Samsung Unveils First PCIe 3.0 x4-Based M.2 SSD, Delivering Speeds of Over 2GB/s

Deathspawner writes: Samsung's SM951 is an unassuming gumstick SSD — it has no skulls or other bling — but it's what's underneath that counts: PCIe 3.0 x4 support. With that support, Samsung is able to boast speeds of 2,150MB/s read and 1,550MB/s write. But with such speeds comes an all-too-common caveat: you'll probably have to upgrade your computer to take true advantage of it. For comparison, Samsung says a Gen 2 PCIe x4 slot will limit the SM951 to just 1,600MB/s and 1,350MB/s (or 130K/100K IOPS), respectively. Perhaps now is a bad time to point out a typical Z97 motherboard only has a PCIe 2nd Gen x2 (yes, x2) connection to its M.2 slot, meaning one would need to halve those figures again.

Comments

PCIe 3.0 availability

[RogueyWon]

It's curious how many relatively recent high-end PCs from prestige-brands don't have PCIe 3.0 slots. Alienware are a particular offender here - they were very slow adopters, quite possibly because a lot of their customers don't actually think to check for this when speccing up a machine.

That said, it's questionable how much it really matters in the real world at the moment. Performance tests on the latest video cards (which can take advantage of PCIe 3.0) have found very little performance gap between 3.0 and 2.0 (and even 1.0) with the likes of the Nvidia 980. The gap is most apparent at extremely high (150+) framerates - which is unlikely to constrain the average gamer, who probably just turns up the graphical settings until his PC can't sustain his target framerate (probably somewhere in the 40-60fps rate) any more.

Re:PCIe 3.0 availability

[Kjella]

With the new Haswell-E processors, the CPU has 40 lanes of PCIe x4. So on a lot of high end x99 motherboards you'll see four PCIe gen3 x16 slots. However, since the CPU only has 40 lanes, this means not all of those "x16" slots are truely using 16 lanes of PCIe. Normally when four cards are plugged in, you'll get slot 0 running at x16, and the other three slots running at x8.

That's not really correct, high end motherboards usually have PLX chips which act like PCIe switches. Like the motherboard the GP listed runs at 16x/16x/16x/16x (or 16x/8x/8x/8x/8x/8x/8x) and only the total is limited to 40.

Re:PCIe 3.0 availability

[drinkypoo]

This SSD is not a PCIe slot form-factor card like you may be used to seeing. It is a M.2 form-factor SSD (see the picture of the drive in the articles). So having a x16 or even a x4 standard PCIe plugin slot will not help at all. The motherboard has to list that has a M.2 slot that is capable of PCIe gen3 speeds.

Bollocks. An M2 to PCIE adapter is twenty bucks.

If you don't want to upgrade your box

[azav]

Then max out the RAM and create a RAM drive.

On my 2010 iMac, I have a 16 GB RAM drive that gets between 3 and 4 GB/s and still have 16 GB of RAM available for my apps.

Check this terminal command out before entering it just to be safe.

diskutil erasevolume HFS+ 'RAM Disk' `hdiutil attach -nomount ram://33554432`

Under Mac OS 10.6.8, and 10.9 the above creates a 17 GB RAM disk.

diskutil erasevolume HFS+ 'RAM Disk' `hdiutil attach -nomount ram://8388608`

This creates a 4.27 GB RAM disk. Enjoy the speed.

Re:If you don't want to upgrade your box

[Anonymous Coward]

No. You should RARELY do this. If you go back and forth from other tasks where you can expect the cache to be re-used and need the absolutely best performance when you come back to those files, then something like this is a good idea. You're essentially committing content to RAM for cases where you know better than your operating system's optimizations.

Re:If you don't want to upgrade your box

[ledow]

Er... welcome to business level warranties.

Which run quite a bit less on business kit when you're buying £300 PCs instead of £2k Macs.

Sorry, but the Apple "service" isn't - it's often a "replace with new". Any business can do that and wait while the replacement comes back from the manufacturer.

Macs at developer courses? Sure. But most of those developers will be doing web-stuff mainly, or forced to use Mac if they want their stuff to compile and work ON a Mac as an end-result.

