Samsung 830 256GB Day 112
(GiB) 2,727,423.40
(TiB) 2,663.50
(Avg) 301.37MB/s
(B1) Wear Leveling Count: 11,823
(B6) Erase Fail Count: 9
(05) Reallocated Sectors: 36864
(POH) 2715
-------------------
Printable View
Samsung 830 256GB Day 112
(GiB) 2,727,423.40
(TiB) 2,663.50
(Avg) 301.37MB/s
(B1) Wear Leveling Count: 11,823
(B6) Erase Fail Count: 9
(05) Reallocated Sectors: 36864
(POH) 2715
-------------------
About 5 hours ago the 330 reported 2 reallocated sectors
B5 and B6 both changed from 0 to 1. (not sure what they are, will check up, both are listed as unknown)
--
Intel 330 120GB
460.54TB Host writes
1.77TiB Host reads
Reallocated sectors : 05 2
Available Reserved Space : E8 100
MWI 47
[B5] 1
[B6] 1
[F1] Total LBAs Written 15090825
[F2] Total LBAs Read 57959
[F9] Total NAND Writes 332438GB
POH 1130
MD5 OK
124.74MiB/s on avg (~48 hours)
--
Kingston SSDNow 40GB (X25-V)
1104.09TB Host writes (36178803*32)
Reallocated sectors : 05 54
Available Reserved Space : E8 99
POH 9863
MD5 OK
34.34MiB/s on avg (~48 hours)
Todays update:
m4
1186.9700 TiB
4959 hours
Avg speed 73.56 MiB/s.
AD 190 to 187
P/E 20446.
C3 4312 to 4457 (after firmware update)
CE 58
MD5 OK.
Reallocated sectors : 00
http://www.diskusjon.no/index.php?ap...tach_id=496978
Intel X25-M G1 80GB
778.54 TiB
24247 hours
Reallocated sectors : 188 to 199
Available Reserved space: 57
MWI= 226 to 219
MD5 =OK
33.41 Mi/s on avg
http://www.diskusjon.no/index.php?ap...tach_id=496976
Intel X25-E 64GB
459.46 TiB
1428-30=1375 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI= 97 to 96
MD5 =OK
91.73 MiB/s on avg
http://www.diskusjon.no/index.php?ap...tach_id=496982
Mtron Pro 7025 32GB
ASU write 99.3457 TiB
Host write 187,7499 TiB
Total host write 235,2854 TiB
Day 37
MWI= 93
MD5 =OK
47.83 MiB/s on avg
http://www.diskusjon.no/index.php?ap...tach_id=496980
Changes since last update:
-The Mtron dropped out 1 time. I'll be pausing the endrurance test on the Mtron until Iæll figured out how to update firmware on it. I need to find some documentation first.
- I'll updatet to the lastest firmware on the m4 to avoid the 5K bug. The update went well and the drive runs normal again. Only change was a smal increase in Cumulative Corrected ecc because it was without power for aprox 1 hour.
http://www.diskusjon.no/index.php?ap...tach_id=496983
@Anvil
Can you update the first post with the new firmware on the m4?
@B.A.T
Will do
@christopher
What's the final verdict on the Plextor?
--
Attachment 128269
The M3P is dead. I couldn't get it running for more than a few moments; It's quite curious actually.
However, today I tried the drive out, and all the data was there, I grabbed some more smart data (a lot more reallocations were seen vs. the last data I posted). Now the drive can be detected by the UEFI, but no dice. It will hang the system now, so looks like it's all gone.
Here is the data I grabbed before death:
Attachment 128288
Attachment 128289
896.699 TiB in host writes. I did manage to figure out what all of the smart attributes are though, so that's a plus.
I really liked this drive, it's sad to see it go. It probably used over 7800 PE cycles, and while I was hoping for more, the drive is just way to damn fast.
sent my contact at Plextor the above info and 3 days later I get this reply
Hi *****,
Plextor SSDs have been through testings by our RDs. For heavy users, the maximum could reach to 20G/ day. It can write about 36.5 TB/ 5 years. TiB written 856 could be used for 115 years.
- Plextor
I also linked then the thread, I'm quite disappointed in the short time it took to get the drive to fail it did go fast while it was kicking.
I know it is disappointing to see a drive die early ... but it did do twice the specified P/E cycles of the NAND.Quote:
Originally Posted by Christopher
So a fairly average result ... but not bad at all in real world terms :)
I wouldn't call it disappointing, shrinking leads to fewer P/E cycles, binning/QC can make a difference.
MWI behaves differently from other drives though, could be that this is just as good as being based on "P/E count".
I'd say it did well!
--
The 330 lost 2 more sectors, current count is up from 2 to 4. (happened during the night)
B5 and B6 looks to go hand in hand with 05 and are both up to 2.
I actually was speaking to a intel rep and showed them this thread and they said they would closely follow this thread and investigate the quick failure rate of the 520 series drive.
Well, I still have that Intel 520, your rep is welcome to contact me regarding it (and getting it back).
I am intrigued about the reasons behind the panic locking death of my Intel 520.
