SSD Write Endurance 25nm Vs 34nm

[my87csx481]

I've been away for quite some time. Read through this thread probably 8-12 months ago. No time right now to go back through it.

Just curious what you guys think is the best general usage drive for under $200 US?

The drive is for a friend. I was looking at Intel 520 cherryville, Samsung 830 or Samsung 840 EVO.

Comments? Suggestions? Any input is appreciated.

Thanks in advance.

[Zaxx]

Sammy 840 Pro if you can settle for 128GB.

[DragoonXX]

Seems to be a shame that this thread looks like its dead. I wonder what ever happened to that poor little trooper X-25

[lucassp]

http://techreport.com/review/25559/t...t-200tb-update

Endurance experiment featuring some newer SSDs. Samsung doesn't look very good there.

[devsk]

Is anybody still having Crucial M4 running this test? Please respond. There is a way to get the device stats from the firmware using smartctl and we can compare that to what we wrote to the disk to see if the erase cycles and MWI exhausted numbers we came up with make sense.

The following are the stats for my crucial m4 256GB running for last 2.5 years in my desktop/home server.

# smartctl -l devstat /dev/sda
smartctl 6.2 2013-07-26 r3841 [x86_64-linux-3.12.5] (local build)
Copyright (C) 2002-13, Bruce Allen, Christian Franke, www.smartmontools.org

Device Statistics (GP Log 0x04)
Page Offset Size         Value  Description
  1  =====  =                =  == General Statistics (rev 2) ==
  1  0x008  4              153  Lifetime Power-On Resets
  1  0x010  4            20579  Power-on Hours
  1  0x018  6      13725052826  Logical Sectors Written
  1  0x020  6        322531254  Number of Write Commands
  1  0x028  6      24940390620  Logical Sectors Read
  1  0x030  6        459425614  Number of Read Commands
  4  =====  =                =  == General Errors Statistics (rev 1) ==
  4  0x008  4                0  Number of Reported Uncorrectable Errors
  4  0x010  4                0  Resets Between Cmd Acceptance and Completion
  5  =====  =                =  == Temperature Statistics (rev 1) ==
  5  0x008  1                0  Current Temperature
  5  0x010  1                0  Average Short Term Temperature
  5  0x018  1                0  Average Long Term Temperature
  5  0x020  1                0  Highest Temperature
  5  0x028  1                0  Lowest Temperature
  5  0x030  1                0  Highest Average Short Term Temperature
  5  0x038  1                0  Lowest Average Short Term Temperature
  5  0x040  1                0  Highest Average Long Term Temperature
  5  0x048  1                0  Lowest Average Long Term Temperature
  5  0x050  4                -  Time in Over-Temperature
  5  0x058  1               70  Specified Maximum Operating Temperature
  5  0x060  4                -  Time in Under-Temperature
  5  0x068  1                0  Specified Minimum Operating Temperature
  6  =====  =                =  == Transport Statistics (rev 1) ==
  6  0x008  4                0  Number of Hardware Resets
  6  0x010  4                0  Number of ASR Events
  6  0x018  4              365  Number of Interface CRC Errors
  7  =====  =                =  == Solid State Device Statistics (rev 1) ==
  7  0x008  1               17~ Percentage Used Endurance Indicator
                              |_ ~ normalized value

I am very interested in these numbers for actual drive used in the endurance test... Thanks a bunch!

[canthearu]

I can understand "smartctl -a /dev/sda" better.

-a will give unmolested statistics from the drive's smart function.

Having software molest the data tends to lose some of the information.

[devsk]

I can understand "smartctl -a /dev/sda" better.

-a will give unmolested statistics from the drive's smart function.

Having software molest the data tends to lose some of the information.

I think you misunderstood. These new stats are part of the standard now and are directly exported by the firmware. The SMART stats attributes are a mess. They were not there in the standard and are generally used incoherently by vendors (seen those Unknown attributes?). ACS-2 came up with SSD Device Statistics standard and they are now available through Device Statistics (GP Log 0x04) from the SSD in the form of various pages. Crucial is one of the few vendors who are correctly filling in those stats. You won't find these stats from SMART attributes. For example, Page 1 (shown below) is missing is completely in SMART attributes on an M4. Kudos to Crucial for being ACS-2 compliant...

# smartctl -l devstat,1 /dev/sda
smartctl 6.2 2013-07-26 r3841 [x86_64-linux-3.12.5] (local build)
Copyright (C) 2002-13, Bruce Allen, Christian Franke, www.smartmontools.org

Device Statistics (GP Log 0x04)
Page Offset Size         Value  Description
  1  =====  =                =  == General Statistics (rev 2) ==
  1  0x008  4              153  Lifetime Power-On Resets
  1  0x010  4            20590  Power-on Hours
  1  0x018  6      13762870401  Logical Sectors Written
  1  0x020  6        323042073  Number of Write Commands
  1  0x028  6      24966539804  Logical Sectors Read
  1  0x030  6        460248105  Number of Read Commands

So, now can you give me those stats from your M4 if you have it running endurance?....

Really, they will help me in my analysis of M4's wear characteristics.

[canthearu]

[/CODE]So, now can you give me those stats from your M4 if you have it running endurance?....

