Sign in with
Sign up | Sign in

Results: Write Endurance

Micron P400m SSD Review: High Endurance MLC Is Here To Stay
By

We typically spend a lot of time evaluating write endurance when we review enterprise-class SSDs. Write endurance is one of the major differentiators separating enterprise and client-oriented drives, after all. As MLC-based storage continues pushing its way into spaces previously filled by SLC NAND, we have to keep a close eye on this difficult-to-benchmark, but still very important variable involved in evaluating solid-state storage.

Micron takes a two-fold approach to ensuring ample write endurance from its P400m. First, it uses 25 nm MLC NAND specifically binned for enterprise purposes. Instead of the 3,000 or 5,000 program/erase cycles you typically get from client-oriented MLC at 25 nm, Micron rates the P400m's memory for 20,000 cycles. Second, Micron uses 70% over-provisioning across all of the P400m's available capacities. Taken together, the 300 GB model should endure 3.5 PB of data written to it over the course of its life. Put differently, Micron says you can write the drive's capacity 10 times per day for five years. 

As always, we perform our tests using sequential workloads with large block sizes. This reduces the impact of write amplification and wear leveling, giving a better indication of actual P/E cycles for the NAND.

Endurance Rating
Sequential Workload, QD=1, 8 MB
Micron P400m
Intel SSD DC S3700
NAND Type
Micron 25 nm MLC
Intel 25 nm HET-MLC
RAW NAND Capacity
340 GB680 GB
264 GB
IDEMA Capacity (User Accessible)
200 GB400 GB
200 GB
Over-provisioning
70%70%32%
P/E Cycles Observed (IDEMA)
34,195
34,195*36,343
P/E Cycles Observed (Raw)
20,113
20,113*27,532
Host Writes per 1% of MWI
68.39 TB
136.78 TB*
72.69 TB
$/PB-Written
$87.73
$68.71*
$64.66
*Estimated


When we compare the specifications of Micron's P400m to Intel's SSD DC S3700, they appear nearly identical. Both products are said to sustain 10 full writes per day for five years. The 200 GB models from both families specify between 3.5 and 3.65 petabytes of total data written to them. At comparable user-accessible capacities (200 GB), we see similar observed P/E cycles (34,000 versus 36,000).

It appears that Intel holds the leads when it comes to write endurance, but Micron compensates by arming its drive with 340 GB of raw NAND (rather than Intel's 264 GB). This is a good academic exercise. However, in the real-world, customers only care about actual performance and actual endurance. In that context, Micron has no trouble keeping pace with its higher level of over-provisioning.

Of course, using over-provisioning to improve write endurance has its drawbacks. In this case, that drawback is cost. Each gigabyte of user-accessible space on the SSD DC S3700 costs $2.35/GB. The P400m will push that closer to $3/GB. When you look at the raw NAND costs, however, the gap disappears; Micron's P400m is $1.76/GB, while Intel's SSD DC S3700 is $1.78/GB. Consequently, when you look at cost per petabyte of data written, the Intel drive maintains a lead.

Although we only tested Micron's 200 GB P400m, we also estimated the 400 GB model's endurance because it's priced lower than the 400 GB competitor from Intel. When you compare 400 GB drives, cost per petabyte written even out, making the 400 GB P400m a worthy alternative to the SSD DC S3700.

Ask a Category Expert

Create a new thread in the Reviews comments forum about this subject

Example: Notebook, Android, SSD hard drive

Display all 16 comments.
This thread is closed for comments
  • 6 Hide
    Nintendo Maniac 64 , February 13, 2013 4:02 AM
    に?

    Seriously though, I wonder why Tom's doesn't run one of their basic "real-world" tests used on consumer SSDs (such as Tom's 7zip test) on one of these professional SSDs just so that we can get an idea how they compare to the consumer-level stuff.

    In particular, Tom's always says that comparing one SSD to another is nearly moot point when you consider the magnitude of improvement an SSD has over a traditional HDD; it would be nice to know if these pro-level SSDs are of a similar magnitude of improvement over consumer SSDs or whether the difference is actually less.
  • -1 Hide
    blazorthon , February 13, 2013 5:41 AM
    Nintendo Maniac 64に?Seriously though, I wonder why Tom's doesn't run one of their basic "real-world" tests used on consumer SSDs (such as Tom's 7zip test) on one of these professional SSDs just so that we can get an idea how they compare to the consumer-level stuff.In particular, Tom's always says that comparing one SSD to another is nearly moot point when you consider the magnitude of improvement an SSD has over a traditional HDD; it would be nice to know if these pro-level SSDs are of a similar magnitude of improvement over consumer SSDs or whether the difference is actually less.


