SSD Write Endurance 25nm Vs 34nm

[Hopalong X]

rofl...you sound a lot like me when I get fed up with something. If you go off the deep end, you gotta post a vid of you wielding a 15lb sledge hammer.

All Christopher has to do is look at an Mtron to make it throw off sparks and go up in smoke. That is before he hooks up the power cable.

Stand clear if he gets access to a power supply!
He will try 220v if that is all he can find. "That SE is going to work this time".

[Christopher]

Frying the drive is just as good as a SE in my opinion. And when drives go lame, the only compassionate thing to do is execute them.

[Hopalong X]

MTRON
Rest In Pieces

[Zaxx]

...the only compassionate thing to do is execute them.

roflmfao....too funny.

[Christopher]

If you can put a horse down for a gimp hoof, then putting a MTRON down is basically is the only considerate course of action

[Christopher]

Samsung 830 256GB Day 91

(GiB) 2,226,819.02
(TiB) 2 174.62

(Avg) 294,07 MB/s

(B1) Wear Leveling Count: 9664

(B6) Erase Fail Count: 7
(05) Reallocated Sectors: 28672
(POH) 2198

Plextor M3P 128GB Day 21

(GiB) 533,900.39
(TiB) 421.39

(Avg) 275.78 MB/s
(POH) 549

-------------------------------------------------------
How many bytes does the 830 have on it?


4 700 701 261 985 sectors * 512 = 2 406 759 046 136 320 bytes. That's.... that is a proper, biblically epic number.

[kaktus1907]

The Plextor is getting speedier, it was ~272MB/s earlier but with your recent updates adds a mb/s on it.. Should i take it as a bad sign?(it was steady 271-272 until 3K run out)
Btw i think its not 400TiB but 500

[[XC] Oj101]

521.39 TiB

[Anvil]

Intel 330 120GB

240.2TB Host writes
Reallocated sectors : 05 0
Available Reserved Space : E8 100
MWI 73
[F1] Total LBAs Written 7870807
[F2] Total LBAs Read 30857 //
[F9] Total NAND Writes 173378GB
POH 618
MD5 OK

123.55MiB/s on avg (~38 hours)

--

Kingston SSDNow 40GB (X25-V)

1044.70TB Host writes (34233896*32)
Reallocated sectors : 05 46
Available Reserved Space : E8 99
POH 9347
MD5 OK

34.71MiB/s on avg (~38 hours)

[canthearu]

Intel 520 60GB - Day 114

Drive hours: 2,681
ASU GiB written: 854,340.34 GiB (834.32 TiB)
Avg MB/s: 93.53 MB/s (237 hours)
MD5: OK

Host GB written (F1): 860,355.69 GiB (840.19 TiB, 27531382 raw)
NAND writes (F9): 609,064 GiB (594.79 TiB)

Reallocated sectors (05): 0
Failure count (AB, AC): 0 program, 0 erase
Raw Error Rate (8B): 104 normalized
Available Reserved Space (AA): 100 normalized
Media Wearout Indicator (E9): 1

Sandisk Extreme G25 - 120gb - Day 14

Drive hours: 335
ASU GiB written: 142,234.58 GiB (138.90 TiB)
Avg MB/s: 121.58 MB/s (237 hours)
MD5: OK

Host GB written (F1): 143,170 GiB (139.81 TiB)
NAND writes (E9): 106,775 GiB (104.27 TiB)

Retired Block Count (05): 4 raw
Failure count (AB, AC): 1 program, 1 erase
Raw Error Rate (01): 105 normalized
Media Wearout Indicator (E7): 73 normalized
Wear Range Delta (B1): 3 raw
Reported Uncorrectable Errors (BB): 0 raw

@Anvil: can you change ASU so that it reads the file for MD5 checking off the disk, rather then out of memory cache. My endurance computer has 8gig RAM so reading in the file for MD5 checking just seems to be checking my memory rather then the disk! Since I fixed my memory a while back, it probably won't find any faults there!

I've just been noticing that the total host reads haven't been increasing on my drives at all! Running HDD-scan though (while ASU is running) does seem to increase the host reads like it should

[Christopher]

The Plextor is getting speedier, it was ~272MB/s earlier but with your recent updates adds a mb/s on it.. Should i take it as a bad sign?(it was steady 271-272 until 3K run out)
Btw i think its not 400TiB but 500

I did botch the last Update, the GiB is correct, but not the TiB.

I expect the drive's averages to increase slightly week by week (I consider this a sign that the drive is finally "broken in"). Both drives are gaining small amounts of speed week by week. I consider it a good thing.

[Kain665]

The 830 hasn't popped any blocks in a very long time...

[Anvil]

@Anvil: can you change ASU so that it reads the file for MD5 checking off the disk, rather then out of memory cache. My endurance computer has 8gig RAM so reading in the file for MD5 checking just seems to be checking my memory rather then the disk! Since I fixed my memory a while back, it probably won't find any faults there!

I've just been noticing that the total host reads haven't been increasing on my drives at all! Running HDD-scan though (while ASU is running) does seem to increase the host reads like it should

I'll check out whats going on, might be why there's been virtually no change to F2 on the 330.

The Kingston has added 2 re-allocated sectors. (in 6-7 hours)

[Christopher]

The 830 hasn't popped any blocks in a very long time...

No, it hasn't. In fact, I'll have to check, but it's been 700TB or so since the last pop.

