SSD Write Endurance 25nm Vs 34nm

[Christopher]

That G1 went out like a boss. 10.5 months of no TRIM is hardly a small victory. In reality, it must have written 5x that much internally. If it had TRIM.......

[Christopher]

Samsung 830 256GB Day 164

(GiB) 4,150,202
(TiB) 4,052
(PiB) 3.98


(Avg) 297.41 MB/s over the past 1300+ hours

(B1) Wear Leveling Count: 17,958

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

(POH) 4098


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

[canthearu]

Yeah, the G1 went pretty well

I just did 24h retention test on sandisk ... didn't worry it at all.

[Zaxx]

I'm sure the Sammy 830 has passed the 4PB mark by now...hats off to Samsung for one solid and steady controller and to Chris for puttin' the screws to it in the quest(s) to see which drives go above and (waaay) beyond expectations. Altho the controller does have a lot of nand to work with vs other drives tested...I'm betting the 128GB and 64GB do just as well with respect to those drive sizes. Getting 2PB out of a 64GB Sammy wouldn't be a huge shocker, tho it'd be damn impressive.

[B.A.T]

Here are todays update:

m4
1531.6987 TiB
6338 hours
Avg speed 74,97 MiB/s.
AD 250 to 245
P/E 26380
C3 7907 to 9220
01 3
CE 90
MD5 OK.
Reallocated sectors : 00


Intel X25-E 64GB
887.50 TiB
2805-30=2775 hours
Reallocated sectors : 205 to 219
Available Reserved space: 93
MWI= 94
MD5 =OK
87.64 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.

[Christopher]

Samsung 830 256GB Day 165

(GiB) 4,184,385
(TiB) 4,086
(PiB) 4.01


(Avg) 297.20 MB/s over the past 1300+ hours

(B1) Wear Leveling Count: 18,106

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

(POH) 4114


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

And that is 4PiB! Haven't had an erase failure in over 2 PiB.

[Hopalong X]

Congrats to Sam(sung) and Chris.

[B.A.T]

Congratz Chris

Next stop 6 PiB.

[Christopher]

Next stop could be never. At this rate, I'll be posing updates for the next 73 years.

[Zaxx]

And that is 4PiB! Haven't had an erase failure in over 2 PiB.

The 830 has got to be the most stable/dependable ssd available. I know that sammy uses their own controller and nand...what process node is the nand? I'm sure it's sync or maybe toggle...probably 2x nm I'm guessing.

[ice_chill]

I am surprised that such a drive can cost so competitively, could it even match SLC ?

[UranusFX]

what the ... ultra durable toughness drive like X25-V!
830's NAND is truly Toggle MLC.

Grateful to Christopher

[canthearu]

Sandisk Extreme G25 - 120gb - Day 93

Drive hours: 2171
ASU GiB written: 929,637.00 GiB (907.85 TiB)
Avg MB/s: 125.54 MB/s (50.60 hours)
MD5: OK

Host GB written (F1): 935,659 GiB (913.73 TiB)
NAND writes (E9): 698,702 GiB (682.33 TiB)

Retired Block Count (05): 4 raw
Failure count (AB, AC): 1 program, 1 erase
Raw Error Rate (01): 109 normalized
Media Wearout Indicator (E7): 10 normalized
Wear Range Delta (B1): 3 raw
Reported Uncorrectable Errors (BB): 0 raw

May I dare dream of having a drive reach 1 PiB (well, without having to do something dumb like breaking out an Intel 510 or something!)

[Christopher]

Since it came up, the 830 is using 27nm Samsung 1.0 Toggle. But IMHO the M4 getting 1.5PiB at 64GB is far more impressive... But it's only a matter of time before the 830 gets there.

[Christopher]

May I dare dream of having a drive reach 1 PiB (well, without having to do something dumb like breaking out an Intel 510 or something!)

Now you know how I felt. Finally, I just said "to hell with the dumb stuff," and got a 256GB 830.

[minpayne]

Since it came up, the 830 is using 27nm Samsung 1.0 Toggle. But IMHO the M4 getting 1.5PiB at 64GB is far more impressive... But it's only a matter of time before the 830 gets there.

