Samsung 950 Pro 256GB RAID Report

What Is RAID And Initial Performance Testing

A Redundant Array of Independent (formally Inexpensive) Disks, or RAID, is often used to combine many physical drives to make one logical volume in an effort to increase performance, capacity and/or data integrity. A case was first made for the technology in a June 1988 white paper called "A Case for Redundant Arrays of Inexpensive Disks (RAID)," presented at the SIGMOD conference. In that paper, the authors suggested the highest-performance mainframe disks could be outperformed by an array of inexpensive drives more commonly found in the PC space. Although failures would rise in proportion to the number of physical devices, by configuring for redundancy, the reliability of an array could far exceed that of any single repository. Prior to this report, five levels of RAID were already in use from various companies, but the technology had yet to be standardized.

The level we're using today is RAID 0. This joins two or more drives to increase performance and storage capacity. The downside to RAID 0 is that it multiplies the risk of a failure. If one drive in a two-disk array dies, the data on both disks is lost.

RAID In The Home PC

The year 2000 was an interesting one for PC enthusiasts. The Parallel ATA (PATA) bus' ceiling rose from 33 MB/s to 66 MB/s. Intel's wildly popular 440BX chipset didn't support the newer specification, so add-in card manufacturers scrambled to support it through discrete controllers. Promise was one such company, and it built two corresponding products: the Ultra66 (an HBA without RAID support) and the FastTrak66 (a PATA RAID controller). Other than a single resistor, the two were identical. Their firmware was unique, but that didn't stop power users from soldering on the missing surface-mount component and flashing the lower-cost card with the flagship's RAID firmware. You can read more about that procedure from Tom's Hardware founder Tom Pabst.

RAID Goes Mainstream

Shortly after the Promise Ultra66 to FastTrak66 mod became popular, manufacturers added RAID support to overclocking-friendly motherboards like Abit's BE6-II. By then, high-end copper-based air coolers and liquid-cooling kits emerged to take Intel's highly-tweakable Coppermine CPUs to new heights. The $300 Pentium III 500MHz could hit 750MHz right out of the box with minimal effort. That 750MHz model should have cost three times more than the 500MHz version. It was a good time to be an enthusiast. Increasing storage performance through RAID was just the next logical step.

We've come a long way since the 440BX. Most modern motherboards have RAID support built right into the chipset. RAID for PCIe-based devices is another matter entirely. It's a newer concept that leverages concepts standardized in the past.

Today, only Intel's Z170 chipset supports booting an operating system from an array of PCIe-attached drives. But not all Z170-based platforms enable the feature. ASRock's Z170 Extreme 7+ goes a step further than most. It supports three M.2 drives for RAID 0 (two-drive performance increase), RAID 1 (two-drive mirroring) and RAID 5 (three-drive performance and parity protection). In this review, we're using the Z170 Extreme 7+ with two 950 Pros in RAID 0 to compare against a single, larger 950 Pro. Our array settings are shown in the image above; we're testing with a 16KB stripe size.

Sequential Read Performance

We're including a few alternatives in these charts, including Samsung's 850 Pro and SanDisk's Extreme Pro, the fastest SATA-based client SSDs out there. SATA only scales to a queue depth of 32, so we can't generate data out to 128 commands deep like the PCIe-based drives (they can actually go to 256 with another 256 commands per queue).

Samsung's 256GB 950 Pro delivers just over 2000 MB/s sequential read performance. The Z170's storage controller communicates with the CPU through the Direct Media Interface (DMI), which shares bandwidth with several other devices. The DMI just isn't capable of doubling the speed of a single 950 Pro because it's limited to less than 4 GB/s between the PCH and CPU.

Sequential Write Performance

The 256GB 950 Pro's sequential write speeds differ from the 512GB model. The higher-capacity model can write at 1500 MB/s, while the smaller drive is limited to 900 MB/s, per Samsung's specifications. Both of the striped 950 Pro arrays double the performance of a single drive, and the two 256GB drives outperform one 512GB 950 Pro by nearly 400 MB/s.

Random Read Performance

The array of 256GB 950 Pros achieves higher random read performance than a single drive, but the increase is only quantifiable at high queue depths. In a normal desktop environment, you simply won't see the benefit. In fact, at low queue depths, where most workloads can be characterized, the RAID array is actually a little slower than a single drive.

Random Write Performance

We also observe lower random write performance at low queue depths in RAID 0. Configuring the DMI for RAID adds latency, which cuts into measurable IOPS. At queue depths of four and up, the array can use the extra device bandwidth.

