SSD Write Endurance 25nm Vs 34nm

[Christopher]

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

It maintains 340 to 350 for most loops now, which makes me concerned that it's getting faster. I remember 340 being the upper ceiling a few months back.

[minpayne]

It maintains 340 to 350 for most loops now, which makes me concerned that it's getting faster. I remember 340 being the upper ceiling a few months back.

What's the maximum temperature your 830 gets during the test?

My new 830 512GB does around 275MB/s average, and 295MB/s instantaneous. Why is it slower than the 256GB model?

[Christopher]

My 256GB under test runs between 50c and 54c. The 512GB is slightly slower in some places, slightly faster In others compared to the 256GB. They share the same number of dice, but the 512GB uses 64Gbit dice while the rest use 32gbit dies. I haven't run the 512GB for any length of time under endurance testing though.

[canthearu]

Sandisk Extreme G25 - 120gb - Day 96

Drive hours: 2239
ASU GiB written: 959,599.64 GiB (937.11 TiB)
Avg MB/s: 125.13 MB/s (39.87 hours)
MD5: OK

Host GB written (F1): 965,764 GiB (943.13 TiB)
NAND writes (E9): 721,201 GiB (704.30 TiB)

Retired Block Count (05): 4 raw
Failure count (AB, AC): 1 program, 1 erase
Raw Error Rate (01): 110 normalized
Media Wearout Indicator (E7): 10 normalized
Wear Range Delta (B1): 4 raw
Reported Uncorrectable Errors (BB): 0 raw

[Anvil]

The 330 has been acting up and so the X25-V is running solo. (the 330 is idle)

I'l get back to tonight with a proper report. (been travelling a lot lately)

[canthearu]

Sandisk Extreme G25 - 120gb - Day 98

Drive hours: 2288
ASU GiB written: 981,462.60 GiB (958.46 TiB)
Avg MB/s: 125.41MB/s (81.37 hours)
MD5: OK

Host GB written (F1): 987,735 GiB (964.58 TiB)
NAND writes (E9): 737,627 GiB (720.34 TiB)

Retired Block Count (05): 4 raw
Failure count (AB, AC): 1 program, 1 erase
Raw Error Rate (01): 112 normalized
Media Wearout Indicator (E7): 10 normalized
Wear Range Delta (B1): 4 raw
Reported Uncorrectable Errors (BB): 0 raw

Will post new graphs in a few minutes

[canthearu]

Bat's poor X25-M 80GB has the final stats on death. I'll update other drives if people post their updates!

Edit 1: Updated BAT's Drives

[B.A.T]

The 330 has been acting up and so the X25-V is running solo. (the 330 is idle)

I'l get back to tonight with a proper report. (been travelling a lot lately)

Good to have you back. We need someone to keep us in line and away from trouble

[canthearu]

Good to have you back. We need someone to keep us in line and away from trouble

Lol, true that, welcome back anvil!

[B.A.T]

Here are todays update:

m4
1570.9176 TiB
6491 hours
Avg speed 75.06 MiB/s.
AD 232 to 223
P/E 27047
C3 15808 to 26044
01 6 to 10
CE 90
MD5 OK.
Reallocated sectors : 00


Intel X25-E 64GB
932.79 TiB
2957-30=2927 hours
Reallocated sectors : 7 to 32
Available Reserved space: 92 to 91
MWI= 94 to 93
MD5 =OK
86.57 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.

[Anvil]

Status of the Intel 330 120GB

831.75TB Host writes
2.81TiB Host reads
Reallocated sectors : 05 1184 +797 (387)
Available Reserved Space : E8 100 ""
MWI 1 (9)
[B5] Program fail count 563 (186)
[B6] Erase fail count 621 (201)
[F1] Total LBAs Written 27254740
[F2] Total LBAs Read 86201
[F9] Total NAND Writes 600556GB // ~586TiB
POH 2605 (1824)
MD5 --

---.--MiB/s on avg (hours)

It's not running any tests (just idling), will have a closer look at it this weekend, most likely on a different computer.

--

Kingston SSDNow 40GB (X25-V)

1210.15TB Host writes (39654321*32)
Reallocated sectors : 05 78 +5
Available Reserved Space : E8 98
POH 10856
MD5 OK

33.15MiB/s on avg (~60 hours)

--

I'll have a look at the events logged by SMARTLog, they should be telling some of the story.

[Christopher]

Samsung 830 256GB Day 172

(GiB) 4,349,371
(TiB) 4,247
(PiB) 4.17


(Avg) 297.64 MB/s over the past 1400+ hours

(B1) Wear Leveling Count: 18,816

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

(POH) 4269


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

The 830 is boring me.

[Christopher]

So I think I'm going to try another drive -- but this time, I'm goinng to do full span random testing with either 4K or 8K. And it will probably be a drive already tested by me or someone else.

[canthearu]

Just a graph update since Anvil and Christopher did updates.

[Hopalong X]

So I think I'm going to try another drive -- but this time, I'm goinng to do full span random testing with either 4K or 8K. And it will probably be a drive already tested by me or someone else.

Do not use the Tesla coil on the new SSD as the first test this time.

Tesla coil is the last test.

[canthearu]

Woohoo, 1 million GB written!

[B.A.T]

Here are todays update:

m4
1578.1809 TiB
6519 hours
Avg speed 75.07 MiB/s.
AD 223 to 219
P/E 27170
C3 26044 to 28392
01 10 to 11 (100 to 99)
CE 90
MD5 OK.
Reallocated sectors : 00


Intel X25-E 64GB
941.05 TiB
2985-30=2955 hours
Reallocated sectors : 32 to 44
Available Reserved space: 91 to 90
MWI= 93
MD5 =OK
86.39 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]

Do not use the Tesla coil on the new SSD as the first test this time.

Tesla coil is the last test.

So many rules... If I call it a death ray, does that change anything?

[Hopalong X]

So many rules... If I call it a death ray, does that change anything?

Sparky

The Death Ray is only to be used for SE'ing uncooperative test subjects.

[Christopher]

Sparky

The Death Ray is only to be used for SE'ing uncooperative test subjects.

The Death Ray is really just a permanent SE. I should license my Death Ray (D-Ray for short) to some of these military spec SSD vendors! I'll make millions!

[minpayne]

The Death Ray is really just a permanent SE. I should license my Death Ray (D-Ray for short) to some of these military spec SSD vendors! I'll make millions!

So the fastest way to wear out the endurance of an SSD is to write a program to repeatedly invoke Secure Erase on it, which translates into a couple of seconds per P/E cycle?

[Christopher]

So the fastest way to wear out the endurance of an SSD is to write a program to repeatedly invoke Secure Erase on it, which translates into a couple of seconds per P/E cycle?

Not exactly... The fastest way to kill a drive is to short it out. I shorted out an MTRON early this year doing some dumb stuff, and that's what Hopalong was talking about.

[canthearu]

So the fastest way to wear out the endurance of an SSD is to write a program to repeatedly invoke Secure Erase on it, which translates into a couple of seconds per P/E cycle?

No, because secure erase would only erase the NAND that was currently used on the drive. If none of the NAND is allocated, then secure erase wouldn't do much.

[canthearu]

The Sandisk Extreme crashed today ... probably a sign that this drive is nearing the end of it's life



Restarting the computer has allowed it to continue working, but it is probably on borrowed time

[kaktus1907]

^^ 750GB total nand writes, Plextor did get premium nands from Toshiba then for extra ~100GB writes then lol..