SSD Write Endurance 25nm Vs 34nm

[B.A.T]

The spec sheet says 16500 IOPS 4K seq writ and 12000 IOPS 4K RW.

[bluestang]

I've been thinking of another drive to test with, so I'm taking suggestions.

What about the Vertex Plus?, the 120GB is cheap and the Indilinx was/is a pretty good sequential performer.
(not sure about the current firmware though)

I think a new Indilinx would be a nice candidate. It's not another SF based drive and I'm curious to how it performs compared to old.

[Christopher]

I was even mulling over an overprovisioned 470...

The new vertex plus is not spectacular. The way I understand it, a vertex plus 60 is equivalent to a 30GB Vertex Original flavor.

The 60GB Vertex Plus weighs in at 180 reads and 90MBs writes. The 120GB would be closer to a 60GB 1G Indilinx in performance.

[Anvil]

Kingston SSDNow 40GB (X25-V)

406.67TB Host writes
Reallocated sectors : 12

MD5 OK

35.38MiB/s on avg (~22 hours)

--

Corsair Force 3 120GB

01 88/50 (Raw read error rate)
05 2 (Retired Block count)
B1 67 (Wear range delta)
E6 100 (Life curve status)
E7 48 (SSD Life left)
E9 201786 (Raw writes)
F1 268723 (Host writes)

106.19MiB/s on avg (~22 hours)

power on hours : 793

edit:

@Christopher

I had a look at it, that's why I suggested the 120GB.
An over-provisioned drive would be great, well, you've still got some time to think about it.

[Christopher]

Mushkin Chronos Deluxe 60 Update, Day 28

05 2
Retired Block Count

B1 24
Wear Range Delta

F1 265051
Host Writes

E9 204440
NAND Writes

E6 100
Life Curve

E7 10
Life Left

Average 127.61MB/s Avg
Intel RST drivers, Asus M4g-z

642 Hours Work (24hrs since the last update)
Time 26 days 18 hours

6 GiB Minimum Free Space


The 470 has an updated version. I'm not positive, but I think it uses the new Samsung 2xnm NAND, so once the old versions go out of stock they'll be hard to find. If a drive should be chosen to test with overprovisioning, I was thinking it should be the same model of another test participant.

I have one of those new SB Celeron Dual Cores on the way to set up an "under the couch rig". I may put the Mushkin on that, or keep it on my main system and run another drive on that eventually.

Still not so much as a peep on the new SF firmware.

[A||uSiOn]

Hey guys after reading a load of pages I have come to one conclusion, the Crucial M4 is one of the most reliable drives or is the most reliable drive out of the lot?

We are about to build a new media server with 16 x SSD's 256GB , now you guys go me thinking on which drives to select.

[One_Hertz]

516TB. 6556 reallocated sectors.

Speed used to peak at 46.5MB/s when the SSD was brand new. Then the speed started rising and it peaked at around 48.5. Now, it is 41.5 max. Really slowing down...

[Christopher]

1_Hz,

Are you using Beta9, or are you still on Beta8?

If the 320 keeps falling in speed, it might last forever if it drops down to 1mb/s... maybe it's Intel's revenge.

[Christopher]

Hey guys after reading a load of pages I have come to one conclusion, the Crucial M4 is one of the most reliable drives or is the most reliable drive out of the lot?

We are about to build a new media server with 16 x SSD's 256GB , now you guys go me thinking on which drives to select.

Well, I'm not really sure what your story is, but if it's a media server, you'd best be served my a non-SF drive. I don't really see how you could go wrong with an M4 (or other Marvell controlled drives like the Intel 510/C300/Plextor M2/Performance3). Putting aside the reliability issues associated with SandForce drives, compressed media files will be slower on SF drives than non compressed/non random files. The penalty is most severe for asynchronous NAND drives (Corsair Force 3, Patriot Pyro, OCZ Agility 3). It's less severe for synchronous NAND SF drives (OCZ Vertex 3, Corsair Force GT). The Toggle NAND SF2200 drives are the best in my opinion, but if you're really serious about putting 16 SATA III SSDs in a server, the Crucial M4 is pretty damn good. But they make a reasonably priced 512GB drive, so maybe just get 8 of those and call it a day. I'm not really sure why a media server needs 16 SSD (some regular-ass-HDDs would probably give you the space and sequential speeds for a media server, but whatever -- it's probably not a Windows Home Server).

[A||uSiOn]

Christopher,

Thank you for that reply cleared a lot of things up

Well its for a content server for a few hundred users at one given time and they want to run 720p content streamed.

Seems 8 x 512GB RAID 5 on a LSi 9265-8i card + BBU with 2 hotspares will work a dream , the only reason we are going with RAID 5 is that LSi card performance in RAID 5 is almost on par with RAID 10 based on a few reviews specifically this : http://thessdreview.com/our-reviews/...ron-c300-ssds/

If you think I can better that let me know, open to suggestions.

[Computurd]

thats a great review

you will need fastpath to replicate those results, and if you have some money for enterprise you can get some amazing results with SLC drives, a la SanDisk Lightning series. With four drives around 289,000 4k random IOPS possible in RAID 5, iirc.