This thread is closed for comments
26 comments
    Your comment
  • Amdlova
    Insane speed. but no gains on WOW.
  • firefoxx04
    Why do you expect gains in video games when everything relevant is loaded in RAM?
  • maxxxt
    Not supported in win 7 SP1. Probably should let people know that too.
  • anort3
    I'm still amazed by how fast my 512GB 950 Pro is.
  • USAFRet
    211300 said:
    Why do you expect gains in video games when everything relevant is loaded in RAM?


    A lot of people do.
    Assuming that the performance gains we saw with spinning disks in RAID 0 automagically does the same with SSD's. It does not.
  • Tibeardius
    Did it have any sort of thermal throttling occur? These pcie m.2 drives can get pretty hot.
  • Integr8d
    "Once upon a time, you could sling a couple of Western Digital Raptors together, fire up a level in Battlefield 2 before anyone else, get the plane and dominate the map."

    Someone just explained 24 months of my life:)
  • HT
    Quote:
    Why do you expect gains in video games when everything relevant is loaded in RAM?

    that's the point of faster drives, loading it all in ram. do you think it magically appears there by itself ?
  • HT
    good article Chris, i'm intrigued by your statement of the samsung driver vs the M$ one, i would've liked to see some numbers comparing the two.
  • Virtual_Singularity
    Interesting article. Also: am a lil' dumbfounded at how much the price of the 850 pro series has dropped since the holidays, a mere 3+ months ago...
  • jt AJ
    Quote:
    good article Chris, i'm intrigued by your statement of the samsung driver vs the M$ one, i would've liked to see some numbers comparing the two.


    iirc i was told samsung driver is better than MS but that might change when MS updates. tbh i dont believe in samsung's software too much.

    overall great article, it is very sad to see going to raid 0 loses performance in most important area, 4k read/4k write takes huge hit and almost no point going into that exception is sequential read. it maybe good for games or video editing, where as everything else its better to go for 1 drive, kinda sad.

    maybe when optane comes out we'll see this problem solved as driver has more time to mature?
  • RedJaron
    1690815 said:
    Quote:
    Why do you expect gains in video games when everything relevant is loaded in RAM?
    that's the point of faster drives, loading it all in ram. do you think it magically appears there by itself ?

    All that does is load levels faster. It doesn't translate to faster in-game framerates. A few world streaming games won't see the stutter when you open a door in a building. But your post could confuse others. A SSD doesn't necessarily "load it all in RAM." That's up to your RAM capacity and the game engine of how many resources get sent to RAM.
  • joex444
    The only difference a drive can make to a game is the load time, and guess what? This article showed the service time, not FPS. There's still no difference. This clearly means that the read speed isn't limited by the storage medium but by how quickly the CPU can parse the binary data and make sense of it. You'd need a faster CPU to lower the service time, not a faster SSD. And again, I'd like to see a comparison against a HDD here just so we get a sense of what an SSD does as the graph says "All SSDs perform equally well" but doesn't prove that SSDs help any relative to a HDD. For sequentially read data, HDDs push 120-160MB/s, if that's already beyond the limit of the CPU for parsing that data then the SSD shouldn't improve any (or the data is not stored sequentially to begin with, however most games use a small number of very large binary files which should be defragmented such that they are sequential).
  • skrewler
    You write "It supports three M.2 drives for RAID 0 (two-drive performance increase)...".

    But RAID 0 is really n-drive performance increase. Why didn't you test with 3 drives as the motherboard supports it?
  • jt AJ
    Quote:
    The only difference a drive can make to a game is the load time, and guess what? This article showed the service time, not FPS. There's still no difference. This clearly means that the read speed isn't limited by the storage medium but by how quickly the CPU can parse the binary data and make sense of it. You'd need a faster CPU to lower the service time, not a faster SSD. And again, I'd like to see a comparison against a HDD here just so we get a sense of what an SSD does as the graph says "All SSDs perform equally well" but doesn't prove that SSDs help any relative to a HDD. For sequentially read data, HDDs push 120-160MB/s, if that's already beyond the limit of the CPU for parsing that data then the SSD shouldn't improve any (or the data is not stored sequentially to begin with, however most games use a small number of very large binary files which should be defragmented such that they are sequential).


    exactly, a faster cpu and a very well coded game to take advantage of newer instruction would greatly help for sure. SSD right now with raid 0 almost doubles the sequential throughput but theres very little advantage to it in using windows. gaming, video editing thats about it, and sequential file copying? unsure about image back up/restoring.

    although OS it self has a lot of files that reads sequential most are random performance, in a case of NVMe PCIE raid 0 it's almost pointless because QD1 to QD2 performance actually drops in comparison to two drives.. like wtf samsung?
  • jt AJ
    Quote:
    You write "It supports three M.2 drives for RAID 0 (two-drive performance increase)...". But RAID 0 is really n-drive performance increase. Why didn't you test with 3 drives as the motherboard supports it?


