MD5 OK


Got the SF-1200 60GB with no LTT in my hands today :D Will post the intake screenshot and info then it's onto some compression tests then when that's done it's back to the endurance testing here :)

SF-1200 nLTT 60GB intake benchmarks and SMART:

SMART:
Attachment 118390

0-fill Anvil Benchmark:
Attachment 118391

46% Comp. Anvil Benchmark:
Attachment 118392

101% Comp. Anvil Benchmark:
Attachment 118393

AS-SSD:
Attachment 118394



Will be running with 35.9GiB of static data (from the 46% compression setting) and 12GB free. Not sure if I'll use 46%, 67%, or a no dedup variety of those for the actual testing.

Onto compression testing now, though :)

Also of note, I lost the extra (F2-F8) SMART values from the C300 when I started back up....I must have done some debug sequence to get them before :wth:

NO M225->Vertex Turbo 64GB update till Monday :(

Lost my connection to my Work PC from my home while stopping app and running wiper and screenshots :mad:

Can't get it to reconnecto and won't be able to get it back up and running till then :down:

@Vapor

:up:

I like the custom drivename :)

One of the weakest 4K QD1 results I've seen on a SF1 series drive, I've got one that hovers around 14MB/s, must be 25nm.

Did you check what NAND and the number of dies?

Hynix 32nm :shakes:

16 of the little buggers.

Quote:

---

Originally Posted by bluestang

NO M225->Vertex Turbo 64GB update till Monday :(

Lost my connection to my Work PC from my home while stopping app and running wiper and screenshots :mad:

Can't get it to reconnecto and won't be able to get it back up and running till then :down:

---

I might be able to add the wiper process to the loop, maybe an option to let it run every n loops.

You should let it run for a day or two just to see how bad it gets.

Quote:

---

Originally Posted by Vapor

Hynix 32nm :shakes:
16 of the little buggers.

---

I find that quite interesting, this will be the first Hynix NAND to be tested, 32nm could go both ways.

--

199.82TB Host writes
Reallocated sector count : 6

I find it interesting too, I just don't think highly of Hynix NAND. Slow, rated for 3000 P/E cycles despite being 34nm-class, very inexpensive...when I think of Hynix NAND, I just don't think of quality.

While I have a hunch the Intel 320 and SF-1200/2200 will be the strongest for endurance (both have parity, SF also can achieve <1.00x WA and SF also has LTT, helping it last longer in terms of days at the expense of performance), but I wouldn't be surprised if this drive were the fastest to die. Maybe further down the road I'll enter an nLTT F60A (IMFT 25nm) or even find an F60 (IMFT 34nm) to enter, but only after one of the two drives I'm running dies (essentially out of SATA ports on the system used for this).

Quote:

---

Originally Posted by johnw

265.419 TiB, 726 hours, sa177: 1/1/21827

---

:eek: MOVEMENT in attribute 178 raw!!!! Changed from 276 to 282. Now adding it to my reporting list:

284.151 TiB, 776 hours, sa177: 1/1/23345, sa178: 72/72/282

Average speed reported by Anvil's app has been steady at about 112MB/s.

The other unknown SMART attribute 235 is still at 99/99/2, just as it was when the SSD was fresh out of the box.

64GB Samsung 470

I wonder if SA178 is reallocated block count? Maybe the SSD shipped with a number of reallocated blocks (from Samsung internal burn-in test?) and only now have there been any new reallocated blocks? I don't know, just guessing.

Right after the above data, I upgraded to the latest version of ASU, Beta4. Before restarting endurance testing, I ran Samsung's SSD Optimization software (Magician), and also ran ASU SSD benchmark and AS-SSD.

Oooo :)

Now is that 6 blocks (equiv) or 1 block (equiv)? Both are multiples of six, it moved by a chunk of six, and the first movement was by a chunk of six, so I'm guessing it's 1 block (equiv) and for some reason the raw value moves in steps of six. Good thing I've been keeping track of SA178 this whole time, don't need to add anything :p:

Quote:

---

Originally Posted by Anvil

I might be able to add the wiper process to the loop, maybe an option to let it run every n loops.

You should let it run for a day or two just to see how bad it gets.

---

Now that would be pretty cool!

Also, back online sooner than expected :)

M225->Vertex Turbo 64GB Update:

5.71TiB, MWI 96, 238 Raw Wear, 37.54MiB/s

Quote:

---

Originally Posted by Vapor

Hynix 32nm :shakes:

16 of the little buggers.

