SSD Write Endurance 25nm Vs 34nm

[canthearu]

Patriot Torqx-2 64GB - Day 70

Drive hours: 1,689
ASU GiB written: 116,081.57 GiB (113.36 TiB)
Avg MB/s: 25.48 MB/s
MD5: OK

Bad blocks: 2 grown, 83 factory
Wear cycle counter: 7268 average erase cycles, 11116 maximum erase cycles.
Total ECC Error Count: 1

Intel 520 60GB - Day 62

Drive hours: 1,487
ASU GiB written: 466,389.60 GiB (455.46 TiB)
Avg MB/s: 95.90 MB/s
MD5: OK

Host GB written (F1): 469,576.19 GiB (458.57 TiB, 15026438 raw)
NAND writes (F9): 332,454 GiB (324.67 TiB)

Reallocated sectors (05): 0
Failure count (AB, AC): 0 program, 0 erase
Raw Error Rate (8B): 118 normalised
Avaliable Reserved Space (AA): 100 normalised
Media Wearout Indicator (E9): 1

Vertex 4 128GB - Day 9

Drive hours: 240
ASU GiB written: 96,551.97 GiB (94.29 TiB)
Avg MB/s: 149.07 MB/s
MD5: OK

Sector GB written (E8): 97,783.18 GiB (95.49 TiB, 205066199033 raw)
Reallocated sectors (05): 0
Raw Read Error Rate (01): 6
Remaining Life (E9): 88

[Christopher]

\\--------------------------------------------//

Samsung 830 256 Day 40

937,036.71 GiB
915.07 TiB

296.21 MB/s Average
930 Hours

4080 Wear Leveling Count
MWI 1

0/0 Erase/Program Fail

OCZ Vertex Turbo 64 Day 58

422,275.36 GiB
412.37 TiB

84.15 MB/s Avg
1427 hours

7550 Avg PE count
-100 MWI

0/0/0 Read/Program/Erase Fail

//--------------------------------------------\\

[B.A.T]

Todays update:
And we start with a new record for the Crucial m4

m4
772.1866 TiB
3291 hours
Avg speed 73.11 MiB/s.
AD gone from 171 to 168.
P/E 13338.
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
551.78 TiB
22624 hours
Reallocated sectors : 128
Available Reserved space: 92
MWI=150 to 144
MD5 =OK
43.47 MiB/s on avg

[Anvil]

\\--------------------------------------------//

Samsung 830 256 Day 40

937,036.71 GiB
915.07 TiB

40 days to write more than the X25-V managed in ~350 days

--

Kingston SSDNow 40GB (X25-V)

906.42TB Host writes
Reallocated sectors : 05 41
Available Reserved Space : E8 99
POH 8134
MD5 OK

33.14MiB/s on avg (~405 hours)

[Christopher]

40 days to write more than the X25-V managed in ~350 days

--

Yes, but will the 830 still be writing in a year from now? Probably not

[Anvil]

I'd be very surprised if the X25-V was still writing in a year from now, we might need an SLC to beat it.

[Christopher]

Actually, the 40 day mark isn't until about 30 hours from now. At that point, the 830 should have about 971,000GiB on it. That means that after 120 days it should have around 2,913,000GiB on it (and right around 12,000PE cycles).

That's moderately frightening.

[Anvil]

Endurance_cr_20120427.png

@Christopher

So, what were the final stats for the Octane?

I'd like to think that the 830, being a 256GB drive was able to get to 3PiB, we'll know within a few months , fast drives yields quick results.
(3PiB would have taken more than 3 years on the X25-V )

[Christopher]

Anvil, the Octane's last stats so far as I can tell is:

311,111.64 GiB
303.82 TiB
159.22 MB/s
Reallocations 238

It died very shortly after the last update, within an hour or so.

-------
The 830 is at 948,311 GiB right now, so it will be another 95 hours or so before it hits 1PiB I think, and only 48 or so till 1PB.

[Anvil]

A 128GB drive dropping out at 64GB "levels" is not what one would expect.

Using newly released drives with almost virgin firmware might not be the best option for Endurance testing, still Host Writes / Capacity ended at about 2600+

Endurance_failed_drives.png

Endurance_PEsummary_20120428.png

[Christopher]

I do feel that the Octane's untimely demise was almost assuredly due to some bum flash, and not the FW or controller. If I were to get another one, and test it, it would probably not die at 311,111GiB.

[Christopher]

\\--------------------------------------------//

Samsung 830 256 Day 41

964,493.65 GiB
941.89 TiB

295.73 MB/s Average
957 Hours

4199 Wear Leveling Count
MWI 1

0/0 Erase/Program Fail

OCZ Vertex Turbo 64 Day 59

430,217.50 GiB
420.13 TiB

84.43 MB/s Avg
1453 hours

7690 Avg PE count
-100 MWI

0/0/0 Read/Program/Erase Fail

//--------------------------------------------\\

The 830 is doing about 24 950 GiB per day at an average speed of 295.71MB/s.

[B.A.T]

Todays update:

m4
779.5108 TiB
3320 hours
Avg speed 73.14 MiB/s.
AD gone from 168 to 164.
P/E 13460.
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
556.13 TiB
22653 hours
Reallocated sectors : 128
Available Reserved space: 92
MWI=144 to 138
MD5 =OK
43.49 MiB/s on avg

[canthearu]

Anvil .... the real surprise is your Kingston (X25-V rebadge).

