Review Index:

Intel SSD 750 Series 1.2TB PCIe and 2.5" SFF Review - NVMe for the Consumer

Subject: Storage
Manufacturer: Intel

Sequential Performance - HDTach, HDTune, File Copy, YAPT (sequential)

We have shifted over to combining our results into two groupings for consumer reviews. First up is sequential performance:


HD Tach will test the sequential read, random access and interface burst speeds of your attached storage device (hard drive, flash drive, removable drive, etc). All drive technologies such as SCSI, IDE/ATA, 1394, USB, SATA and RAID are supported. HDTach tests sequential performance by issuing reads in a manner that was optimized more for HDD access, but this unique method has proven useful in evaluating the sequential response time of SSDs. The accesses are relatively small in size (2k), and are issued with a single working thread (QD=1). The end result is that devices with relatively IO high latency will not reach their ultimate rated speed.

View Full Size

Looks like Intel just punched everyone else in the gut to start off this fight. HDTach is not an ideal sequential read test for an SSD, as it performs very small read commands that are issued sequentially (rather than queued). Despite this, the SSD 750 turns in the highest figures we've ever seen in this test.


HDTune tests a similar level of features as compared with HDTach, but with a different access pattern. Thus provides us with an additional set of benchmark numbers to compare between storage configurations. CPU utilization has proven negligible with modern processing horsepower, and is no longer included. Additionally, we do not include write performance due to HDTune's write access pattern not playing nicely with most SSDs we have tested it on.

View Full Size

We're not sure why HDTune has become as inconsistent as it has with these faster SSDs, but it appears to not mesh well with NVMe devices. Speeds are still good, but we know the SSD 750 is much more capable than what this test reports. Included here mostly as a data point, but we will be discontinuing use of this test in the near future.

PCPer File Copy Test

Our custom PCPer-FC test does some fairly simple file creation and copy routines in order to test the storage system for speed.  The script creates a set of files of varying sizes, times the creation process, then copies the same files to another partition on the same hard drive and times the copy process.  There are four file sizes that we used to try and find any strong or weak points in the hardware: 10 files @ 1000 MB each, 100 files @ 100 MB each, 500 files @ 10 MB each and 1000 files at 1 MB each.

View Full Size

Intel cleans house on file creations right until the 1000 x 1MB file creation, where it slows a bit. This is because its enterprise pedigree heavily optimizes for 4k and higher random access, and writing very small files with our tool means a lot of <4k file table updates, slowing things down a bit.

View Full Size

Overall the fastest we've ever seen the copy test complete. Watching the batch file run was like listing a large directory at the command line. It simply flew by. Since writes during a copy operation can be cached by Windows, the small file creation slow down we saw above no longer had an impact on the SSD 750.


YAPT (yet another performance test) is a benchmark recommended by a pair of drive manufacturers and was incredibly difficult to locate as it hasn't been updated or used in quite some time.  That doesn't make it irrelevant by any means though, as the benchmark is quite useful.  It creates a test file of about 100 MB in size and runs both random and sequential read and write tests with it while changing the data I/O size in the process.  The misaligned nature of this test exposes the read-modify-write performance of SSDs and Advanced Format HDDs.

View Full Size

YAPT has always done a great job of maxing out SSDs, and the same applies here. The SSD 750 takes the crown with nearly 2.6 GB/sec reads...

View Full Size

...but falls short of the Phoenix Blade on sequential writes. This is because Intel was more conservative in its write performance in the interest of lower per-IO latency. This means that overall IOPS of the SSD 750 is higher in mixed workloads as it is not saturating itself with writes.

Video News

April 2, 2015 | 12:08 PM - Posted by Allyn Malventano

A quick note on the samples we included in this testing:

  • The OCZ RevoDrive 350 was left out to make room for other devices. In its place was the G.Skill Phoenix Blade, which is equivalent (and faster) hardware.
  • The SSD P3700 was not included in these results to make room for other competitors. While a bit faster than the SSD 750, it is not bootable in most consumer systems and therefore not an appropriate comparison for this review.
April 3, 2015 | 11:09 PM - Posted by Anonymous (not verified)

I just bought the latest ASUS Z10PE10-D8 WS MOBO. Can you tell me for the 2.5 form factor will I need that adapter SFF? Plus does ASUS need to update the bios for this board to be able to boot from a NVMe SSD?