IMO the Samsung 830 is far more impressive, as it has a lot more flash chips inside! The failure of each flash chip is statistically independent, which means it's more difficult to get a large drive with good quality flash chips together, than to get a small drive. Think of the cost of display/monitor - the price does not grow linearly with respect to area, but grows exponentially!

[Christopher]

IMO the Samsung 830 is far more impressive, as it has a lot more flash chips inside! The failure of each flash chip is statistically independent, which means it's more difficult to get a large drive with good quality flash chips together, than to get a small drive. Think of the cost of display/monitor - the price does not grow linearly with respect to area, but grows exponentially!

Actually, that's a good point. The endurance of the drive is governed by the worst NAND die, so the probability of one die failing is 4x greater than a 64GB 830 -- which have 2x the dice of ONFI drives with 64gbit dice.

[ZOMGVTEK]

I don't think you can typically expect the drive endurance to increase linearly with capacity in real world testing, but I'm no expert. If you ask me the 830 is the most impressive drive to date. 4PiB is no joke.

Just wanted to comment that this thread made me go run out and grab an 830 for my notebook. It's performing well, even on my ancient 2011 SATA II hardware. My only complaint is that it gets the palm area where the drive resides considerably warmer than before. I was writing 150GB or so to the drive at 200+MB/s, but I never noticed that area get that warm before.

[B.A.T]

Here are todays update:

m4
1541.8668 TiB
6378 hours
Avg speed 74,88 MiB/s.
AD 245 to 239
P/E 26553
C3 9220 to 11450
01 3 to 5
CE 90
MD5 OK.
Reallocated sectors : 00


Intel X25-E 64GB
899.39 TiB
2844-30=2814 hours
Reallocated sectors : 219 to 237
Available Reserved space: 93 to 92
MWI= 94
MD5 =OK
87.36 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.

[Christopher]

The 830 gets warm; under prolonged writing mine stays around 52c, and it's in open air. The 830 is great for laptops since it only draws .38w at idle -- though load consumption is as much a
6w for the larger capacities.

[B.A.T]

Here are todays update:

m4
1547.8740 TiB
6401 hours
Avg speed 74,90 MiB/s.
AD 239 to 236
P/E 26656
C3 11450 to 11796
01 5
CE 90
MD5 OK.
Reallocated sectors : 00


Intel X25-E 64GB
906.38 TiB
2868-30=2838 hours
Reallocated sectors : 237 to 248
Available Reserved space: 92
MWI= 94
MD5 =OK
87.19 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.

[Fackamato]

Great thread.

At 08:42 this morning the G1 pulled his last breath and went away silently. It just couldn't take any more after 10.5 months of torture. I tried every trick in the book but I can't get it to connect again.

R.I.P Intel X25-M G1 80GB

Is it possible to run secure erase on the X25-M G1? (from another host)
If so, what would happen, would it be saved in some way?

[B.A.T]

The drive is undetected in bios. Tried it from 3 different machines. The G1 is truly dead.

[Christopher]

Samsung 830 256GB Day 168

(GiB) 4,265,360
(TiB) 4,165
(PiB) 4.09


(Avg) 297.53 MB/s over the past 1400+ hours

(B1) Wear Leveling Count: 18,454

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

(POH) 4198


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

The 830's instantaneous estimated speed is 350MB/s for most of the loop! If it weren't for the average being almost 1500 hours, avg/MBs would most likely be quite high.

[B.A.T]

350 MiB/s write speed after 4 PiB is ridiculous fast and impressive.

Here are todays update:

m4
1554.0292 TiB
6425 hours
Avg speed 74,93 MiB/s.
AD 236 to 232
P/E 26760
C3 11796 to 15808
01 5 to 6
CE 90
MD5 OK.
Reallocated sectors : 00


Intel X25-E 64GB
913.48 TiB
2892-30=2862 hours
Reallocated sectors : 248 to 255 to 7
Available Reserved space: 92
MWI= 94
MD5 =OK
87.01 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.