SSD Write Endurance 25nm Vs 34nm

[Christopher]

Oh wait... I was completely wrong. There was a reallocation event, but no actual reallocated sectors(?). Look at the M3p's SMART data again. Reallocated sector count still equals zero.



And, the 830 has another reallocation At 2.34PiB:

[kaktus1907]

yeah is such a thing even possible?
what's the chance CDI reads smart datas wrong, what's Plextool saying ?

[Christopher]

yeah is such a thing even possible?
what's the chance CDI reads smart datas wrong, what's Plextool saying ?

PlexTool reports "SSD Health" at 98%.



There is a 1.04FW available on Plextor's site, perhaps SMART data makes more sense now? It could just be that CDI reports reallocted sector count incorrectly, and it's really a different attribute.

I tried CDI5 (apparently CDI 5 can be downloaded in a more... aethetically feminine version)

[Anvil]

Intel 330 120GB

325.30TB Host writes
1.34TiB Host reads
Reallocated sectors : 05 0
Available Reserved Space : E8 100
MWI 63
[F1] Total LBAs Written 10659453
[F2] Total LBAs Read 43793
[F9] Total NAND Writes 234809GB
POH 816
MD5 OK

123.90MiB/s on avg (~67 hours)

--

Kingston SSDNow 40GB (X25-V)

1067.74TB Host writes (34987817*32)
Reallocated sectors : 05 50
Available Reserved Space : E8 99
POH 9547
MD5 OK

33.78MiB/s on avg (~67 hours)

--

I'll be keeping the normal edition of CDI

1.04 has been available for some time, looks like it fixes some issues with the update process.

[Christopher]

Yeah, CDI5 "anime edition" is better than HDSentinel's "Cheeses of the World edition", but not as fantastic as SSDLife's "lolcats version".

[kaktus1907]

This is probably completely off but i'll bite
Plextool SSD health = 100 - BB (?)

The Time spent between BB 1 to 2 was ~13 hours and it writes ~950 GiB/h, therefore total write is 12380GiB.. There is left 98*12380 = 1214192GiB(1.186TiB).. it's already written ~700TiB, 1900TiB would be too good to be true for a 128GB 24nm MLC drive ?

[Christopher]

Kaktus, I don't know that it works like that. The M3P is just a mystery, so I'd hate to make any predictions.

It is up to a 277.63MB/s average though, so it's getting faster.

[B.A.T]

Todays update:

m4
1113.8730 TiB
4665 hours
Avg speed 74.52 MiB/s.
AD 233 to 229
P/E 19180.
C3 3323
CE 58
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
744.02 TiB
23955hours
Reallocated sectors : 91 to 100
Available Reserved space: 63
MWI= 12
MD5 =OK
32.68 MiB/s on avg


Intel X25-E 64GB
367.28 TiB
1136-30=1106 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI= 97
MD5 =OK
92.04 MiB/s on avg


Mtron Pro 7025 32GB
ASU write 72.2915 TiB
Host write 135,6659 TiB
Total host write 183,2015 TiB
Day 24
MWI= 95 to 94
MD5 =OK
47.82 MiB/s on avg

[Christopher]

Samsung 830 256GB Day 100

(GiB) 2,449,403.13
(TiB) 2 391.99

(Avg) 295.11 MB/s

(B1) Wear Leveling Count: 10,625

(B6) Erase Fail Count: 9
(05) Reallocated Sectors: 36864
(POH) 2413

Plextor M3P 128GB Day 28

(GiB) 745,415.09
(TiB) 727.94

(Avg) 277.70MB/s
(POH) 763

(C4) Reallocation Event Count: 2

[AMD_Freak]

oh no seeing red on that Samsung hoping for 3,000,000GB

[Christopher]

oh no seeing red on that Samsung hoping for 3,000,000GB

Well, it did make it the better part of a PiB (not the first PiB, but the second) without a reallocation, so I suppose I was due for two more

UPDATE
uh oh, the M3P now has 6 reallocations. MWI = attribute B1 - 100. It's now down to 97%, but I'm not filled with confidence. The M3p is probably past 6500 PE cycles.

[canthearu]

UPDATE
uh oh, the M3P now has 6 reallocations. MWI = attribute B1 - 100. It's now down to 97%, but I'm not filled with confidence. The M3p is probably past 6500 PE cycles.