April 2, 2015 | 12:23 PM - Posted by Jason (not verified)

why didn't intel use NVMe over sata express? the use of the sff seems like an odd choice, compared to the more compatible and consumer friendly sata express.

April 2, 2015 | 12:50 PM - Posted by Allyn Malventano

SATA Express is limited to PCIe x2 (typically 2.0). This SSD easily saturates that link.

April 2, 2015 | 05:30 PM - Posted by Anonymous (not verified)

SATA express just seems to be DOA. It is a large connector and just seems like a kludge. I don't like using this weird m.2 adaptor. They reall should just create an official standard for PCI express cable connections if one doesn't already exist.

April 4, 2015 | 12:00 PM - Posted by Bill (not verified)

what is up with this horrid boot time....worst out of 14 other drives. Plextor M6 is 15 seconds, this 1200 dollar 750 is 34 seconds...whats up, see review here.

April 4, 2015 | 02:50 PM - Posted by Allyn Malventano

That's a BIOS issue. 

April 21, 2015 | 04:32 AM - Posted by mvitkun
I looked at Asus' ROG blog and it seems to have some unclear information.
I was hoping you could clear it up for me

all ASUS X99 and Z97 motherboards offer full support for both PCIe and M.2 NVMe storage devices.

It then goes on to say that only X99 supports hyperkit.

With the optional ASUS Hyper Kit expansion card, owners of ASUS X99 motherboards are free to attach 2.5-inch NVMe storage devices via the SFF-8639 (Mini-SAS HD) connector provided by Hyper Kit.

All the Z97 boards are listed as N/A for compatibility...does that mean that the bios will be updated to allow compatibility with the hyperkit, but that they aren't currently compatible?

Or does it mean that all their boards are compatible with NVMe m.2 SSDs, but hyperkit will only be supported by X99?

April 2, 2015 | 12:30 PM - Posted by Anonymous (not verified)

My question is regarding that non-pcie ssd, why didnt Intel choose a 3.5" form factor for that drive it would have allowed for better thermals and most of the enterprise / enthusiast cases have 3.5" mounting options by default. other than that in future I hope all ssds drop that ahci protocol for the faster nvme or some other protocol.

April 2, 2015 | 12:52 PM - Posted by Allyn Malventano

Most cases also have 2.5" mount points. Intel already had their enterprise variant in the standard 2.5" SFF-8639 form factor (used to fit more units into a smaller rack mount chassis). Also, the thermals are handled well enough with the smaller form factor, so no need for them to engineer a larger one.

April 2, 2015 | 12:32 PM - Posted by Anonymous1 (not verified)

How does it work with 86 dies? Isn't it supposed to be multiple of 18 since it's an 18 channel controller?

April 2, 2015 | 12:56 PM - Posted by Allyn Malventano

Intel's controller does not require fixed die count per channel. They can stagger them however they please.

April 2, 2015 | 01:00 PM - Posted by cyberwire

Any comparison of CPU usage vs sata/other implementations?

April 2, 2015 | 01:13 PM - Posted by Allyn Malventano

This is very tricky to measure accurately as the IOPS figures are vastly different between the two types for any given test. Data I've seen in NVMe presentations suggests the CPU overhead is reduced to ~35% of an equivalent AHCI device for each IO.

April 2, 2015 | 01:00 PM - Posted by Anonymous (not verified)

Well its insanely fast but does it really matter?
Do you "feel" any difference between a maxed out SATA and a 4xPCIe NVMe SSD?