I was only dissappointed with the M3P because I really thought it would reach 1PIB. It was using twice the number of PE cycles as the 830 with half as much flash. It's still a fantastic drive.
--Good thing I have another!
I don't know whether the MWI on the Plextor was spot on or whether it was just a coincidence. I don't think it was the erase failures, but it was the read failures that did it in (IMHO). MWI is based on reallocation events, but it only had 15 read failures.
That makes two of us...maybe they'll bite.Quote:
Originally Posted by canthearu
Intel 330 120GB
475.67TB Host writes
1.81TiB Host reads
Reallocated sectors : 05 4
Available Reserved Space : E8 100
MWI 45
[B5] 2
[B6] 2
[F1] Total LBAs Written 15586863
[F2] Total LBAs Read 59207
[F9] Total NAND Writes 343364GB
POH 1165
MD5 OK
125.04MiB/s on avg (~35 hours) // looks like there is a small increase in write speed
--
Kingston SSDNow 40GB (X25-V)
1108.31TB Host writes (36317077*32)
Reallocated sectors : 05 54
Available Reserved Space : E8 99
POH 9898
MD5 OK
34.77MiB/s on avg (~35 hours)
Sandisk Extreme G25 - 120gb - Day 37
Drive hours: 885
ASU GiB written: 381,370.22 GiB (372.43 TiB)
Avg MB/s: 125.23 MB/s (322.50 hours)
MD5: OK (verified off disk now)
Host GB written (F1): 383,510 GiB (374.52 TiB)
NAND writes (E9): 286,169 GiB (279.46 TiB)
Retired Block Count (05): 4 raw
Failure count (AB, AC): 1 program, 1 erase
Raw Error Rate (01): 118 normalized
Media Wearout Indicator (E7): 24 normalized
Wear Range Delta (B1): 4 raw
Reported Uncorrectable Errors (BB): 0 raw
Anvil: On my sandisk extreme
B5 = Program Fail Count
B6 = Erase Fail Count
So I'm pretty sure that is what your Intel 330 is measuring too (given it is a sandforce based drive)
Samsung 830 256GB Day 114
(GiB) 2,718,314.72
(TiB) 2,716.13
(Avg) 300.65MB/s
(B1) Wear Leveling Count: 12,055
(B6) Erase Fail Count: 9
(05) Reallocated Sectors: 36864
(POH) 2767
-------------------
@canthearu
You may be right, although Intel did rearrange quite a few attributes.
Intel 330 120GB
491.36TB Host writes
1.85TiB Host reads
Reallocated sectors : 05 6
Available Reserved Space : E8 100
MWI 43
[B5] 3
[B6] 3
[F1] Total LBAs Written 16101001
[F2] Total LBAs Read 60503
[F9] Total NAND Writes 354688GB
POH 1201
MD5 OK
125.06MiB/s on avg (~71 hours)
--
Kingston SSDNow 40GB (X25-V)
1112.46TB Host writes (36453240*32)
Reallocated sectors : 05 54
Available Reserved Space : E8 99
POH 9934
MD5 OK
33.95MiB/s on avg (~71 hours)
--
Attachment 128352
The Plextor is shown one last time as TiB written was quite a bit up from the previously reported value.
Samsung 830 256GB Day 115
(GiB) 2,811,688.22
(TiB) 2,745.79
(PiB) 2.70
(Avg) 302.38 MB/s
(B1) Wear Leveling Count: 12,186
(B6) Erase Fail Count: 9
(05) Reallocated Sectors: 36864
(POH) 2795
-------------------
The 830 has been averaging over 300MB/s for the past four days! Crazy.
If I were to run another 128GB M3P but artificially reduce PE/day, I think it would make it to 1.5PiB.
I've been thinking about making an option for reducing throughput, not set yet but it might be interesting.
Not sure if the Plextor fits the "Drives that did not meet MWI spec" chart as MWI is handled differently, I'm leaning towards putting it in the standard chart.
Attachment 128354
Attachment 128355
Great work. I find this thread fascinating. Also I thought Intel 520 would go on forever, but it looks like the Sandforce controller is a let down, as it died without even reporting relocated sector counts first.
Don't forget though, the 520 is a desktop drive based on MLC. It won't receive the absolute best NAND, as it has already been turned into eMLC for their enterprise products. It's tested life, at roughly twice the specifications that Intel put on the nand (10k cycle real life, 5k cycle rated life) was still quite good.Quote:
Originally Posted by ice_chill
It's far better than OCZ's NAND, but compared to Samsung 830 it's dissapointing for the reason that the Intel SSD costs considerably more than Samsung 830.Quote:
Originally Posted by canthearu
Lets not get ahead of ourselves, the samsumg hasn't even reached the same amount of writes as the Intel 520 yet ... at least on a per gig of accessable space basis. The samsung drive is 256gig, the intel was a 60gig.Quote:
Originally Posted by ice_chill
A bit early to be decrying how much the Intel drive sucks, especially when there isn't a hope in hell that any desktop user will reach the MWI on either the intel or the samsung.
In any case, both the intel 520 and the samsung 830 is a great drive.