SSD Write Endurance 25nm Vs 34nm

[Ao1]

I know there is a way to determine actual WA on a SF drive, but that it's covered under strict NDA (though surely someone has figured it out by now). It's possible that the Chonos D and F3 could have much higher WA than it would appear, and perhaps its hidden away in the Smart data for anyone who knows what they're looking for to see.

Maybe there was a good reason that the original SF drives uses 28 percent OP.

I’m inclined to agree. Compression obviously helps but something else appears to be offsetting the benefit. At a guess it is the fact that SF (unlike anyone else) do not use DRAM.

[minpayne]

The 830 is going down for a retention test as of 10:53PM GMT -6 31 Jan 2012.

Nice contrast against Ao1's! I bet yours is going to be dead for sure if powerless for a week or so, while his would survive the 3-month enterprise standard

[bulanula]

Any news of that Crucial C300 ? Anyone got another one that we can test ?

[bluestang]

Yes, yes they are.

But all of the factory bad blocks were laundered, and thrown back into the good pool, so really I think of it as only 11 erase fails... Based on similar drives, like my Vertex LE 100, 128 Turbo that I've peeped out, it's gonna look just like the original. Which is what I'm pretty sure it is, just tarted up with some new FW and a new plastic dress.

I wonder how it tracks with yours. I think my drive is good bit slower, but yours was run without TRIM for quite some time.

I did look at your drive's last stats and saw that your 1.6FW drive and my 1.7FW are using the same number of PE cycles vs. GB written, just a bit over 56GB/PE cycle.

Ok, for comparison...

Yes, mine ran on XP for 16 days with 53.38 TB written in that time. MWI was 64 and Avg Erase 1830. Ran on Windows 7 from that point on.

I had 8 Initial/Factory Bad Blocks to start with. I threw my first Erase Failure at 87.51 TiB with MWI 49 and Avg Erase at 2571.

At 11 Erase Fails I was at 701.57 TB and Avg Erase 12727. Also Program Fails was at 2.

[Christopher]

Ok, for comparison...

Yes, mine ran on XP for 16 days with 53.38 TB written in that time. MWI was 64 and Avg Erase 1830. Ran on Windows 7 from that point on.

I had 8 Initial/Factory Bad Blocks to start with. I threw my first Erase Failure at 87.51 TiB with MWI 49 and Avg Erase at 2571.

At 11 Erase Fails I was at 701.57 TB and Avg Erase 12727. Also Program Fails was at 2.

...........

Yeah. It could be a refurb, that much is true. But after the factory burn in, by the time the drive had a partition on it, it already had additional blocks bad at runtime. That's never a good sign. I spent some time asking around the OCZ forums, but I didn't get anywhere. People thought that I was just retarded for wanting to know about this drive's behavior. It was at 97MWI out of the box, but just from the couple of erase fails.

I also bought a Vertex EX SLC 64GB from another place at the same time as the Turbo. It was at 89MWI out of the box, and had 209 Initial bad blocks on it's Samsung SLC. But it was on 1.3FW, had a metal case, and some drives did ship with wrong MWI just because of early FW. I dflashed it to 1.7 and it's quite impressive.

[Christopher]

@minipayne,

Seriously, a week? I give it three days at the outside. It's writing about 11600GB a day now. I did some benches, and it's actually hitting 172MBs 4MB writes in ASU. Reads... well, they're getting slower still. I think Samsung accidentally throttled reads instead of writes.

@Ao1,

Given my 120GB Chronos Deluxe's performance profile, it's sustained writes are somewhat ridiculous. I think you could look at it's internal WA, perhaps on a workload basis, or maybe overall, through examination of a variety of attributes (that SF will disclose if you sign their NDA).

The SF processor is doubtlessly doing something fruity internally when bombarded with massive amounts of writes.

[bluestang]

Interesting. Can't wait to see how far your VT goes.

[Anvil]

Kingston SSDNow 40GB (X25-V)

680.89TB Host writes
Reallocated sectors : 05 22
Available Reserved Space : E8 99
POH 6136
MD5 OK

32.79MiB/s on avg (~148 hours)

--

A bit busy these days, will try to catch up later this week.

[Hopalong X]

Anvil

I love it. The old "turtle" just keeps chugging along.

It may be the only one to surpass 1PB by a large margin. Eventually.

[Christopher]

The X25-V has written it's own capacity 17,430.784 times. It's probably over 18,000PE cycles.

[Christopher]

OCZ Vertex Turbo 64 Update Day 15

TiB 94.6472
GiB 96918.77
83.56 MBs Avg

Avg Erase Count
1739
MWI 66 down from 69


1 Program Fail
32 Erase Fail (up from 29)
0 Read Fail
343 Hours




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

I'm taking a trip over the weekend, so the 830 resumes when I get back.