    The difference is much, much less in terms of performance difference. Tom's has told us this time and time again.
  • 4 Hide
    mayankleoboy1 , February 13, 2013 6:49 AM
    Nintendo Maniac 64に?Seriously though, I wonder why Tom's doesn't run one of their basic "real-world" tests used on consumer SSDs (such as Tom's 7zip test) on one of these professional SSDs just so that we can get an idea how they compare to the consumer-level stuff.In particular, Tom's always says that comparing one SSD to another is nearly moot point when you consider the magnitude of improvement an SSD has over a traditional HDD; it would be nice to know if these pro-level SSDs are of a similar magnitude of improvement over consumer SSDs or whether the difference is actually less.


    In desktop loads, very very less difference.
    In server loads, huge difference. Plus, these server SSD's wil maintain high speeds even after large amounts of data is continuously being written.
  • 2 Hide
    mayankleoboy1 , February 13, 2013 6:55 AM
    theoretical question : how much life would 100% provisioning give ? So you ship a 512GB MLC drive, but the usable is only 256GB. The rest 256GB is for getting better wrtite endurance.
    How would this compare to a true SLC SSD ?
  • -1 Hide
    blazorthon , February 13, 2013 7:19 AM
    mayankleoboy1theoretical question : how much life would 100% provisioning give ? So you ship a 512GB MLC drive, but the usable is only 256GB. The rest 256GB is for getting better wrtite endurance. How would this compare to a true SLC SSD ?


    It'd probably still be inferior overall and not even be cheaper at that point. Over-provisioning is only good for mitigating MLC's disadvantages over SLC AFAIK, not replacing SLC.
  • 1 Hide
    blazorthon , February 13, 2013 7:23 AM
    mayankleoboy1In desktop loads, very very less difference.In server loads, huge difference. Plus, these server SSD's wil maintain high speeds even after large amounts of data is continuously being written.


    Even in server workloads, there are many desktop drives where the performance difference is still not great. For example, Vector is right up there at the tops of the charts with Samsung 840 Pro in performance and many cheaper alternatives are often not far behind in performance. Endurance is another matter, but that wasn't the question. If you want a seriously significant performance difference like with HDDs versus cheap consumer SSDs, you have to consider SSD RAID and/or extreme PCIe storage.

    Also, many consumer drives have no trouble keeping performance over time with lots of data written. That's also not an enterprise-only feature.
  • 0 Hide
    mayankleoboy1 , February 13, 2013 12:10 PM
    edit
  • 0 Hide
    drewriley , February 13, 2013 12:41 PM
    mayankleoboy1theoretical question : how much life would 100% provisioning give ? So you ship a 512GB MLC drive, but the usable is only 256GB. The rest 256GB is for getting better wrtite endurance. How would this compare to a true SLC SSD ?


    Some back of the napkin calculations - If you have a consumer 512GB MLC SSD with no over-provisioning, and the MLC was high grade consumer, you could expect 5K P/E cycles. In order to mimic the P400m, which is 35K P/E cycles, the usable space would be ~75GB. By dong that, you would be paying roughly $7/GB for usable storage(assuming you paid $512 for your consumer drive). For SLC, you are looking at 17GB of usable space and $30/GB.

    No matter how you look at it, consumer grade MLC will never get close to eMLC and SLC in terms of write endurance, unless the price goes down by orders of magnitude compared to eMLC and SLC.
  • 2 Hide
    drewriley , February 13, 2013 12:54 PM
    blazorthonEven in server workloads, there are many desktop drives where the performance difference is still not great. For example, Vector is right up there at the tops of the charts with Samsung 840 Pro in performance and many cheaper alternatives are often not far behind in performance. Endurance is another matter, but that wasn't the question. If you want a seriously significant performance difference like with HDDs versus cheap consumer SSDs, you have to consider SSD RAID and/or extreme PCIe storage.Also, many consumer drives have no trouble keeping performance over time with lots of data written. That's also not an enterprise-only feature.