[Christopher]

Samsung 830 256GB Day 92

(GiB) 2,249,907.95
(TiB) 2 197.17

(Avg) 294,70 MB/s

(B1) Wear Leveling Count: 9763

(B6) Erase Fail Count: 7
(05) Reallocated Sectors: 28672
(POH) 2220

Plextor M3P 128GB Day 22

(GiB) 555,342.60
(TiB) 542.33

(Avg) 275.62 MB/s
(POH) 571

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

[Anvil]

I've made changes to make sure that the MD5 testfile is not available in the disk-cache, will be available later today.

I'll try to trace back what caused the change.

[Anvil]

Intel 330 120GB

256.17TB Host writes
Reallocated sectors : 05 0
Available Reserved Space : E8 100
MWI 71
[F1] Total LBAs Written 8394035
[F2] Total LBAs Read 31991 // looks OK
[F9] Total NAND Writes 184903GB
POH 655
MD5 OK

123.44MiB/s on avg (~28 hours)

--

Kingston SSDNow 40GB (X25-V)

1049.03TB Host writes (34374762*32)
Reallocated sectors : 05 48
Available Reserved Space : E8 99
POH 9385
MD5 OK

33.21MiB/s on avg (~23 hours)

--

Some of the drives aren't up to date. (just a few days off)

Endurance_cr_20120620.png

[Anvil]

Make sure to download the new build, it no longer allows the MD5 testfile to be cached.


Download link

[Christopher]

Samsung 830 256GB Day 92

(GiB) 2,278,395.09
(TiB) 2 224.99

(Avg) 294,47MB/s

(B1) Wear Leveling Count: 9886

(B6) Erase Fail Count: 7
(05) Reallocated Sectors: 28672
(POH) 2248

Plextor M3P 128GB Day 22

(GiB) 582,785.07
(TiB) 568.12

(Avg) 276.03 MB/s
(POH) 599

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

[Anvil]

Intel 330 120GB

260.43TB Host writes
1010GiB Host reads
Reallocated sectors : 05 0
Available Reserved Space : E8 100
MWI 71
[F1] Total LBAs Written 8533905
[F2] Total LBAs Read 32332 // looks OK
[F9] Total NAND Writes 187990GB
POH 665
MD5 OK

123.59MiB/s on avg (~38 hours)

--

Kingston SSDNow 40GB (X25-V)

1050.18TB Host writes (34412198*32)
Reallocated sectors : 05 48
Available Reserved Space : E8 99
POH 9395
MD5 OK

33.19MiB/s on avg (~33 hours)

--

[canthearu]

I've upgraded one of my instances of ASU .... the new version does not seem to be helping WRT actually reading the MD5 off disk

[B.A.T]

We have another kingston here :P

But seriously, these Mtron drives are proving to be seriously unreliable/fragile. Just highlights the importance of firmware quality.

SLC + bad firmware = still a bad drive

It feels very much like the Kingston V+100. It's dropping out every 2-3 day but luckily for me the high WA burns through P/E.

Todays update:
I've been a little busy with visitors from my family the last couple of day. Here are fresh numbers

m4
1075.7994 TiB
4514 hours
Avg speed 73.59 MiB/s.
AD gone from 6 to 255 to 251
P/E 18536.
CE 58
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
724.83 TiB
23805 hours
Reallocated sectors : 172 to 200
Available Reserved space: 74 to 73
MWI= 42
MD5 =OK
39.97 MiB/s on avg


Intel X25-E 64GB
320.06 TiB
986-30=956 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI= 98 to 97
MD5 =OK
90.96 MiB/s on avg


Mtron Pro 7025 32GB
ASU write 56.1501 TiB
Host write 104,79630 TiB
Total host write 152,3318 TiB
Day 18
MWI= 96 to 95
MD5 =OK
47.79 MiB/s on avg


The Mtron has dropped out 1 time tha last 3 days. The rest is ok and running fine.

[Anvil]

I've upgraded one of my instances of ASU .... the new version does not seem to be helping WRT actually reading the MD5 off disk

Hmmm

You should update all instances of the app. (as the old app could leave a copy of the source-file in the buffer)
Is the source file located on the same drive as is being tested?
Is there some AV that could interfere? (by scanning the file and leaving it in the buffer, if so, disable scanning on MD5 extensions or the TEST folder)

I'll have another look but there should be no reason to the file being buffered.

There will be an update one of these days as generating the checksum has been optimized. (a bit faster)

[Christopher]

Samsung 830 256GB Day 93

(GiB) 2,304,864.77
(TiB) 2,250.83

(Avg) 294.68MB/s

(B1) Wear Leveling Count: 10 000

(B6) Erase Fail Count: 7
(05) Reallocated Sectors: 28672
(POH) 2273

Plextor M3P 128GB Day 23

(GiB) 607,939.62
(TiB) 593.69

(Avg) 276.41 MB/s
(POH) 624

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

[canthearu]

Hmmm

You should update all instances of the app. (as the old app could leave a copy of the source-file in the buffer)
Is the source file located on the same drive as is being tested?
Is there some AV that could interfere? (by scanning the file and leaving it in the buffer, if so, disable scanning on MD5 extensions or the TEST folder)

I'll have another look but there should be no reason to the file being buffered.

There will be an update one of these days as generating the checksum has been optimized. (a bit faster)

Ah, moving the source file to another drive has fixed the md5 testing

I will post an update in a few hours ... let speeds somewhat stabalize again.!