Did you really expect it to last 25k write cycles when you first started testing? At this rate it is going to outlive the corsair 120gig ... it still has heaps and heaps of spare sectors to use.

[canthearu]

Patriot Torqx-2 64GB - Day 72

Drive hours: 1,738
ASU GiB written: 120,454.57 GiB (117.63 TiB)
Avg MB/s: 25.52 MB/s
MD5: OK

Bad blocks: 2 grown, 83 factory
Wear cycle counter: 7541 average erase cycles, 11548 maximum erase cycles.
Total ECC Error Count: 1

Intel 520 60GB - Day 64

Drive hours: 1,535
ASU GiB written: 482,626.01 GiB (471.31 TiB)
Avg MB/s: 95.90 MB/s
MD5: OK

Host GB written (F1): 485,898.63 GiB (474.51 TiB, 15548756 raw)
NAND writes (F9): 344,005 GiB (335.94 TiB)

Reallocated sectors (05): 0
Failure count (AB, AC): 0 program, 0 erase
Raw Error Rate (8B): 118 normalised
Avaliable Reserved Space (AA): 100 normalised
Media Wearout Indicator (E9): 1

Vertex 4 128GB - Day 11

Drive hours: 299
ASU GiB written: 121,372.80 GiB (118.53 TiB)
Avg MB/s: 146.02 MB/s
MD5: OK

Sector GB written (E8): 122,733.71 GiB (119.86 TiB, 257391250415 raw)
Reallocated sectors (05): 0
Raw Read Error Rate (01): 6
Remaining Life (E9): 85

[Rubycon]

Fascinating thread and info so far.
Any cut and dry conclusions here as far as best drive to select for heavy use and combination of speed?
The 830 looks like a beast. A dozen of those in RAID0 would take quite a beating for a while.

[Christopher]

Fascinating thread and info so far.
Any cut and dry conclusions here as far as best drive to select for heavy use and combination of speed?
The 830 looks like a beast. A dozen of those in RAID0 would take quite a beating for a while.

The 830 is a beast. But it's also 256GB and writes an average of 296MB/s, which is like a 64GB writing 74MB/s avg. But if you really want something that can handle abuse, the X25-V is pretty much unmatched.

[B.A.T]

Todays update:

m4
784.6358 TiB
3340 hours
Avg speed 73.14 MiB/s.
AD gone from 164 to 161.
P/E 13546.
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
559.18 TiB
22674 hours
Reallocated sectors : 128
Available Reserved space: 92
MWI=138 to 133
MD5 =OK
43.46 MiB/s on avg


I'll shut down the testrig for a couple of hours to do some maintenace tonight. Everything will be back running when I go to bed.

[Christopher]

Samsung 830 256 Day 41

989,593.91 GiB
966.40 TiB

296.68 MB/s Average
981 Hours

4307 Wear Leveling Count
MWI 1

0/0 Erase/Program Fail



OCZ Vertex Turbo 64 Day 59

437,336.26 GiB
427.08 TiB

84.62 MB/s Avg
1477 hours

7817 Avg PE count
-100 MWI

0/0/0 Read/Program/Erase Fail

--------------------------------
981 hours to 989,593 GiB? More of that!

The Vertex Turbo is doing quite well, though considering it's really still working on MWI, it should be. It really should be considered as MWI 22 at the moment. Getting close, but still a couple of weeks away from true MWI 0.

[Anvil]

Endurance_cr_20120429.png

Kingston SSDNow 40GB (X25-V)

912.51TB Host writes
Reallocated sectors : 05 41
Available Reserved Space : E8 99
POH 8185
MD5 OK

33.36MiB/s on avg (~34 hours)

--

I had to start using the new ASU release candidate so the drive had a 5 minute break last night.
ASU RC1 will most likely be available next weekend.

[Anvil]

Anvil .... the real surprise is your Kingston (X25-V rebadge).

Did you really expect it to last 25k write cycles when you first started testing? At this rate it is going to outlive the corsair 120gig ... it still has heaps and heaps of spare sectors to use.

I don't know if expect is the right word for it, hoping fits better.

There is another drive out there (just like mine) that sort of inspired this test, we couldn't know for sure that it was for real but it looks that way.
(the other drive got past 20K write cycles, last report was over a year ago, don't know what happened to it)

[B.A.T]

And I'm 11000 P/E behind you. Wonder how much I can expect from the m4?!? It'll take another 3-4 months before we know....

[Christopher]

I have high hopes for the Turbo, but it seems like it would very much be the luck of the draw. Your chances of getting a "representative" X25-v are probably better. Bluestang's M225 did well, the first Turbo I did was sub par, and the current Turbo is looking great.

If I started running my 3 Vs today, I'd wager they'd all still be working in 2013.

[Anvil]

Christopher,

Will we ever see the 830 64GB again?

[Christopher]

Christopher,

Will we ever see the 830 64GB again?

I don't think we will. It's probably time to remove it from the active roster.


That thing chewed through PE cycles like nobody's business, though.