    You are correct. Many drives, such as the 840 PRO, would perform great on the enterprise performance tests. There are also a lot of consumer drives that would have problems over time if they aren't allowed to TRIM every so often. We limit the scope of our testing because the main use cases for these drives are enterprise. Will some companies use them in workstations, absolutely. The same can be said for high-end RAID cards too. Nearly every consumer drive would perform very poorly when you take into account write endurance and other enterprise features.
  • 0 Hide
    themdg , February 13, 2013 3:06 PM
    This technology is moving so fast...hard to keep up. I feel like these articles could use that line from Tommy Boy and get the same point across (to me, at least).

    "These hard drives are really cool...you're not even gonna believe it..."
  • -2 Hide
    sanilmahambre , February 13, 2013 5:19 PM
    Oh boy! this awesome ssd is the best in the world.
  • 0 Hide
    dalethepcman , February 13, 2013 7:44 PM
    The funny thing about "consumer" ssd performance for enterprise is that HDD bandwidth was never a problem that faster drives could solve, seek time and iops were. The reason bandwidth was never a problem is that even using traditional spinning drives you could easily max throughput of a 16gb/s fiber channel (that's only 100 drives.)

    Although these devices seem to be targeted for enterprise use, ill be sticking with 3par for the foreseeable future.
  • 0 Hide
    f-14 , February 14, 2013 5:31 PM
    Nintendo Maniac 64に?Seriously though, I wonder why Tom's doesn't run one of their basic "real-world" tests used on consumer SSDs (such as Tom's 7zip test) on one of these professional SSDs just so that we can get an idea how they compare to the consumer-level stuff.In particular, Tom's always says that comparing one SSD to another is nearly moot point when you consider the magnitude of improvement an SSD has over a traditional HDD; it would be nice to know if these pro-level SSDs are of a similar magnitude of improvement over consumer SSDs or whether the difference is actually less.


    they will get back to you on that in 5 years as this article is about durability of this drive mostly with new firm ware and higher binned modules.

    it's the difference between recreational paintball quality and tournament winter paintball grade. not about speed, just reliability in the long term. the article even states the 300MB up/down per second transfer speeds in the specs if you care to read.
  • 0 Hide
    f-14 , February 14, 2013 5:39 PM
    forgot to mention, in paintballs they're all great coming right out of the factory, but with rec ball the shells aren't as thick and degrade after 3 months and tend to break in your barrel or breach both rec and winter tournament grade will develop flat spots it's just a matter of how bad and after how long. the rec is made with a thin shell and the tournament is made with thicker shell, winter grade is double shelled. to get a thicker double shell is optimal but that decreases the chances of them breaking open and in
  • 0 Hide
    mark0718 , February 18, 2013 2:44 AM
    Re mayankleoboy1 and others:
    Some consumers MLCs ship seem to run a SLC if only half of the space is allocated and
    seem to automatically switch to MLC as more space is used and not freed by TRIM.
    Thus, the consumer can buy a large drive but only partition a few MB less than the stated
    capacity and get SLC performance until/unless he needs more capacity.

    Note: while the read and write performance seems to be as good as real SLC, it isn't
    known to the public what the endurance is.

    The reason why a consumer might want to configure the way is the "enterprise" SLC stuff
    (and even some "enterprise" MLC stuff) costs 10 times as much as consumer stuff,
    so using only half available space is still only 1/5 the cost of "enterprise".

    It also isn't publically know if the endurance of the cells is the same if operated as
    SLC. (My quess is that the endurance would be higher.)
  • 0 Hide
    blazorthon , February 18, 2013 1:45 PM
    mark0718Re mayankleoboy1 and others: Some consumers MLCs ship seem to run a SLC if only half of the space is allocated andseem to automatically switch to MLC as more space is used and not freed by TRIM.Thus, the consumer can buy a large drive but only partition a few MB less than the statedcapacity and get SLC performance until/unless he needs more capacity.Note: while the read and write performance seems to be as good as real SLC, it isn'tknown to the public what the endurance is.The reason why a consumer might want to configure the way is the "enterprise" SLC stuff(and even some "enterprise" MLC stuff) costs 10 times as much as consumer stuff, so using only half available space is still only 1/5 the cost of "enterprise". It also isn't publically know if the endurance of the cells is the same if operated asSLC. (My quess is that the endurance would be higher.)


    Te consumer MLC drives, even when used in that way, are still far from equaling enterprise SLC, let alone beating it. Also, consumer SSDs that do anything like that such as Vertex 4 and Agility 4 don't adjust based on the capacity of the drives IIRC. As I recall, they constantly rewrite the data as MLC regardless of the capacity used to avoid the issue of huge speed bumps as the drives fill past certain capacity points known to some older firmware versions.