Printable View
Well, maybe in a few years one could get a 512GB drive at a reasonable price for using in this thread :)
Are there any reviews on the Phison that you have noted?
(a few benchmarks could tell if it's worth spending time and money on something using that controller)
I checked in on the 330, not much have changed except for TiB written, it's down to 122.x MB/s over a 12-14 hour period.
Todays update:
m4
932.0644 TiB
3929 hours
Avg speed 74.35 MiB/s.
AD gone from 80 to 77.
P/E 16059.
MD5 OK.
Reallocated sectors : 00
http://www.diskusjon.no/index.php?ap...tach_id=491903
Intel X25-M G1 80GB
644.13 TiB
23241 hours
Reallocated sectors : 159 to 161
Available Reserved space: 90
MWI=124
MD5 =OK
42.66 MiB/s on avg
http://www.diskusjon.no/index.php?ap...tach_id=491907
Intel X25-E 64GB
104.83 TiB
328-30=298 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI=99
MD5 =OK
101.16 MiB/s on avg
http://www.diskusjon.no/index.php?ap...tach_id=491905
No movement on mwi for the E even after 100TiB of writing, wonder when the first drop will come. Besides that all 3 have been very stabile the last 12 days. Writespeeds are barely moving so it looking good. :)
BIOS update may have resolved my stability problems (I was running a VERY early bios version, the very initial version), running linx in background to check.
Edit: Vertex 4 at 214 reallocations. Still working though
That's good news, using the initial bios version often is risky business.
214 reallocations, that should translate to 214 bad blocks, not sure how many MBs per block on your drive. (there should be plenty left)
The Crucial m4s are showing factory bad block count which in some cases (on some of my drives) exceeds your current bad block count. (if we assume 1 reallocation consumes 1 block)
Yeah, I jumped on the Z77 chipset very early because I wanted Intel's USB3 rather then a 3rd party chipset. But things are starting to be ironed out now:Quote:
Originally Posted by Anvil
a) Updated BIOS seems to have fixed stability.
b) Updated Intel graphics drivers have helped.
c) Updated RST drivers (11.1 rather then 11.0) means smartlog is working right now.
d) Updated crystaldiskinfo now shows thresholds like it should.
But one good thing is that boot times are excellent!
holy crap is that SLC flash?Quote:
Originally Posted by Christopher
Kingston SSDNow 40GB (X25-V)
980.23TB Host writes
Reallocated sectors : 05 43
Available Reserved Space : E8 99
POH 8782
MD5 OK
34.47MiB/s on avg (~24 hours)
--
Intel 330 120GB
10.27TB Host writes
Reallocated sectors : 05 0
Available Reserved Space : E8 100
[F1] Total LBAs Written 336465
[F2] Total LBAs Read 14536
[F9] Total NAND Writes 7396GB
POH 58
MD5 OK
122.72MiB/s on avg (~24 hours)
Here's the initial benchmarks, FOB performance.
Attachment 127063
Attachment 127064
Attachment 127065
No, that's Samsung Toggle mode NAND.Quote:
Originally Posted by Cyber-Mav
The capacity of the drive leads to those figures.
No, it's just 256GB of thunder and lightning. It's comparable to a 64GB drive with 409TiB (~419,580 GiB). Not terrible, but not earth shattering either. At least, not yet.Quote:
Originally Posted by Cyber-Mav
I got tired of the drives I was running not putting in the kind of performance I was hoping for, so I had to reach for a bigger hammer in the 830 256GB.
does toggle mode nand as in the plextor m3 have more durability than regular nand such as that in the crucial m4?
Not necessarily; they're still governed by the same principles which limit endurance. With that said, the companies which make NAND and SSDs tend to find the best flash out of their fab facilites to put in their own drives. Intel/Micron(Crucial) and Samsung make flash, and this flash can go either in their own products, or get sold to other companies. Usually, they keep the best stuff for themselves. (Note that Toshiba does make flash, and does make drives, but they primarily are OEM units and some enterprise oriented offerings, but the same still probably applies to them)Quote:
Originally Posted by Cyber-Mav
The Plextor M3P is a fine product, but it might actually have less endurance than the M4. Perhaps we'll find out soon, I think I'm going to test one.