April 2, 2015 | 01:11 PM - Posted by Allyn Malventano

It really depends on what you are doing with it, but the fact that latency is at least half of any given SATA IO, that largely adds to the 'feel' of the device. As a point of comparison, this is much closer to the feel of a RAM drive.

April 29, 2015 | 01:01 AM - Posted by Anonymous (not verified)

This is easily explained this way:
The slowest part in your computer is your HDD/SSD currently. If you increase that speed you increase the entire computer speed.
You well not see benefit if SSD getting faster if that SSD has the speed of your RAM.

April 2, 2015 | 01:05 PM - Posted by Polycrastinator (not verified)

That's awesome. Does a consumer/enthusiast really need this sort of speed, though? A colleague's immediate reaction was "I want one of these for my ESXi server," and I have to agree on that score. Still feels like something best suited for specialist applications. Especially with that connector on the 2.5" model.

April 2, 2015 | 01:17 PM - Posted by Allyn Malventano

I would agree if the cost/GB was higher. For anyone wanting PCIe-level SSD performance, this comes in at roughly half the cost of any competing solution that comes anywhere close on sequentials, and it wipes the floor with them in random IO and extremely low latency. This would be a good equivalent for those who have previously been comfortable with an SSD 4-drive RAID-0 (but this is still faster).

April 2, 2015 | 04:40 PM - Posted by BBMan (not verified)

The general consumer is likely going to wait until these are within whistling distance of an 850 EVO (.35-.40 $/GB). The general consumer- even many prosumers- are not going to care unless they have enterprise level operations.

Still, it's pretty impressive if you are spending for ego purposes.

April 2, 2015 | 01:09 PM - Posted by Polycrastinator (not verified)

How hot do these get? Especially the 2.5" one? Does it need a fan blowing past it, or could you stick it in a corner of your case without any direct airflow without issue?

April 2, 2015 | 01:15 PM - Posted by Allyn Malventano

The 1.2GB model can peak at 22W during continuous heavy writes, so you would certainly not want it sitting in a 'dead spot' of your case if you intend to load it to that degree. For easier / intermittent use, it would probably be ok with indirect ventilation.

April 2, 2015 | 01:25 PM - Posted by OuchClonk (not verified)

'O geezz I wonder how many of these dang new fangal social services will Ryan make me join in order to Enter this competition[Like Twitter... Yes Ryan Some Don't have FB's and TWIT'sey accounts. or forgot the password ;(] will Ryan make me join in order to Enter this competition..Anyway looking forward to this.. I'll start diggin up that twatter pword:.

April 2, 2015 | 01:35 PM - Posted by Anonymous (not verified)

1. When will these be available for purchase (not pre order)?

2. Why did they use a green pcb instead of a better looking black one?

May 3, 2015 | 04:03 PM - Posted by Anonymous (not verified)

yeah i miss the skulltrail logo as well a black pcb. didnt the 730 series have one at least?

April 2, 2015 | 01:38 PM - Posted by qqqvid

I'd buy this for a dollar : )

April 2, 2015 | 01:47 PM - Posted by Buyers

*hnnnggh* Oh man, like i needed another reason to upgrade from my aging z68 and sandy bridge setup.

April 2, 2015 | 01:59 PM - Posted by Shambles (not verified)

Looking at some other reviews I see that although this is a fast drive it doesn't seem to improve game load times whatsoever. What's the bottleneck now that drives aren't the slowest link. The CPU? If the CPU can push benchmarks that much faster shouldn't it be loading games that much faster as well?

May 3, 2015 | 04:09 PM - Posted by Anonymous (not verified)

nowadays games are made with consoles in mind so you have fixed hardware to reach your frametimes. beside some fancy pc centric MMO's i dont know any software to need better streaming performance storage-wise.

April 2, 2015 | 02:33 PM - Posted by HypnoSteak

Those drives are sweet

April 2, 2015 | 03:24 PM - Posted by razor512

Does the new controller also have the the hard write limit e.g., bricking/ read only mode after a specific amount of writes?