Well, that is still more then twice the NAND's specified lifetime. BAT's M4 just seems to be an anomaly! If Micron was selling eMLC memory, BAT's m4 would be made with it

[canthearu]

Sandisk Extreme G25 - 120gb - Day 23

Drive hours: 549
ASU GiB written: 234,306.45 GiB (228.81 TiB)
Avg MB/s: 122.47 MB/s (127 hours)
MD5: OK (verified off disk now)

Host GB written (F1): 235,714 GiB (230.19 TiB)
NAND writes (E9): 175,824 GiB (171.70 TiB)

Retired Block Count (05): 4 raw
Failure count (AB, AC): 1 program, 1 erase
Raw Error Rate (01): 116 normalized
Media Wearout Indicator (E7): 54 normalized
Wear Range Delta (B1): 4 raw
Reported Uncorrectable Errors (BB): 0 raw

[Christopher]

This is not good.

[kaktus1907]

wow another 22 in 24hr.. plextool is at 94% or 76%?

btw HDTune says Reported Uncorrectable Errors for BB

[Anvil]

Intel 330 120GB

335.79TB Host writes
1.36TiB Host reads
Reallocated sectors : 05 0
Available Reserved Space : E8 100
MWI 62
[F1] Total LBAs Written 11003031
[F2] Total LBAs Read 44644
[F9] Total NAND Writes 242385GB
POH 841
MD5 OK

123.97MiB/s on avg (~92 hours)

--

Kingston SSDNow 40GB (X25-V)

1070.54TB Host writes (35079465*32)
Reallocated sectors : 05 50
Available Reserved Space : E8 99
POH 9572
MD5 OK

33.58MiB/s on avg (~92 hours)

--

@christopher
It doesn't have to mean much at all, if it does it would be a bit early for a 128GB drive, way better than specs though.

[Christopher]

Generally speaking, 24 reallocations in as many hours isn't a good thing. We shall see.

It might even be erase or program failures ala the M4, with no reallocations occurring yet.

[Anvil]

Endurance_cr_20120628.png

Endurance_PEsummary_20120628.png

[B.A.T]

Todays update:

m4
1118.5135 TiB
4683 hours
Avg speed 74.74 MiB/s.
AD 229 to 227
P/E 19259.
C3 3323
CE 58
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
746.23 TiB
23973 hours
Reallocated sectors : 100 to 104
Available Reserved space: 63 to 62
MWI= 12
MD5 =OK
32.24 MiB/s on avg


Intel X25-E 64GB
372.94 TiB
1153-30=1123 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI= 97
MD5 =OK
89.78 MiB/s on avg


Mtron Pro 7025 32GB
ASU write 73.2138 TiB
Host write 137,6927 TiB
Total host write 185,0331 TiB
Day 25
MWI= 94
MD5 =OK
47.76 MiB/s on avg


The Mtron dropped out again today. Runs fine now but I had to restart the rig.

[Christopher]

Okay -- now we have reallocated sectors. 335872 to be exact. If we multiply that by 512, we get 171966464 bytes, or 164MB. I'm not sure how that works out, because if you divide 335872 by the 41 reallocation events, you get 8192 sectors per reallocation. 8192 sectors is 4MB, so each erase block is 4MB? That can't be right.

[UranusFX]

Not, not good...
Increase count up of re-allocation sectors too more rapidly.
To predict. SMART spit out in the middle or long-term span?

[neeyuese]

Okay -- now we have reallocated sectors. 335872 to be exact. If we multiply that by 512, we get 171966464 bytes, or 164MB. I'm not sure how that works out, because if you divide 335872 by the 41 reallocation events, you get 8192 sectors per reallocation. 8192 sectors is 4MB, so each erase block is 4MB? That can't be right.

24nm Toshiba Toggle DDR 2 NAND erase block is 2MB , but if you use 2 plane algorithm (increase write speed) , Plane A block fail will cause Plane B block (same address) also useless. Maybe logical block size is 4MB.

[Christopher]

565,248 sectors and 69 blocks. M3P MWI is down to 88.

I was hoping it would stabilize.

[Christopher]

In two hours it went from 69 to 106 blocks.

[B.A.T]

Todays update:

m4
1124.1442 TiB
4705 hours
Avg speed 74.54 MiB/s.
AD 227 to 223
P/E 19354.
C3 3323
CE 58
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
748.85 TiB
23995 hours
Reallocated sectors : 104 to 112
Available Reserved space: 62
MWI= 12
MD5 =OK
34.46 MiB/s on avg


Intel X25-E 64GB
379.78 TiB
1175-30=1145 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI= 97
MD5 =OK
90.14 MiB/s on avg


Mtron Pro 7025 32GB
ASU write 76.8286 TiB
Host write 144,4311 TiB
Total host write 191,9666 TiB
Day 26
MWI= 94
MD5 =OK
47.81 MiB/s on avg