Samsung 830 256GB Day 67
1,603,779.67GiB
1,566.19 TiB
288.07 MB/s Average
1590 Hours
6967 Wear Leveling Count
MWI 1
6/0 Erase/Program Fail
Used Reserved Block Count: 12/24572 sectors
I stopped the 830 and migrated to RC, all is well.
Vertex 4 looks dead:
Attachment 127104
Smart values at end of life:
Attachment 127105
Drive has now disappeared from windows. I'll attempt to resurrect.
[device is not ready] never was a good sign!
It's worth a bit of effort of course, it was showing the signs of a dying drive...
At 61% remaining life it is quite the disappointment. (if this is it)
Well, I got it running again. It is now Zombie Vertex 4.
A couple of restarts got it re-detected.
Remaining life is probably closer to 31% given the drive life reset when I upgraded firmware. In any case, not good at all.
Zombie Vertex 4 stopped again ... same position and error.
Wondering if I should try secure erase.
This sounds very familiar... See the two VTs and the Octane...Quote:
Originally Posted by canthearu
~30% is better, nevertheless, Remaining Life is there for a reason.
What about giving it a some extra over-provisioning?
Todays update:
m4
938.1534 TiB
3953 hours
Avg speed 74.35 MiB/s.
AD gone from 77 to 74.
P/E 16162.
MD5 OK.
Reallocated sectors : 00
http://www.diskusjon.no/index.php?ap...tach_id=491993
Intel X25-M G1 80GB
647.67 TiB
23265 hours
Reallocated sectors : 161
Available Reserved space: 90
MWI=124
MD5 =OK
42.71 MiB/s on avg
http://www.diskusjon.no/index.php?ap...tach_id=491997
Intel X25-E 64GB
113.14 TiB
352-30= 322 hours
Reallocated sectors : 0
Available Reserved space: 100
MWI=99
MD5 =OK
101.12 MiB/s on avg
http://www.diskusjon.no/index.php?ap...tach_id=491995
Attachment 127106
So much for the marketing gimmick otherwise known as the Ndurance 2.0. Is Ndurance just a throttling process? Surely someone has a grasp by now on what is going on with the V4 sustained write speeds?
That is what I've been thinking for the last week.Quote:
Originally Posted by Ao1
Ndurance 2.0 = almost lasts as long as the NAND is designed for.
What is even more confounding is how they are making the marvell controllers perform so poorly for endurance when the they have typically experienced perfectly fine endurance control.
It's all about the flash. Period.
For sure it would appear from testing in this thread that the grade of NAND used in Intel/ Micron/ Samsung SSD’s is superior to what others like $OCZ can buy on the open market, but that does not explain everything. So far the 520 has behaved much better than any other SF based drive. The premature block failures of the Octane/ Petrol drives are occurring before the P/E cycles have hardly been touched. It’s almost like something is accelerating/ inducing damage to the NAND. (Maybe it’s the overclock).
So what gives with the V4? Is it a form of throttling that causes write speed to drop by half? Someone must know, but no one is saying anything. If it was an Intel/ Micron/ Samsung/ Toshiba product the world and his dog would be screaming about it from the rooftops.
They are still sceaming trying to get their Vertex 2's to work. :shocked: :ROTF:Quote:
Originally Posted by Ao1
Kingston SSDNow 40GB (X25-V)
982.99TB Host writes
Reallocated sectors : 05 45 up 2
Available Reserved Space : E8 99
POH 8806
MD5 OK
33.84MiB/s on avg (~48 hours)
--
Intel 330 120GB
20.37TB Host writes
Reallocated sectors : 05 0
Available Reserved Space : E8 100
[F1] Total LBAs Written 667628
[F2] Total LBAs Read 14536
[F9] Total NAND Writes 14696GB
POH 81
MD5 OK
122.49MiB/s on avg (~48 hours)