[Christopher]

If it turns out that you need the speed more than capacity, you can always over provision 512GB drives by a metric f**k tonne and double the endurance. I'm certainly not qualified to speak on these matters, but most of the accesses in a media server should be reads -- possibly 90% reads and 10% writes? Endurance may not be a huge issue, but it is something to think about. Intel's new 700 series enterprise drives using HET-MLC NAND (and a substantial amount of Over Provisioning too) is sort of the middle ground [between a M4 and higher end SLC solutions], but the M4 512 is the same price as the 100GB Intel 710 roughly. I say that if you can determine how much storage you need, what kind/how many writes per day, and find that a high capacity MLC drive with some extra provision could be cost effective, then that's the way to go. If you only needed 300GB per drive of 512GB, that extra space could double (or possibly more) endurance by itself and help keep the drive performance from degrading without TRIM. If your server ends up being super write heavy, an enterprise solution might be what you need, especially if the writes are random and/or small.

[A||uSiOn]

Yeah writes will be very minimal and mostly reads after the initial write of the data. This data will once completely when new content comes available every 90 days. So it will be completely rewritten 4-5 times a year, hardly nothing.

It is going to be accessed by schools streaming content around the country.

The over provisioning is something I will look into but how would one go about calculating the benefits of say 412GB out of the 512GB?

[Computurd]

tbh i do not think you would need the over-provisioning for the model you are using.

[Christopher]

Over provisioning might be more necessary just for keeping performance consistent over time. 20 to 28 percent is somewhat normal for an enterprise drive, so that will give you more endurance and better steady state performance to boot. The 512GB would really be a 480GB as the Crucial uses about 7% spare area. So really anything below that as additional area would just be gravy. I can't really say what kind of performance degradation happens to 8 drives together over time, but I imagine over the course of a year it could get pretty messy, and just going to 14% percent OP could stop quite a bit of it from happening. That's not my department though, perhaps CT could elaborate.

The whole OP issue is pretty interesting, and that's why I wanted to try a second endurance testing drive set up OPed. I believe much of the additional TBW rating of the new 710 drives is due to the heavy OP, so it stands to reason that if I were to get a 470 and added some extra OP that I might be able to add 50 percent to the lifetime.

[johnw]

Over provisioning might be more necessary just for keeping performance consistent over time. 20 to 28 percent is somewhat normal for an enterprise drive, so that will give you more endurance and better steady state performance to boot. The 512GB would really be a 480GB as the Crucial uses about 7% spare area. So really anything below that as additional area would just be gravy. I can't really say what kind of performance degradation happens to 8 drives together over time, but I imagine over the course of a year it could get pretty messy, and just going to 14% percent OP could stop quite a bit of it from happening. That's not my department though, perhaps CT could elaborate.

Wow, it is hard to count the number of mistakes and bad advice in this one paragraph!

[Christopher]

Wow, it is hard to count the number of mistakes and bad advice in this one paragraph!

Yeah, well at 3:00 AM you get me, and not necessarily a sober me at that.

[A||uSiOn]

johnw,

Can you please explain why? Not here to start a heated discussion but would be great to hear your side

[Ao1]

^ http://intelstudios.edgesuite.net/id...2/SSDS002.html (14.45s)

[Anvil]

Kingston SSDNow 40GB (X25-V)

407.94TB Host writes
Reallocated sectors : 12

MD5 OK

34.33MiB/s on avg (~34 hours)

--

Corsair Force 3 120GB

01 92/50 (Raw read error rate)
05 2 (Retired Block count)
B1 69 (Wear range delta)
E6 100 (Life curve status)
E7 47 (SSD Life left)
E9 205009 (Raw writes)
F1 273020 (Host writes)

106.19MiB/s on avg (~22 hours)

power on hours : 805

Wear Range delta is still increasing!
I'll consider giving it a break for some hours, last time that did no good though.
(won't happen until it's 2 days or more or when the new firmware is released)

[Anvil]

^ http://intelstudios.edgesuite.net/id...2/SSDS002.html (14.45s)

Yeah, in essence that's what the Intel drives do. (increase random write endurance and speed)
Over-provisioning can still make a difference as it's hard to tell if (some) of the writes are converted to random writes even if they are sequential by nature.

What impact over-provisioning makes on other controllers is more of a mystery.

[bluestang]

Funny you guys are talking about SSDs in a server. I need to build a new server in a few-to-6 months (waiting on new LGA-2011 XEONs) and also want to use SSDs. It will be a SQL 2008 R2 DB server with the minimum requirements of 6 total SSDs in 3 separate RAID0 plus a couple of enterprise SAS HDDs for backups. DB size could be anywhere around 500MB to 2GB with up to 10 concurent users. I was thinking the M4's or the Intel 520 (depending on the controller used).

[johnw]

johnw,

Can you please explain why? Not here to start a heated discussion but would be great to hear your side

It is off topic for this thread. Why not start a thread with your question? Also, it would be a good idea to put in the first post of the thread as much information as possible about your expected usage of the server (budget, capacity, peak and average load, access patterns, needed reliability and uptime, server hardware, network connection, OS, software, etc.)

[bluestang]

M225->Vertex Turbo 64GB Update:

541.24 TiB (595.10 TB) total
1475.04 hours
10966 Raw Wear
118.74 MB/s avg for the last 15.98 hours (on W7 x64)
MD5 OK
C4-Erase Failure Block Count (Realloc Sectors) at 10.
(1=Bank 6/Block 2406; 2=Bank 3/Block 3925; 3=Bank 0/Block 1766; 4=Bank 0/Block 829; 5=Bank 4/Block 3191; 6=Bank 7/Block 937; 7=Bank 7/Block 1980; 8=Bank 7/Block 442; 9=Bank 7/Block 700; 10=Bank 2/Block 1066)

[bulanula]

I will stick to the C300 as the most endurant for now. Maybe something will beat it but for now ( seeing as the SF drives can't even write a bit before dying out ) I bet it will be the most endurant.