SSD Write Endurance 25nm Vs 34nm

[B.A.T]

Thanks for that
With 5 sec between each loop the avg increases to 112 MiB/s. I'll try a couple of loops with 2 GB freespace since it so much OP.

[Christopher]

Considering how much like the Gen1 80GB it is, I'd say it's performance is pretty damn good.

[B.A.T]

It really is
Maybe I need to treat it like the G1 with regular maintenace. I'll know in a couple of days.

[canthearu]

Hmm, the X25-E would probably exceed 10PiB before failing .... gonna be a long wait if it is.

[Anvil]

Let it run for some time (until i settles), it is interesting seeing how that drive develops the next few days.
~100MB/s w/o TRIM is pretty good for this workload, as the Intel does no background GC it wont benefit from extra pausing though.
The only cleaning the X25-V gets is the occational cleaning using the Toolbox, usually a few weeks between those.
My drive is still bootable so on it's 1 year day of testing I might try booting it. (3 more days)

[B.A.T]

I'm curios about how the wmi will perform. Then it's easyer to predict how much time this will take

[Anvil]

Hmm, the X25-E would probably exceed 10PiB before failing .... gonna be a long wait if it is.

I'm expecting/hoping for more.

The 64GB is rated for 2 petabytes of random writes, it'll be a long wait.

[B.A.T]

With an estimated avg speed of 90 MiB/s (I know I'm conservative here) I'll write 2,6 PiB a year.
Anvil has been testing for a year in 3 days and I'm only 6 weeks away from my first year. I'll think I'm up for the challenge

[Anvil]

I've left the Vertex Turbo in the chart this time, as it wasn't updated in the previous chart.

Endurance_cr_20120512.png

Endurance_PEsummary_20120512.png

[Christopher]

Now I'm paranoid about the erase failures on the 830, and I can't stop checking in on it. It made it a good deal beyond MWI before the first reallocation, but now I'm gun shy. I already had to bury the Vertex Turbo in an unmarked grave.

[Anvil]

I try to check in on the V a few times per day, it's not much one can do.

I have tried using the default windows driver (still raid mode) and the V was undoubtedly slower, within ~35-40 hours it had dropped below 30MiB/s.

Using the Intel 10 series driver (currently using 10.8) is clearly a winner vs the default one, the V can go for weeks at ~33MiB/s

[canthearu]

The Torqx 2 has popped 8 more blocks since yesterday ... I'm preparing the funeral now.

[Anvil]

That is a bit early, less than 150 TiB.

The Intel 520 looks to be doing fine.

[Christopher]

As long as those blocks aren't all on the same die or plane you should be golden. At least, that's what I keep telling myself.

[B.A.T]

Todays update:

m4
866.3879 TiB
3671 hours
Avg speed 74.63 MiB/s.
AD gone from 117 to 114.
P/E 14945.
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
606.44 TiB
22984 hours
Reallocated sectors : 154 to 155
Available Reserved space: 90
MWI=124
MD5 =OK
41.94 MiB/s on avg


Intel X25-E 64GB
14,87 TiB
71-30=41 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI=99
MD5 =OK
102.55 MiB/s on avg

[AMD_Freak]

I haven't read all 173 pages but is there any drives in Raid0, Curious on the effects of no TRIM

[Christopher]

I haven't read all 173 pages but is there any drives in Raid0, Curious on the effects of no TRIM

No, but the two G1 Intel drives aren't using TRIM. Generally speaking, without TRIM W.A. is higher, average speed is lower. If you look in the charts, the drives without TRIM are noted (see Anvil's update a few posts back).

[Christopher]

Samsung 830 256GB Day 56

1,338,797.80 GiB
1,307.41 TiB

293.95 MB/s Average
1325 Hours

5815 Wear Leveling Count
MWI 1

4/0 Erase/Program Fail
Used Reserved Block Count: 8/16384 sectors

[Anvil]

Kingston SSDNow 40GB (X25-V)

951.52TB Host writes
Reallocated sectors : 05 42 (up 1)
Available Reserved Space : E8 99
POH 8532
MD5 OK

33.33MiB/s on avg (~20 hours)

[B.A.T]

Todays update:

m4
873.0461 TiB
3697 hours
Avg speed 74.57 MiB/s.
AD gone from 114 to 111.
P/E 15058.
MD5 OK.
Reallocated sectors : 00


Intel X25-M G1 80GB
610.23 TiB
23010 hours
Reallocated sectors : 155 to 156
Available Reserved space: 90
MWI=124
MD5 =OK
42.20 MiB/s on avg


Intel X25-E 64GB
23,96 TiB
97-30=67 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI=99
MD5 =OK
101.42 MiB/s on avg

[Anvil]

Kingston SSDNow 40GB (X25-V)

953.08TB Host writes
Reallocated sectors : 05 42
Available Reserved Space : E8 99
POH 8546
MD5 OK

33.28MiB/s on avg (~33 hours)

Just 1 day to go!

[Christopher]

Samsung 830 256GB Day 57

1,362,943.72 GiB
1,331.00 TiB

290.72 MB/s Average
1349 Hours

5919 Wear Leveling Count
MWI 1

4/0 Erase/Program Fail
Used Reserved Block Count: 8/16384 sectors

[Anvil]

Kingston SSDNow 40GB (X25-V)

954.25TB Host writes
Reallocated sectors : 05 43 (1 up)
Available Reserved Space : E8 99
POH 8556
MD5 OK

33.26MiB/s on avg (~43 hours)

Tonight, it will be 1 year since the test started. (in about 15 hours)
(it started the 16th of May 2011 at 02:20 AM)

[B.A.T]

Probably the toughest trial an ssd has endured

And quite a testimony of perseverance

[canthearu]

Patriot Torqx-2 64GB - Day 90

Drive hours: 2,036
ASU GiB written (first run): 137,907.07 GiB (134.68 TiB)
ASU GiB written (current run): 6,268.46 GiB (6.12 TiB)
ASU GiB written (Total): 144,175.53 GiB (140.8 TiB)
Avg MB/s: 20.57 MB/s
MD5: OK

Bad blocks: 20 grown, 83 factory
Wear cycle counter: 9067 average erase cycles, 13875 maximum erase cycles.
Total ECC Error Count: 1
SATA PHY error count: 250

Intel 520 60GB - Day 82

Drive hours: 1,919
ASU GiB written: 611,308.40 GiB (596.98 TiB)
Avg MB/s: 94.92 MB/s
MD5: OK

Host GB written (F1): 615,608.66 GiB (601.18 TiB, 19699477 raw)
NAND writes (F9): 435,813 GiB (425.60 TiB)

Reallocated sectors (05): 0
Failure count (AB, AC): 0 program, 0 erase
Raw Error Rate (8B): 115 normalised
Avaliable Reserved Space (AA): 100 normalised
Media Wearout Indicator (E9): 1

Vertex 4 128GB - Day 29

Drive hours: 718 (556 + 162)
ASU GiB written: 310,567.51 GiB (303.29 TiB)
Avg MB/s: 125.40 MB/s
MD5: OK

Sector GB written (E8): 313,245.90 GiB (305.90 TiB, 486156078177 + 170768184054) raw)
Reallocated sectors (05): 1
Raw Read Error Rate (01): 6
Remaining Life (E9): 52 (82 current - 30)

Lots more bad blocks on the Torqx 2, and the vertex 4 popped it's first bad block!