---

Do you know the NAND part number?

Quote:

---

Originally Posted by ao1

do you know the nand part number?

---

h27ubg8t2atr

Attachment 118447

Quote:

---

Originally Posted by johnw

284.151 TiB, 776 hours, sa177: 1/1/23345, sa178: 72/72/282

---

294.054 TiB, 802 hours, sa177: 1/1/24157, sa178: 71/71/290

Average speed reported by Anvil's app has been steady at about 112MB/s.

The other unknown SMART attribute 235 is still at 99/99/2, just as it was when the SSD was fresh out of the box.

64GB Samsung 470

sa178 continues to increase (raw value) and the normalized value went from 72 to 71. Interesting, maybe the flash is finally starting to wear out?

SA178 having a reading of 290 answers the 6 vs. 1 question...it looks like 6 blocks lost and now 14 blocks lost :)

Do you know when it ticked over to 71 normalized? If it happened at 285, then it could be an indication of how the normalized value is calculated and how much NAND failure is 'acceptable' (with an allocation of 1000 blocks to fail and every 10 blocks it moves 1 step).

No, I have only been checking it once a day, so I don't know any more details. Maybe later this week I will try recording the SMART attributes several times in one day, if I get a chance.

m4 update:

194.0943 TiB
648hours
Avg speed 88.70 MiB/s.
AD gone from 251 to 243.
P/E 3394.
MD5 OK.
Still no reallocated sectors

Attachment 118465

C300 Update

133.72TiB, 55MWI, 2260 raw wear, speed is down to 61.6MiB/sec (likely due to the SF-1200's introduction), reallocations still at 2048/1, MD5 OK

SF-1200 nLTT Update

9.688TiB host writes, 3.781TiB NAND writes, 100 MWI, 60.5 raw wear, 62.35MiB/sec @ 46% setting, no reallocations, haven't started MD5 testing yet.

Very good with the no LTT drive. Can the end user modify the firmware and remove LTT is my question :) ?

Quote:

---

Originally Posted by Vapor

h27ubg8t2atr

---

I can't find data to match the last two letters, but other than that here are the specs:

Page size: 8,640 bytes (8,192 + 448)
Block size: 256 pages (2M+112 bytes
Plane size: 1,024 blocks
P/E: 3,000

I'm really looking to see what happens without throttling getting in the way :D

Attachment 118496

204.43TB Host writes
Reallocated sectors 6
MD5 OK

33.00MiB/s on avg for the last 51.7 hours.

M225->Vertex Turbo 64GB Update:

10.31TiB, MWI 92, 421 Raw Wear, 45.30MiB/s for the last 22.1 hours (<-stopped, ran wiper, then restarted)
No MD5 Testing yet.

Also, noticed change in C3 from 0 to 1 and CE from 1 to 214.

Attachment 118499

m4 update:
New milestone passed 200TiB with no reallocated sectors :)

200.3614 TiB
668hours
Avg speed 88.71 MiB/s.
AD gone from 243 to 240.
P/E 3505.
MD5 OK.
Still no reallocated sectors

Attachment 118535

Very impressive :)

Also of note, your WA has been creeping upward for a long time now. At 20TiB, it was 1.0626x; at 40TiB, it was 1.0727x; at 91TiB, it was 1.078x; at 130TiB, it was 1.089x; at 200TiB, it's 1.0933x. Before 20TiB (when calculation is less precise), it climbed from 1.028x to 1.0626x pretty smoothly and quickly....it's been a very smooth climb the entire time actually.

Quote:

---

Originally Posted by johnw

294.054 TiB, 802 hours, sa177: 1/1/24157, sa178: 71/71/290

---

302.868 TiB, 826 hours, sa177: 1/1/24873, sa178: 71/71/290

Average speed reported by Anvil's app has been steady at about 113MB/s.

The other unknown SMART attribute 235 is still at 99/99/2, just as it was when the SSD was fresh out of the box.

64GB Samsung 470

300TiB broken :D

Quote:

---

Originally Posted by Vapor

Very impressive :)

Also of note, your WA has been creeping upward for a long time now. At 20TiB, it was 1.0626x; at 40TiB, it was 1.0727x; at 91TiB, it was 1.078x; at 130TiB, it was 1.089x; at 200TiB, it's 1.0933x. Before 20TiB (when calculation is less precise), it climbed from 1.028x to 1.0626x pretty smoothly and quickly....it's been a very smooth climb the entire time actually.