April 2, 2015 | 04:43 PM - Posted by Allyn Malventano

None that we are aware of. These should not go into read-only mode unless they exhaust all spare cells. The P3700 series would brick in that condition, but that is because Intel enterprise firmwares are designed to fail on any significant fault (so they can be easily identified and replaced). The 750 has a more consumer friendly firmware.

April 2, 2015 | 03:25 PM - Posted by ZoyosJD

Nice review, really like what I'm hearing in regards to performance. BTW, just checked Amazon to see the pre-order prices sitting over $1/GB at $488 and $1230 respectively without shipping included.

Should we be expecting updates (firmware or new iterations) that will allow systems take advantage of of the drives full potential (ex. multi-threading, raid)?

April 2, 2015 | 04:39 PM - Posted by Allyn Malventano

These already multithread extremely well (it was the fault of our test methods pegging before the 750). RAID boot support is extremely unlikely on the current generation of systems.

April 2, 2015 | 03:46 PM - Posted by Anonymous (not verified)

SSDs don't need to get faster, they need to get cheaper. $0.86/GB is pathetic.

April 2, 2015 | 04:37 PM - Posted by Allyn Malventano

Faster $0.86/GB SSDs drive down the cost of lower performing SSDs even further.

April 2, 2015 | 07:48 PM - Posted by jimecherry

lets hope that the interface change doesn't complicate the usual pricing trends.

April 29, 2015 | 12:56 AM - Posted by Anonymous (not verified)

They do need to get faster. HDD/SSD is the slowest part in any PC. Increase that speed and you increase the entire PC speed.

April 2, 2015 | 07:15 PM - Posted by Anonymous (not verified)

Where's that ARM douche in these comments?

April 2, 2015 | 09:53 PM - Posted by Dean (not verified)

Sounds like single threaded software might be the upcoming bottleneck in PC's if the SSD market keeps growing?

April 3, 2015 | 04:21 AM - Posted by Cyclops

So Allyn, NVME drives have a single controller correct? That should make them more reliable than stuff like the Phoenix blade.

Isn't the Phoenix blade essentially four SSDs in RAID0? Wouldn't that be a much more volatile solution than the 750?

Not to mention that the blade is using ISCSI that's limited to 32 queues which is where random performance and actual real world usage comes in.

April 3, 2015 | 02:33 PM - Posted by Allyn Malventano

Yes, those competing units have more points of failure. Since the Blade is handing off IOs to multiple controllers, it can actrually scale to QD>32, but it still has its limits, and is significantly more latent per IO.

April 4, 2015 | 02:49 AM - Posted by Cyclops

Good info, thanks.

April 3, 2015 | 10:20 AM - Posted by Erik Mulder (not verified)

I've been looking for the 400GB AIC version but so far i've found them for about $100 above the listed price on this website...
newegg and amazon is where i found it so far.

April 3, 2015 | 12:26 PM - Posted by TheFutureIsHere (not verified)

Have you tried booting windows from one of these? Does windows boot and load apps much faster than with a SATA SSD?

April 29, 2015 | 12:58 AM - Posted by Anonymous (not verified)

You do not need to try to verify this will be true.

April 3, 2015 | 01:44 PM - Posted by nagus (not verified)

If intel is only guaranteeing compatibility with Z97 and X99 platforms... how are we going to test for compatibility with everything else?

I'm on X79 and would love this drive.

I have a UEFI bios but have no idea if this will work...


April 4, 2015 | 12:03 PM - Posted by Bill (not verified)

This report says it needs UEFI version 2.3.1

April 10, 2015 | 12:05 PM - Posted by Allyn Malventano

That's only part of the requirement. That UEFI revision alone does not guarantee that a 750 will be bootable. 

April 3, 2015 | 02:58 PM - Posted by another brian (not verified)

A couple of times when performance suffered on the 750 you made a reference to "4k". Can you explain 4k and how it affects this drive? Does it have anything to do with how the drive is formatted or the BIOS?

