SSD Write Endurance 25nm Vs 34nm

[Anvil]

Kingston SSDNow 40GB (X25-V)

1365.35TB Host writes 44739793*32)
Reallocated sectors : 05 768 -
Available Reserved Space : E8 86 (86)
POH 12475
MD5 OK

31.53MiB/s on avg (~92 hours)

--

[B.A.T]

Here are todays update:

m4
1984.9957 TiB
8143 hours
Avg speed 73.10 MiB/s.
AD 253 to 241.
P/E 34177
C3 456606202 to 614011911
01 167920 to 239518 (1)
CE 90
MD5 OK.
Reallocated sectors : 00


Intel X25-E 64GB
1.35 PiB
4607-30=4577 hours
Reallocated sectors : 79 to 157
Available Reserved space: 59 to 56
MWI= 91
MD5 =OK
77.59 MiB/s on avg


Mtron Pro 7025 32GB
Paused until new firmware is updated.

OCZ Vertex 1 120GB
Will start after an update to the latest firmware.

[canthearu]

Samsung 840 Day 5

Drive Hours: 124
ASU GiB Written: 43,190.93
Avg MB/s: 110.47 (60.86 hours)
MD5: OK

Wear Leveling Count (B1): 345 raw (65 normalized)

Reallocated blocks (B3,05): 0 (0 sectors)
Failure count (B5, B6): 0 program, 0 erase
Uncorrectable Error Count: 0
ECC Error Rate (C3): 0

[canthearu]

Samsung 840 Day 8

Drive Hours: 193
ASU GiB Written: 68,231.36
Avg MB/s: 112.75 (3.52 hours)
MD5: OK

Wear Leveling Count (B1): 545 raw (46 normalized)

Reallocated blocks (B3,05): 0 (0 sectors)
Failure count (B5, B6): 0 program, 0 erase
Uncorrectable Error Count: 0
ECC Error Rate (C3): 0

[hollywood]

Amazing work gentlemen. Saw this article referenced on AnandTech.

Funny too, as my Intel 320 120GB is only 2.03 TiB in total writes. Guess at this useage, it should last well into the 2030's...

[Christopher]

Samsung 830 256GB Day 250

(GiB) 6,223,052
(TiB) 6,077
(PiB) 5.97


(Avg) 298.38 MB/s over the past 2000+ hours

(B1) Wear Leveling Count: 26,901

(B5) Program Fail Count: 0
(B6) Erase Fail Count: 25
(B7) Runtime Bad Block Count: 461
(05) Reallocated Sectors: 1,888,256

Read Failures: 436

(POH) 6177

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

I couldn't wait until exactly 6PiB for the next update, so here it is: 5.97PiB

[ice_chill]

samsung 830 256gb day 250

(gib) 6,223,052
(tib) 6,077
(pib) 5.97


(avg) 298.38 mb/s over the past 2000+ hours

(b1) wear leveling count: 26,901

(b5) program fail count: 0
(b6) erase fail count: 25
(b7) runtime bad block count: 461
(05) reallocated sectors: 1,888,256

read failures: 436

(poh) 6177

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

i couldn't wait until exactly 6pib for the next update, so here it is: 5.97pib

epic!

[Kain665]

Guys, 175k IOPS from a SATA drive.

[gordo]

Christopher, how many sectors are there on the 830 to re-allocate to? Can we extrapolate out the (fairly steady?) reallocation rate and predict when it should run out of space and finally keel over?

For the record, 6PiB is insane. My 830 is averaging 8.5 GB writes/day. At that rate, the NAND would last until the year 4016 -- assuming it didn't get fried by a lightning surge, hurricane, earthquake, flood, nuclear bomb or meteor strike first.

[Demo]

My 240Gb ForceGT suddenly failed after 5 months of use, it only had about 3Tb of host write on it.. It was working fine until a reboot and it became unpartitioned, it can be detected by the bios/os/parted magic but cannot be accessed.. I/O error of some sorts. What could have possibly caused this ?

[Christopher]

My 240Gb ForceGT suddenly failed after 5 months of use, it only had about 3Tb of host write on it.. It was working fine until a reboot and it became unpartitioned, it can be detected by the bios/os/parted magic but cannot be accessed.. I/O error of some sorts. What could have possibly caused this ?

Read failures can cause data corruption, leading to hosed partitions. It's happened a few times to me.

[Christopher]

Samsung 830 256GB Day 252

(GiB) 6,282,908
(TiB) 6,135
(PiB) 6.03


(Avg) 298.49 MB/s over the past 3400+ hours

(B1) Wear Leveling Count: 27,159

(B5) Program Fail Count: 0
(B6) Erase Fail Count: 27
(B7) Runtime Bad Block Count: 542
(05) Reallocated Sectors: 2,220,032

Read Failures: 515

(POH) 6133

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


And that's what 6PiB looks like.

[canthearu]

My 240Gb ForceGT suddenly failed after 5 months of use, it only had about 3Tb of host write on it.. It was working fine until a reboot and it became unpartitioned, it can be detected by the bios/os/parted magic but cannot be accessed.. I/O error of some sorts. What could have possibly caused this ?