And "same day" only works if you're a) near an Apple store, b) it's open, and c) at the goodwill of Apple. Not something to base your business on versus a 4hr on-site service response with 24-hour turnaround for less than half the price difference between a PC and Mac of equivalent spec.

Re:If you don't want to upgrade your box

[ledow]

Last time I used a RAM drive, it was on the contents a floppy disk. My brother was sick of slow compile times and worked out how to use the university DOS computers to produce a RAM drive. Autoexec.bat created it and copied his files into it, and then it ran like greased lightning.

But that was back when 1.44Mb of RAM was a lot and he was lucky enough to be somewhere where every computer had that spare.

Last time I saw it when when making a single-floppy Linux distribution that copied itself into RAM because it was often used on diskless workstations. Just like almost every Ubuntu install disk can do now if you select Live CD from the boot menu.

But on ordinary desktop OS? Since Windows 95, RAMDisks have been dead. Since then, we've been using RAM better to cache all recent filesystem accesses. There's very, very, very, very little that will ever benefit from a RAMDisk over just having that RAM as filesystem cache automatically anyway. You still have to read the data from permanent storage anyway, and once you've done that, it's in RAM until you start to fill up RAM. Read it often enough and it will never drop out of the cache. If you're not reading it often enough, why the hell bother to RAMDisk it?

And you lose NOTHING if the machine dies mid-way. With a RAMDisk, any changes you make are gone.

Please. Stop spreading absolute "gold-plated-oxygen-free" junk advice like this.

Anyone who wants to do this can do it with any bit of freeware on any machine. But why they would bother is beyond me. Hell, next you'll be telling me to enable swapfiles and put them on the RAMDisk....

Re:If you don't want to upgrade your box

[drinkypoo]

Not something to base your business on versus a 4hr on-site service response with 24-hour turnaround for less than half the price difference between a PC and Mac of equivalent spec.

Yeah, I had one of those 24-hour warranties, from HP, on a $2000 Elitebook. First of all, you only get 24-hour turnaround if you are located within one of their service areas, if they have to drive more than two hours to reach you then you get service when they get around to it. After they show up, days after you called, if they botch the repair (they only bring enough parts to make whatever the knowledge base says are the most common repairs) then you have to wait some more days while they get more parts and try again. If that fails, you finally get a new machine. The up side is that for anyone but Apple, the models change enough to where near the end of a three year warranty, they won't have back stock to replace it. So they'll replace the unit with a refurb'd unit of another model with all equal or superior specifications. I went from a mediocre Core Duo to one of the higher-end C2Ds. Then I sold it and bought a fleet of netbooks. None of them are HPs.

Re:If you don't want to upgrade your box

[azav]

Well, shame on me. I've been doing it for 3 years on a daily basis.

I have my RAM drive rsynched to an SSD partition that is the same-ish size.

And here's one area where you're incorrect. Safari loads web pages. Each page loads javascript. Many of these leak over time or simply never purge their contents. I often end up with 8 GB used in Safari. Safari alone is a citizen that doesn't play by these rules because each page that loads is a prisoner of the javascript that loads and often doesn't handle memory freeing properly.

When I use my RAM as a drive, I get near INSTANT builds on OS X.

This matters to me more than your claims of "all modern operating systems taking full advantage of the RAM". If the operating system takes full advantage of the RAM, it may not be to my best benefit.

For example, Apple apps now by default do not quit when you close the last document. They merely stay in memory, hide the UI and then need to be relaunched to enable the UI again. Why does this matter? For TextEdit, if I want to open a document form the open menu if I close the last document and click elsewhere, this forces me to reopen the app because the OS fake closes the app (really only hiding the UI) while the rest of the app stays memory resident.

So, I have to relaunch the app. This takes more time and ONLY just renables the UI. How much memory does this save on my 32 GB machine? 1 MB. Now, that's certainly not taking full advantage of the RAM. It's a case of the OS designers thinking that "he wanted to quit the app, so we'll do it for him". But I didn't want to quit the app. The computer is not taking full advantage of the RAM in this case. That's not what I wanted it to do.