April 3, 2015 | 08:22 PM - Posted by Anonymous (not verified)

I am assuming that the 4k number is because most modern file systems use 4k (4096 byte) cluster or block size. So under normal circumstances, all accesses will be aligned to 4k boundaries.

April 3, 2015 | 04:51 PM - Posted by Anonymous (not verified)

Where to buy?

April 4, 2015 | 11:55 AM - Posted by Bill (not verified)

Newegg has it as a pre order

April 4, 2015 | 03:44 AM - Posted by Anonymous (not verified)

Hi, I heard from some reviews that the 750 ssd has the worst boot up timing? Can anyone verify this as there is not much mentioned on the boot up timing...

I heard is incredibly slower compared to even SATA drive, (techreport review shows 51 secs boot up time for 8.1, compared to 30s for other SATA drives) hence is there any verification to this

April 4, 2015 | 11:53 AM - Posted by Bill (not verified)

Yes, what is up with this horrid boot time....worst out of 14 other drives. Plextor M6 is 15 seconds, this 1200 dollar 750 is 34 seconds...whats up, see review here.

April 4, 2015 | 01:37 PM - Posted by Anonymous (not verified)

For that price, that SSD better get boot up in 10 secs or less.

No one like to spent time waiting for PC to restart slowly while trying to fine tune their BIOS for bio flash, OCing etc. Not at this price

April 4, 2015 | 02:55 PM - Posted by Allyn Malventano

BIOS boot support is in early stages for these, and the longer times are due to BIOS implementation - not the direct fault of the 750.

April 4, 2015 | 07:05 PM - Posted by Bill (not verified)

Those guys in your intel video said it needs the UEFI 2.3.1, is that something we can download if board doesnt have it. Im getting the ASUS X99 Deluxe/3.1. Got call into asus , but have not heard back, they were researching it.

April 10, 2015 | 12:07 PM - Posted by Allyn Malventano

In defense of the 750, that linked test is not an apples to apples comparison. The boot time of the 750 was not on the same motherboard as the other times tests in their chart.

April 4, 2015 | 06:10 PM - Posted by Anonymous (not verified)

Is there a reason Intel has made the firmware for the 3rd gen(?) controller so unoptimized for 0.5-2.0kb transfer sizes? It's been like this on every single drive that has used the controller(s3700,s3500,730,p3600,p3700) from them.

What do they gain by keeping the current controllers firmware optimized like this? It's not like these transfer sizes aren't seen in enterprise and consumer environments.

Also, am I safe to assume that like the s3700, the intel 750 will actually gain performance by disabling trim? That's another interesting quirk that the controller has.

April 7, 2015 | 11:16 PM - Posted by jgstew

All SSDs are better at 4k or larger writes than anything under 4k. This is due to the way flash memory is programmed. Additionally, file system cluster sizes tend to be 4k or larger as well. There is little need to optimize for any transfer less than 4k.

April 4, 2015 | 08:57 PM - Posted by Pagan Mutant (not verified)

The 2.5" version requires the ASUS Hyper Kit.
So ... Where do we get one from?

I found it stated somewhere that the SSD comes with that "cable" ... which is worthless without something to plug it into.

JJ from ASUS implied that one comes with one of the ASUS motherboards, yet I see none of them on the ASUS site showing that Hyper Kit in their list of accessories nor anywhere else on the site, for that matter.

Search for the Hyper Kit on the ASUS site.
It's like they've never heard of it ... other than merely quoting someone's review of the Intel 750 Series SSD ... and how it requires the ASUS Hyper Kit (and thus we're back where we started from).

I've Googled it and all I get are the similar SSD Review references ... nothing stating where one can get one from, nor does it show up on Amazon, etc..

When anyone out there successfully completes a purchase and receipt of a 2.5" version, could you please post a note to let us all know what was (and was not) included in the package.
(Did it include that Hyper Kit?)


April 4, 2015 | 10:37 PM - Posted by veritronx (not verified)

It has been mentioned a few times that Intel is working with motherboard manufacturers to include the connector, the only one that comes with one right now is the new x99 sabertooth which is what pcper used to test that drive.