[Anvil]

The X25-V has written it's own capacity 17,430.784 times. It's probably over 18,000PE cycles.

It's somewhat close to 18K times it's own capacity

Let's hope it lasts for a year

[Christopher]

Slow is the way to go.

[Anvil]

The thing is that it really isn't that slow.

I'm using the X25-V as a boot drive on 3 different setups and 99% of the time there are no signs of the capped write speed, random writes are great (for OS usage) and I rarely install apps so copy speed is not an issue
I will however be switching to other drives as they are low on capacity, 64GB or 80GB will do for 2012.

edit:

How are you doing B.A.T, it's been a few days since your last report

[Christopher]

Anvil,

I still use my Vs, even though I have have many faster, larger drives just laying around. And it's not like I have a lot of systems, either; Just a desktop, laptop, and endurance rig.

I am of the opinion that if you want the most TBW, either get a slow writer or artificially cap the write speed in this test. If I capped the 830 at 5TB a day, I think it could make it well over a PB; It would take longer, but you'd get more TBW. You can write faster, but get less TBW.

The X25-M of BATs could last until the next ice age.



-------------------------------------------------------------------------------------------------------------------------
I also have a new MTRON 16GB in on the way, so I hope to have an SLC drive back in the test late next week.

It's a 7000 PRO 16GB 2.5". I have a 16GB 3.5", but it won't really fit in the testing rig. I have pictures of it's pcb in case anyone is dying to see what it looks like inside.

[bulanula]

Try and take out the X25-V and see if it dies just like the others. I have noticed that as soon as you take out these drives and leave them for like a week unpowered the NAND cells loose charge and your SSD becomes a paperweight. So much for the read only after the MWI is over.

[B.A.T]

How are you doing B.A.T, it's been a few days since your last report

I've been on a work trip and I'd planned to update during this but with only edge mobile coverage I couldn't use teameviewer.
On the otherside work was very nice


Todays update.
Kingston V+100


It dropped out again. I'll get it up running when I'm home tonight.

Intel X25-M G1 80GB
308,22 TiB
20700 hours
Reallocated sectors : 00
MWI=121 to 116
MD5 =OK
46.67 MiB/s on avg


m4
354.6262 TiB
1325 hours
Avg speed 81.41 MiB/s.
AD gone from 167 to 148.
P/E 6247.
MD5 OK.
Reallocated sectors : 00

[DragoonXX]

Try and take out the X25-V and see if it dies just like the others. I have noticed that as soon as you take out these drives and leave them for like a week unpowered the NAND cells loose charge and your SSD becomes a paperweight. So much for the read only after the MWI is over.

Biggest thing with these drives is that nobody is stopping at MWI=0 where according to ISO standard- they should retain data for a year (or 3 months for enterprise?) they are going WELL beyond MWI=0. and IIRC, the x25-v went through a week data retention test in the last month and came through with flying colours.

[bulanula]

Biggest thing with these drives is that nobody is stopping at MWI=0 where according to ISO standard- they should retain data for a year (or 3 months for enterprise?) they are going WELL beyond MWI=0. and IIRC, the x25-v went through a week data retention test in the last month and came through with flying colours.

Well I think we ought to test the retention as well not only the write endurance. I bet that the X25-V will be totally dead after one week without any power going to its NAND.

Maybe we should see how much past the MWI you can go while still being able to read your data for 1 year after etc. ?

[Anvil]

bulanula

The thread started out as an Endurance test, not a retention test.
If we hadn't started out the way we started the thread wouldn't have evolved the way it has, awareness of retention testing is one if the things that has come out of the testing, so far.

There is some retention testing (have you stopped reading the thread or is it just bits), that's how several of the contestants dropped out of the test.
Ao1 is having a 3 month retention test as well.

As for the 1 year retention test, just bring me a time machine and I'll be happy to do some testing

If you'd like to enter the test the price of the SSD is just petty cash, the commitment lies in spending the hours and for some of the contestants using their main rig for testing so please don't use "we" until you actually start testing some drive.

[Hopalong X]

Anvil

Wish I had access to a "Thank You" button.

Never have figured out why I don't have one.

[Ao1]

Biggest thing with these drives is that nobody is stopping at MWI=0 where according to ISO standard- they should retain data for a year (or 3 months for enterprise?) they are going WELL beyond MWI=0. and IIRC, the x25-v went through a week data retention test in the last month and came through with flying colours.

Come back on the 26th April and you will see if the 830 has been able to retain data after three months of no power after getting to MWI 1.

[Anvil]

Anvil

Wish I had access to a "Thank You" button.

Never have figured out why I don't have one.

Thanks!

As for the "Thank You" button, you need 100 posts IIRC.

[Christopher]

B.A.T.,

Nice picture!

[B.A.T]

Thank you!
Some days winter shows herself from her best side