Maybe I have apps in the background that are doing stuff, but I want them to pause completely if another app is running in the foreground. Maybe I want ALL Safari pages to suspend their javascript when in the background, but the app still can still process downloads as if it's running at normal priority.

See, there are many cases where the computer's OS will not take proper advantage of the RAM and the processing power since it can not mirror the user's intentions. Even in cases where it tries to, it often gets them wrong. And in some cases where it does (Safari javascript), the computer ends up eating processing power and RAM for tasks that the user doesn't want it to be placing priority on. And in some of these cases, it can't allocate RAM and processing power properly, because it can't if it relies on other programmers writing their javascript competently and acting as good citizens.

I can cordon off a small chunk of my computer's RAM (since I have way more than enough) and direct it to do pretty damn much just want I want it to do.

That's why I bought it. I don't want the OS to prioritize things the way it wants to. I want to tell (parts of) the OS to prioritize things the way I want it to.

Cheers.

Re:If you don't want to upgrade your box

[guruevi]

There exists no PC that costs $300 that will match up to a $2k Mac. Even if you plunk down $700 for a Mac Mini with AppleCare, it will be hard to find a similar machine with a similar service contract (think Dell Gold Service Contract).

Apple will come to you within 24h or ship you a new machine overnight but even after the warranty expires you can still call them and they will answer you. I have dealt with Dell, HP and Lenovo, it doesn't even come close.

Re:Starts to get reasonable

[geantvert]

No! I would say: FORTUNATELY I don't think that's going to happen.

2000MB/s is still one or two orders of magnitude slower than DDR3/4 (between 20000 and 60000MB/s) so I clearly don't want direct mapping.
Let the OS cache the SSD sectors in RAM pages and everything will be fine (and from the user point of view, that won't change anything)

"just 1,600MB/s and 1,350MB/s"

[Red_Chaos1]

Because that's just so terrible, right? :p

Re:Starts to get reasonable

[Lunix Nutcase]

There's nothing unfortunate about it. Access times for SSD is around .1 ms vs at worst around 15 nanoseconds for DDR3 RAM. You do realize how significant of a performance impact that would be, right?

Daring us to use it

[WaffleMonster]

How long can you sustain these kinds of I/O rates before burning the thing out?

Awesome it is so fast yet like LTE with tiny data caps utility appears to be substantially constrained by limitations on use.

For subset of people with workloads actually needing this kind of performance how useful is this? Reads can be cached by DRAM which is quite cheap.

For those who don't really need it I can understand how it would be nice to have.

Re:Daring us to use it

[grumpy_old_grandpa]

> How long can you sustain these kinds of I/O rates before burning the thing out?

If you were to sustain 1550 MByte/s write for 1 year, you'd write a total of 48 PB. (1550*60*60*24*365/1000/1000/1000), or 0.13 PB/day. In Techreport's endurance test, only two drives made it past 1.5 PB. So, if that is the bar, the drive would last only 11 days.

However, that would give you no time to read the data you'd written. Since you're not likely to write at max speed 24/7, the drive should last considerably longer.

http://techreport.com/review/2...

Don't overlook NVME!

[Anonymous Coward]

The big news here is that this is pretty much the first consumer available SSD that supports NVME! There are some super-expensive pro devices that do NVME but they likely alone cost more than your whole high-end gaming rig.

http://en.wikipedia.org/wiki/NVM_Express

NVME is the interface that replaces AHCI, which was designed for spinning rust devices that can really only read and write one bit at a time. Flash based devices don't have to wait for moving parts and thus can access many things at once.

AHCI was designed for magnetic drives attached to SATA. NVME is explicitly designed to accommodate fast devices directly connected to PCI express. Take a look at the comparison table on the wikipedia page linked above. Multiple, deep, queues and lots of other features to remove bottlenecks that don't' apply to flash based storage.

How useful NVME currently is to consumers, though, is different. Only really new operating systems can boot from NVME devices. (Windows 8.1 or later. Don't know the current state of linux support but I bet at least someone's got a patched version of the kernel and grub if there's not mainline support already) And most most motherboards don't properly support NVME booting yet either. (Ive heard reports that some do with a BIOS/firmware update but it's currently really spotty)

Don't get too excited

[Solandri]

Samsung is able to boast speeds of 2,150MB/s read and 1,550MB/s write.