---

Interesting observations! Maybe SSD is trying to leverage the wear more aggressively thus doing more static data rotation and increasing writes?

The more I think about it, the more it sounds like sizable chunk of data being unaccounted for in the early lifetime of the drive. If there were 100GB unaccounted for in the beginning, its impact as a percentage would get smaller and smaller as writes increased, basically asymptotic behavior. I haven't done any math to see what size that would be, but it might match up nicely with the attempts of getting the right amount of static data on the drive (or even the 39GB itself).

That might be it. First I ran 1 run of Anvils benchmark, ATTO and AS SSD. Then I copied the static of 39,3GiB and started the endurance test. I don't know how much AS SSD and ATTO writes but Anvils app is supposed to keep the write to an minimum.

Very good to see the 300 TB barrier being broken so quickly by the Samsung !

C300 Update

139.04TiB, 53 MWI, 2350 raw wear, reallocations still at 2048/1, MD5 OK


SF-1200 nLTT Update

13.563TiB Host writes, 7.53TiB NAND writes, 99 MWI, 120.5 raw wear, wear range delta 1.

Interesting.

The V2 (34nm) took 11.60TiB before the MWI dropped to 99%. Out of that 11.6TiB approx 4TB was 0fill.
The V3 (25nm) took 11,078 GiB (NAND writes) before the MWI dropped to 99%.

207TB Host writes
Reallocated sectors 6
MD5 OK

34.22MiB/s on avg for the last 21 hours.

M225->Vertex Turbo 64GB Update:

14.14TiB, MWI 89, 557 Raw Wear, 49.70MiB/s for the last 21.25 hours (<-stopped, ran wiper, then restarted)
No MD5 Testing yet.

Also, C3 from 1 to 2 and CE from 214 to 235.

Attachment 118571

EDIT: When should I start MD5 Testing?

Whoa, that is a savage reduction in the MWI considering the drive has 5K P/E. Personally I would not bother with MD5 Testing until the MWI gets very low or unless C3/ CA/ CB start changing significantly.

Looks like the health status is established by D0 (avg erase count) / PE specification count.

I now alot of people notice their drive life drop by 50% or more when they went from Indilinx FW 1916 to FW 2030.
It changed the P/E count from 10k to 5k to cover other NAND I believe.

I think we're good for 10k P/E myself :)

Did you post which type of NAND the SSD is using? (Product ID) If not do you know? If you could let me know I can try to find the product literature. Then we can be sure about the P/E spec.

NAND is Samsung K9LBG08U0M-PCB0

EDIT: Not sure if this helps ya... http://www.datasheetarchive.com/inde...DSA0038979.pdf

Looks like it is 51nm and at some stage the P/E cycle specification was changed from 5k down to TBD

Attachment 118572

m4 update:
Getting close to Anvil now :)
The rigg was down 2 hours last night due to some maintenance.

206.0306 TiB
684hours
Avg speed 91.69 MiB/s.
AD gone from 240 to 236.
P/E 3601.
MD5 OK.
Still no reallocated sectors
Attachment 118573

Quote:

---

Originally Posted by Ao1

Looks like it is 51nm and at some stage the P/E cycle specification was changed from 5k down to TBD

---

I have these notes for some reason in my M255 junk. Don't know where they came from though:

revision history

endurance

April 12th 2006 : 10,000 times

Dec. 22th 2006 : 10,000 times -> 5,000 times

Jan. 12th 2007 : 5,000 times -> TBD

C300 Update

142.01TiB, 52MWI, 2401 raw wear, 2048/1 reallocations, MD5 OK


SF-1200 nLLT 60GB Update

17.188TiB host writes, 10.063TiB NAND writes, 98 MWI, 161 raw wear, wear range delta of 3. Still doing compression tests so still no static data or MD5 testing.

Quote:

---

Originally Posted by johnw

302.868 TiB, 826 hours, sa177: 1/1/24873, sa178: 71/71/290

---

312.510 TiB, 851 hours, sa177: 1/1/25623, sa178: 71/71/292

Average speed reported by Anvil's app has been steady at about 113MB/s.

The other unknown SMART attribute 235 is still at 99/99/2, just as it was when the SSD was fresh out of the box.

64GB Samsung 470

sa178 raw increase of 2 this time