    there are aritcles out there tested raid 0 with 3 SSD on another website i think pcper? but still the same, sequential goes way up, random performance drop, kinda sad.
  • CRamseyer
    Quote:
    The only difference a drive can make to a game is the load time, and guess what? This article showed the service time, not FPS. There's still no difference. This clearly means that the read speed isn't limited by the storage medium but by how quickly the CPU can parse the binary data and make sense of it. You'd need a faster CPU to lower the service time, not a faster SSD. And again, I'd like to see a comparison against a HDD here just so we get a sense of what an SSD does as the graph says "All SSDs perform equally well" but doesn't prove that SSDs help any relative to a HDD. For sequentially read data, HDDs push 120-160MB/s, if that's already beyond the limit of the CPU for parsing that data then the SSD shouldn't improve any (or the data is not stored sequentially to begin with, however most games use a small number of very large binary files which should be defragmented such that they are sequential).


    We don't show HDDs because the results are very different. If you want I can make a post with all of the benchmarks in the forums in a few days. I have the data, I just need to build the charts. Give me a couple of days and I'll whip the charts up.
  • CRamseyer
    Quote:
    You write "It supports three M.2 drives for RAID 0 (two-drive performance increase)...". But RAID 0 is really n-drive performance increase. Why didn't you test with 3 drives as the motherboard supports it?


    We only had two identical drives at the time. We now have three drives. I have a review coming with three drives in RAID 5.
  • RedJaron
    1888934 said:
    We don't show HDDs because the results are very different. If you want I can make a post with all of the benchmarks in the forums in a few days. I have the data, I just need to build the charts. Give me a couple of days and I'll whip the charts up.

    This would actually be great, Chris. I don't think you need a huge comprehensive list, just a simple one that shows the reviewed drive against a couple of the more popular current SSDs and some common 7200 and 5400 3.5" drives ( a WD Black, Blue, and Green should be good ). That'd give a good idea of the relative real-world performance gain.

    On a slightly unrelated tangent, I would love to see a new HDD review. I think it'd be great to see how much spindle drives have improved in the last few years. I mean with platter density and other improvements, can current 5400 drives match or outperform older 7200 drives?
  • Dax corrin
    Quote:
    "Once upon a time, you could sling a couple of Western Digital Raptors together, fire up a level in Battlefield 2 before anyone else, get the plane and dominate the map." Someone just explained 24 months of my life:)


    I did that with 3 74 GB Raptors in Battlefield 2 before I could afford an SSD big enough.
  • skrewler
    Quote:
    Quote:
    You write "It supports three M.2 drives for RAID 0 (two-drive performance increase)...". But RAID 0 is really n-drive performance increase. Why didn't you test with 3 drives as the motherboard supports it?
    We only had two identical drives at the time. We now have three drives. I have a review coming with three drives in RAID 5.


    Looking forward to it. Would be great if you could include 3 drives in RAID 0 too! I've got an Asus board right now, but as I snagged up 2x 512GB 950 Pros the other day at $270/each I'm thinking of switching to the Asrock motherboard and buying one more 950 pro.

    It would also be nice if you could compare the 850 evo and Crucial (MX something). It would be interesting to see how cheaper drives perform in RAID 0 versus the 850 Pro.

    edit: the m.2 editions of the 850 evo and crucial obviously.
  • skrewler
    has anyone used one of those msata -> m.2 adapters? do they affect performance at all? i wanted a 1tb m.2 drive but samsung only offers the 850 evo for msata.
  • LanVlam
    It makes no sense because the article says the sata3 speed is the bottleneck at the point the cpu communicates with the controller. I have 1 256gb evo 850 and it capped at 500mb. My sabertooth z77 died and i play on an i7 920 @ 2.8ghz and everything runs great even i lack pci-e 3 (gtx970) and have steady 240mb/s because sata 2. Asus p6t deluxe. I am amazed though i can still play gta v on great quality at 60fps. Msi afterburner tells me the pci-e bus is 88% max used. Great that i kept that mb. I only use 3x4gb 1600mhz ram in triple channel. Ivy and sandy were 4x4gb in dual
  • skrewler
    2215564 said:
    It makes no sense because the article says the sata3 speed is the bottleneck at the point the cpu communicates with the controller. I have 1 256gb evo 850 and it capped at 500mb. My sabertooth z77 died and i play on an i7 920 @ 2.8ghz and everything runs great even i lack pci-e 3 (gtx970) and have steady 240mb/s because sata 2. Asus p6t deluxe. I am amazed though i can still play gta v on great quality at 60fps. Msi afterburner tells me the pci-e bus is 88% max used. Great that i kept that mb. I only use 3x4gb 1600mhz ram in triple channel. Ivy and sandy were 4x4gb in dual



    I meant using the 850 Evo m.2 versions, not SATA...