Probably firmware fault.

Since you can still see the drive in the BIOS, try booting parted magic and secure erasing the drive. That might reset it and get it going again.

[DeadlyFire]

Samsung 830 256GB Day 252

(GiB) 6,282,908
(TiB) 6,135
(PiB) 6.03


(Avg) 298.49 MB/s over the past 3400+ hours

(B1) Wear Leveling Count: 27,159

(B5) Program Fail Count: 0
(B6) Erase Fail Count: 27
(B7) Runtime Bad Block Count: 542
(05) Reallocated Sectors: 2,220,032

Read Failures: 515

(POH) 6133

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


And that's what 6PiB looks like.

Nothing short of legendary! Kudos to you for running this test for so long and cheers to Samsung for this determined brute of a drive, this thread was a huge factor for me in buying several Samsung 830s

[Demo]

Probably firmware fault.

Since you can still see the drive in the BIOS, try booting parted magic and secure erasing the drive. That might reset it and get it going again.

Already did, it would give out an I/O error and would not perform the secure erase.

[canthearu]

Samsung 840 Day 13

Drive Hours: 321
ASU GiB Written: 111,733.01
Avg MB/s: 105.73 (5.02 hours)
MD5: OK

Wear Leveling Count (B1): 894 raw (11 normalized)

Reallocated blocks (B3,05): 0 (0 sectors)
Failure count (B5, B6): 0 program, 0 erase
Uncorrectable Error Count: 0
ECC Error Rate (C3): 0

Drive has nearly reached it's specified lifetime writes.

[Anvil]

Kingston SSDNow 40GB (X25-V)

1388.55TB Host writes 45500018*32)
Reallocated sectors : 05 768 -
Available Reserved Space : E8 86 (86)
POH 12688
MD5 OK

31.92MiB/s on avg (~210 hours)

--

I'll have a look at the charts asap. (tonight or tomorrow)

[F@32]

Samsung 840 Day 13

Drive Hours: 321
ASU GiB Written: 111,733.01
Avg MB/s: 105.73 (5.02 hours)
MD5: OK

Wear Leveling Count (B1): 894 raw (11 normalized)

Reallocated blocks (B3,05): 0 (0 sectors)
Failure count (B5, B6): 0 program, 0 erase
Uncorrectable Error Count: 0
ECC Error Rate (C3): 0

Drive has nearly reached it's specified lifetime writes.

Thanks for testing this one. Eagerly awaiting further progress. Ordered one yesterday and wonder if it will be good enough. I doubt I will be writing 10Gb a day on it, so I hope it will last fine as a main OS drive.

[Christopher]

Nothing short of legendary! Kudos to you for running this test for so long and cheers to Samsung for this determined brute of a drive, this thread was a huge factor for me in buying several Samsung 830s

Now that you mention it, it is moderately epic.

[Zaxx]

Far to modest my man...I guess it'll take 10 PiB to earn the title of truly epic. lol

Too lazy to search the thread....did you buy this drive 'off the shelf' so to speak or did samsung give it to ya?

[Christopher]

Far to modest my man...I guess it'll take 10 PiB to earn the title of truly epic. lol

Too lazy to search the thread....did you buy this drive 'off the shelf' so to speak or did samsung give it to ya?

It's just a retail box 830... Came with a copy of Arkham City, too. I don't game all that much, but I just recently got around to playing it -- it was moderately enjoyable. To put things in perspective though, BATs 64GB M4 is further along when you take capacity and speed into account... The 830 is four times as fast, but four times as big, too.

[canthearu]

Samsung 840 Day 14.5ish

Drive Hours: 344
ASU GiB Written: 123,784.38
Avg MB/s: 105.20 (37.63 hours)
MD5: OK

Wear Leveling Count (B1): 990 raw (1 normalized)

Reallocated blocks (B3,05): 0 (0 sectors)
Failure count (B5, B6): 0 program, 0 erase
Uncorrectable Error Count: 0
ECC Error Rate (C3): 0

[eddieobscurant]

Why crystaldiskinfo only shows 2tb being written? Is it because the drive is tlc and unsupported or does it count it differently than A.S.U.?

[canthearu]

Why crystaldiskinfo only shows 2tb being written? Is it because the drive is tlc and unsupported or does it count it differently than A.S.U.?

Counter is broken. It counts up fine to 2TiB, then stops counting. Firmware bug.

[eddieobscurant]

Counter is broken. It counts up fine to 2TiB, then stops counting. Firmware bug.

thank you

and thank you all for testing these drives for us. I'm planning to get a 500gb ssd for my laptop, now that the prices are "normal" and i was between the 830(proven stability/endurance) and the 840(faster random reads, minimal slower writes than the 830, for the 500gb version) since the 840 pro is too expensive for me, but i was worried about the endurance of the tlc nand since my laptop will also have some torrents with average 20gb/day , 24/7. But since the 120gb manages more than 120tb, i don't think i'll have a problem with the 500gb for 3-5 years.