May 3, 2015 | 03:49 PM - Posted by Anonymous (not verified)

Just some fancy routing or are there any performance slowdowns to expect when using the Hyper drive kit? Did they use it in this test or was the plugin card used? I think the autor was not very specific wich one was used.

April 10, 2015 | 11:54 AM - Posted by EnjoyFebruary (not verified)

So my question is - Why isn't there just a male M.2 PCIe connector on the 2.5" version? Maybe I'm not understanding the electrical requirements correctly.

April 10, 2015 | 07:52 PM - Posted by Allyn Malventano

The drive is physically too large to fit an M.2 slot. 

April 29, 2015 | 12:55 AM - Posted by Anonymous (not verified)

Finally a good SSD. I will be getting one for sure. This will change the way we use computers.
I hated the bottleneck all other SSDs have and this solves this problem! What more can you ask for? I am also tempted to see the next version and how much faster it can get.

May 6, 2015 | 12:05 PM - Posted by BBMan (not verified)

Nice review. Indeed a fast storage solution- if you find it necessary.

One of the things I find not compared is the power consumption. From what I can tell, many SSDs are far better than this. It is an edge to consider if your into that.

Personally, I'd like to see more efficient and mobile solutions soon.

May 6, 2015 | 06:31 PM - Posted by kunglao

I want to win this DRIVE

May 6, 2015 | 08:12 PM - Posted by pnadeau1

Nice storage device.. I really like that m.2 connector...

May 6, 2015 | 10:14 PM - Posted by Strykir

Amazing something to throttle my game even more bring it on says the king......woot

May 7, 2015 | 12:50 AM - Posted by Anonymous (not verified)

I want to win this!

May 7, 2015 | 04:36 PM - Posted by glfnchf24

Would like to see someone take one of these and have Dave Dougdale over at edit some video for us to see how it would improve our workflow. Left a message at PCDIY for JJ to do it, especially since he did a build with him. Need to know if it will help in my endeavors for video editing.

May 8, 2015 | 07:14 AM - Posted by skline00

What a ssd! Would love to run it in my Asus Rampage Extreme V 5960x rig!

May 8, 2015 | 08:31 AM - Posted by Neppie86 (not verified)

This is the first time i see an International Contest. between Linus,Paul's hardware, Tek Syndicate and you guys. Keep up the good job guys and thanks for the review. =D

May 8, 2015 | 09:10 AM - Posted by john fodo (not verified)

Sweet!But will this work on my Amd ASUS SABERTOOTH 990FX R2.0 MB?

May 9, 2015 | 09:27 AM - Posted by alannathanson

Looks sweet!

May 9, 2015 | 07:33 PM - Posted by vando1952

Being an Australian where everything, especially broadband is so bloody slow, this would really make my PC sing.

May 9, 2015 | 07:34 PM - Posted by vando1952

Being an Australian, where everything is a mile behind, especially the internet. This SSD would make me very very happy and my PC sing.

May 9, 2015 | 10:41 PM - Posted by Tim Stanton (not verified)

This is the type of thoughtful, in-depth analysis I enjoy when I come to learn about the latest in PC tech. Thanks

May 10, 2015 | 03:08 AM - Posted by Anonymous (not verified)

this 750 model ssd looks and specs are AWESOME!!!

May 12, 2015 | 05:19 AM - Posted by cmd_fuller (not verified)

Very nice, big thanks to intel and pc per for keeping up the great work

May 12, 2015 | 09:52 AM - Posted by Mark_GB

I think I can hear the Samsung 840 EVO 1TB SSD crying inside of my case...