1. These are sequential speeds. They're only relevant when you're dealing with large files. Unless your job is working with video or disk images or other large files, the vast majority of your files are going to be small, and the IOPS matters more. 130k/100k IOPS is really good, but only about a 10%-20% improvement over SATA3 SSDs. It translates into 520/400 MB/s at queued 4k read/writes best case. Current SATA3 drives are already surpassing 400 MB/s queued 4k read/writes.

2. Like car MPG, the units here are inverted from what actually matters. You don't say "gee, I have 5 gallons in the tank I need to use today, how many miles can I drive with it?", which is what MPG tells you. You say "I need to drive 100 miles, how many gallons will it take?" which is gal/100 miles. Yes they're just a mathematical inverse, but using the wrong one means the scaling is not linear. If you've got a 100 mile trip:

A 12.5 MPG vehicle will use 8 gallons
A 25 MPG vehicle will use 4 gallons (a 4 gallon savings for a 12.5 MPG improvement)
A 50 MPG vehicle will use 2 gallons (a 2 gallon savings for a 25 MPG improvement)
A 100 MPG vehicle will use 1 gallon (a 1 gallon savings for a 50 MPG improvement)

See how the fuel saved is inversely proportional to the MPG gain? As you get higher and higher MPG, it matters less and less because MPG is the wrong unit. If you do it in gal/100 miles it's linear. (This is why the rest of the world uses liters / 100 km.)

An 8 gal/100 mile vehicle will use 8 gallons.
A 4 gal/100 mile vehicle uses 4 gallons (a 4 gallon savings for a 4 gal/100 mi improvement)
a 2 gal/100 mile vehicle uses 2 gallons (a 2 gallon savings for a 2 gal/100 mi improvement)
a 1 gal/100 mile vehicle uses 1 gallon (a 1 gallon savings for a 1 gal/100 mi improvement)

The same thing is true for disk speeds. Unless you've got a fixed amount of time and need to transfer as much data as you can in that time, MB/s is the inverse of what you want. The vast majority of use cases are a fixed amount of MB that needs to be read/written, and the time it takes to do that is what you're interested in because that's time you spend twiddling your thumbs. If a game needs to read 1 GB to start up:

A 100 MB/s HDD will read it in 10 sec
a 250 MB/s SATA2 SSD will read it in 4 sec (a 6 sec savings for a 150 MB/s improvement)
A 500 MB/s SATA3 SSD will read it in 2 sec (a 2 sec savings for a 250 MB/s improvement)
A 1 GB/s PCIe SSD will read it in 1 sec (a 1 sec savings for a 500 MB/s improvement)
This 2 GB/s PCIe SSD will read it in 0.5 sec (a 0.5 sec savings for a 1000 MB/s improvement

Again, the actual time savings is inverted from the units we're using to measure. We really should be benchmarking HDDs and SSDs by sec/MB.

A 10 sec/MB HDD will read 1 GB in 10 sec
A 4 sec/MB SATA2 SSD will read it in 4 sec (a 6 sec savings for a 6 sec/MB improvement)
A 2 sec/MB SATA3 SSD will read it in 2 sec (a 2 sec savings for a 2 sec/MB improvement)
A 1 sec/MB PCIe SSD will read it in 1 sec (a 1 sec savings for a 1 sec/MB improvement)
This 0.5 sec/MB PCIe SSD will read it in 0.5 sec (a 0.5 sec savings for a 0.5 sec/MB improvement)

That's nice and linear. You see that the vast majority of your speedup comes from switching from a HDD to a SSD - any SSD, even the old slow first gen ones. The next biggest savings is switching to a SATA3 SSD. Beyond that the extra speed is nice, but don't be mislead by the huge MB/s figures - the speedup from PCIe drives will never be as big as those first two steps from a HDD to a SATA SSD. Manufacturers just report performance in MB/s (instead of sec/MB) because it exaggerates the importance of tiny increases in time saved, and thus helps them sell new and improved (and more expensive) products. Review sites also report in MB/s because if you report in sec/MB, the benchmark graphs are boring and the speedup from these shiny new SSDs is barely perceptible.