Really, they will help me in my analysis of M4's wear characteristics.

I'm not running any endurance test. However, your drive is showing, using that current smart command, far more wear than expected for a 250gig drive that has only done 6TB of writes

I wanted to see the full stats so I could see if your drive was correctly reporting writes and wear.

[Zaxx]

Imo, durability of current generation SSD's is not a valid issue anymore for 95% of users unless it's being used in a busy server. I have an OS drive (Vertex 2 60Gb 34nm) which also catches d/ls that has been in use 24/7 for well over 2 years and it's only accumulated 7 TB. I'm not saying durability should be a non-factor cuz now that the prices have come down massively, a lot more people are using them for general storage...not just for the OS. I lucked up and won an M500 1TB drive from SSD Review and it will be seeing a sht load of primary storage duty. Even so, Crucial rates it at 40GB of drive writes a day for 5 years...I'll never even come close to that...lol

[Anvil]

Surprise surprise

My X25-V is still alive

I'll be back with stats this weekend.

(I retired the drive from the test on Christmas eve, it's currently at 4% Health.)

At 2 years 7 months run-time, it's going to be a hard one to beat.

[[XC] Synthetickiller]

Not sure if this was posted before, but I came across this article:

Link


Seems intel is the only one right now (unless there's some hidden gems out there) that can handle power interrupts durring read/writes.


With my Samsung Pros not earning their PRO name, I'm thinking of switching to intel whenever I do a new build & either selling or using as a scratch drive.

[sequoia464]

Not sure if this was posted before, but I came across this article:

Link


Seems intel is the only one right now (unless there's some hidden gems out there) that can handle power interrupts durring read/writes.


With my Samsung Pros not earning their PRO name, I'm thinking of switching to intel whenever I do a new build & either selling or using as a scratch drive.

Not really certain, but I thought that Crucial uses capacitors in their new drives to help mitigate power loss issues (a crucial M4 was used in the article, not the newer M500). Maybe someone that knows more than I do about this can comment.

[Zaxx]

None of the drives tested have caps/super caps...they are speaking on power cycling the drives at their power cable. As mentioned, ALL drives we're just fine when the box power was interrupted/cycled. Wouldn't put much stock in that write up. Only way it would hurt you is if your sata power block suddenly 'broke' connection...verrry unlikely. If the psu loses power. it's caps still provide enough of a 'delay' that an ssd will normally be able to complete it's current write/flush.

Afaik, the M4's do not have caps but the M500s do...not that I want to test mine.

[DragoonXX]

Surprise surprise

My X25-V is still alive

I'll be back with stats this weekend.

(I retired the drive from the test on Christmas eve, it's currently at 4% Health.)

At 2 years 7 months run-time, it's going to be a hard one to beat.

it lives?!!?!? ok...I think my pair of X-25s are going to stay in every system until they eventually die...or sata is replaced, lol

[Zaxx]

That's insane! All I can say is what a drive...who'd have thunk?

[devsk]

Some interesting work being done by TR at http://techreport.com/review/26058/t...on-after-600tb

I wish they had one Crucial drive in there but I have a strong feeling that Crucial drives will do better than most they included in there.

They use Anvil's code.

[Computurd]

So sad this glorious thread and testing has come to a halt. Who is down for a 3D NAND bake-off. The Masked CKRyan, where art thou?

[eddieobscurant]

it came to a halt because

a) it's proven that under normal or even extreme workload you won't wear out your ssd

b) some manufacturers already give a 10 year warranty (so you can imagine how long ssds will live)

c) there was a "design fault" with these tests. After a while and thousands of p/e cycles you could continue write data to the cells and keep wearing them out , but the cells couldn't preserve the data if you keep them out of current, even for a few hours. So what was the point of the test after that.

Everything has a time, and the time of testing the endurance of ssds has passed. Although it was one of my favorites too.

techreport had a similar articles a few weeks ago

http://techreport.com/review/26523/t...-to-a-petabyte

[UranusFX]

new generation SSD Endurance test has still operating. (it's not in my testing)

-Toshiba HG6q 128GB (Marvell, Tosh A19nm MLC) already dead, 338TB.
-Plextor M6S 128GB (Marvell 88SS9188, Tosh A19nm MLC) dead, 331TB.
-Samsung 840 EVO 120GB (Samsung MEX, Sam 21nm TLC?) still alive, 205TB.
-Intel SSD 530 120GB (SandForce SF-2281, intel 20nm MLC) still alive, 170TB.
-Panasonic PR-SSB120GAK 120GB (OCZ Barefoot 3, Tosh A19nm MLC?) still alive, 180TB.
-Silicon Power Slim S55 120GB (Phison PS3108, Tosh ETT 19nm MLC?) still alive, 135TB.
-ADATA SP600 128GB (JMicron JMF661, intel 25nm MLC?) still alive, 155TB.
-Transcend SSD 340 128GB (JMicron JMF667H, Micron 20nm MLC) still alive, 185TB.

http://xn--ssd-hz3g941m.com/?p=371 (japanese)

[canthearu]

I couldn't help myself.

I'm doing a samsung 850 endurance test over at:

http://forums.whirlpool.net.au/forum....cfm?t=2311278