May 12, 2015 | 10:27 PM - Posted by maxart111

thanks for the ifo

May 13, 2015 | 06:17 AM - Posted by adam (not verified)

the specs are awesome

May 13, 2015 | 11:30 AM - Posted by Ricardo Borzani (not verified)

WOW what a incredible speed these ssd have! change my sansung EVO for this is like going from HD to SSD

May 13, 2015 | 09:38 PM - Posted by Andy Petrie (not verified)

I want one for my new rig!!! great review

May 14, 2015 | 05:42 PM - Posted by Barta (not verified)

Good to see manufacturers supporting reviewers with direct and indirect sponsoring.

May 14, 2015 | 06:43 PM - Posted by Jason (not verified)

My samsung SSD drive now working love to try intel SSD 750 Series PCIe / NVMe (400GB)

May 15, 2015 | 02:34 AM - Posted by Anonymous (not verified)

Great review I really need one of these.

May 15, 2015 | 03:35 AM - Posted by vando1952

That would really help my cause....LOL

May 15, 2015 | 06:37 PM - Posted by DrWattsOn

This is fabulous tech, I just wonder how soon it will become mainstream-enough to easily find hardware.
I hear they're working on a drive so fast it finishes all your work before the computer boots!
Seriously, though, this is currently just a millionaire's toy (and since in USA our Corporations are legally individuals, that means their toy).
Great review from the Genius Malventano, and terrific input from the poster's to the comments, all of whom contribute to learning more about this tech.
Thanks. Keep up the good work, all. Shout out to Shrout also of fame.

May 24, 2015 | 10:00 AM - Posted by MRFS (not verified)

> Great review from the Genius Malventano

I second that! Genius indeed!!


May 16, 2015 | 04:32 PM - Posted by olin (not verified)

I did not like the M.2 slot implementation of the 2.5" drive. but as we need a new interface, till then I would go for the PCI express version.

May 22, 2015 | 04:50 AM - Posted by Anonymous (not verified)

hope to win the hardrive

May 22, 2015 | 04:54 AM - Posted by Anonymous (not verified)

win win

May 22, 2015 | 10:34 PM - Posted by MRFS (not verified)

Allyn and Ryan,

You guys are THE BEST!

Many thanks for the great video discussion.

It helped a lot to see you simplify the cabling topology with a clear demonstration.

Also, Jeremy wanted me to ask Allyn this next question:

Over here:

I asked:
Jeremy, from an architectural point of view, could you compare this device with a compatible add-on RAID controller with support for the usual RAID modes? What happens to this device if one of its component flash chips fails? I admit my bias tilts me towards my past experiences with inexpensive RAID controllers in RAID 0 mode, for MAX SPEED. In your opinion, would it be safer overall to wait for x4 and x8 NVMe RAID controllers and to connect 2x, 4x or more 2.5" NVMe SSDs, in order to exploit the redundancy obtained by RAID arrays in modes 5 and 6? Thanks! MRFS

In reply, Jeremy wrote:

"Check Al's review and comments for more indepth that I can really give you. That said there is orders of magnitude of difference in cost between the two solutions, you would have to desperately need this speed to justify the price."


Lastly, are there presently any x4 or x8 PCIe NVMe RAID controllers with multiple ports for that SFF-8639 connector?


May 23, 2015 | 04:03 PM - Posted by MRFS (not verified)

Here's the cabling topology that comes to my mind:


We bought one of these SATA-II 3G RAID controllers several years ago:
(4 per Mini-SAS channel up to 16 (SATA I or II Hard Drives) Internal Connectors)

p.s. is there an easy way to imbed graphics in a Reply here?


May 22, 2015 | 10:47 PM - Posted by MRFS (not verified)

If anyone is interested, here is a review with photos of a Supermicro server with room for multiple 2.5" NVMe SSDs:,2-878.html

Start reading at:

"NVMe Hot Swap Capabilities"


"NVMe has made a massive impact in the server space, specifically for applications where low latency and high queue depths are the norm. Applications such as databases and real-time analytics are seeing massive speed-ups from the technology."

"... the PCIe x4 2.5" form factor drives are made to fit into similar spaces as their SAS/SATA counterparts."

"One can see that these fit into standard Supermicro 2.5" to 3.5" converters so a major aspect of these drives is fitting into familiar infrastructure. These drives can be inserted and removed similar to traditional disks. Modern OSes are able to handle these drives and use them in hot swap applications such as RAID arrays."

May 23, 2015 | 09:22 AM - Posted by MRFS (not verified)

And so, as many prosumers have already done with 2 x 6G SSDs, we can reach your preferred capacity of 800GB with 2 x 400GB 2.5" Intel 750 SSDs in RAID 0.

Now, where do we find a host controller with at least 2 x SFF-8639 ports?

Am I dreaming (again)?


May 26, 2015 | 09:34 AM - Posted by MRFS (not verified)


Supermicro AOC-SLG3-2E4R NVMe AOC card, Standard LP, 2 internal NVMe ports, x4 per port, Gen-3

Only $150 at Newegg.

May 23, 2015 | 09:26 AM - Posted by MRFS (not verified)

NOTE the roadmap implied by "24 Gb/s"

There are multiple using generations based on performance.

12 Gb/s SFF-8637
24 Gb/s SFF-8638

May 23, 2015 | 09:36 AM - Posted by MRFS (not verified)

MSI Preparing SFF-8639 Adapter Card for Motherboards

"There is no (measurable) performance difference between a four-lane PCIe Gen 3 link routed via a PCIe expansion slot or an SFF-8639 connector. The biggest difference is compatibility; many small form factor and multi-VGA systems simply cannot surrender a PCIe slot to anything other than a graphics card, so housing an ultra-fast SSD elsewhere may be the only viable option."

May 23, 2015 | 09:45 PM - Posted by Richard (not verified)

I want one!

May 23, 2015 | 09:57 PM - Posted by Richard (not verified)

Great SSD

May 25, 2015 | 11:25 AM - Posted by MRFS (not verified)

JJ at ASUS says that heat is a factor with the 2.5" Intel 750:

I wonder if Icy Dock is preparing a 5.25" enclosure for 4 x Intel 750s?

The Icy Dock model Fits 7, 9.5, 12.5, 15mm height drive:

May 25, 2015 | 12:37 PM - Posted by MRFS (not verified)

For comparison purposes, we got these numbers from an inexpensive Highpoint RocketRAID x8 model 2720SGL PCIe RAID controller:

ATTO on 4 x Samsung 128GB model 840 Pro SSDs:

ATTO on 1 x Samsung 256GB 850 Pro SSD:

We are happy with these numbers, because the bulk of our I/O here is batch database updates e.g. drive images written to all data partitions, XCOPY updates to a large HTML database etc.

XCOPY also works fine over a LAN e.g.:

xcopy folder X:\folder /s/e/v/d

We've also experimented with OS hosting on the same RAID controller, using 4 x Samsung SSDs and also 4 x Intel SSDs: the 4 x Samsung 840 Pro on a PCIe 2.0 motherboard (ASUS P5Q Deluxe) are VERY SNAPPY, particularly with an overclocked quad-core Intel CPU.


May 25, 2015 | 01:05 PM - Posted by MRFS (not verified)

p.s. JJ reports "up to 1,200 MB/s [sequential] WRITE performance" (at 2:00 on the counter).


May 29, 2015 | 12:52 PM - Posted by mohamedkamel

Nice! I want one for my new build.

May 29, 2015 | 01:00 PM - Posted by TekHead

I WANT IT(^_^)

May 30, 2015 | 03:49 AM - Posted by vando1952

That would make mine the fastest PC in OZ!!!! LOL

May 31, 2015 | 05:42 AM - Posted by Anonymous (not verified)

Looks interesting. I hope the price will drop soon though.

May 31, 2015 | 05:42 AM - Posted by Anonymous (not verified)

Looks interesting. I hope the price will drop soon though.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Lines and paragraphs break automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd> <blockquote><p><br>
  • Web page addresses and e-mail addresses turn into links automatically.

More information about formatting options

This question is for testing whether you are a human visitor and to prevent automated spam submissions.