Performance Consistency

We've been looking at performance consistency since the Intel SSD DC S3700 review in late 2012 and it has become one of the cornerstones of our SSD reviews. Back in the days many SSD vendors were only focusing on high peak performance, which unfortunately came at the cost of sustained performance. In other words, the drives would push high IOPS in certain synthetic scenarios to provide nice marketing numbers, but as soon as you pushed the drive for more than a few minutes you could easily run into hiccups caused by poor performance consistency. 

Once we started exploring IO consistency, nearly all SSD manufacturers made a move to improve consistency and for the 2015 suite, I haven't made any significant changes to the methodology we use to test IO consistency. The biggest change is the move from VDBench to Iometer 1.1.0 as the benchmarking software and I've also extended the test from 2000 seconds to a full hour to ensure that all drives hit steady-state during the test.

For better readability, I now provide bar graphs with the first one being an average IOPS of the last 400 seconds and the second graph displaying the standard deviation during the same period. Average IOPS provides a quick look into overall performance, but it can easily hide bad consistency, so looking at standard deviation is necessary for a complete look into consistency.

I'm still providing the same scatter graphs too, of course. However, I decided to dump the logarithmic graphs and go linear-only since logarithmic graphs aren't as accurate and can be hard to interpret for those who aren't familiar with them. I provide two graphs: one that includes the whole duration of the test and another that focuses on the last 400 seconds of the test to get a better scope into steady-state performance.

Steady-State 4KB Random Write Performance

Given the higher over-provisioning and an enterprise-oriented controller, it's no surprise that the SSD 750 has excellent steady-state random write performance. 

Steady-State 4KB Random Write Consistency

The consistency is also very good, although the SSD 750 can't beat the 850 Pro if just focusing on consistency. When considering that the SSD 750 provides nearly three times the performance, it's clear that the SSD 750 is better out of the two. 

Intel SSD 750 1.2TB (PCIe 3.0 x4 - NVMe)

At the initial cliff the performance drops to around 15K IOPS, but it quickly rises and seems to even out at about 22-23K IOPS. It actually takes nearly an hour for the SSD 750 to reach steady-state, which isn't uncommon for such a large drive but it's still notable. 

I couldn't run tests with added over-provisioning because NVMe drives don't support the usual ATA commands that I use to limit the LBA of the drive. There is similar command set for NVMe as well, but I'm still trying to figure out how to use them as there's isn't too much public info about NVMe tools.

Intel SSD 750 1.2TB (PCIe 3.0 x4 - NVMe)
Introduction, The Drive & The Test AnandTech Storage Bench - The Destroyer
Comments Locked


View All Comments

  • knweiss - Thursday, April 2, 2015 - link

    Kristian, you wrote "for up to 4GB/s of bandwidth with PCIe 3.0 (although in real world the maximum bandwidth is about 3.2GB/s due to PCIe inefficiency)". Is this really true? PCIe 2.0 uses 8b/10b encoding with 20% bandwidth overhead which would match your numbers. However, PCIe 3.0 uses 128b/130b encoding with only 1.54% bandwidth overhead. Could you please explain the inefficiency you mentioned? Thanks in advance!
  • DanNeely - Thursday, April 2, 2015 - link

    The real world number includes the bandwidth consumed by PCIe packet headers, NVME packet headers, NVME command messages, etc. Those are over and above the penalty from the encoding scheme on the bus itself.
  • IntelUser2000 - Thursday, April 2, 2015 - link

    The 4GB bandwidth takes into account the encoding scheme.

    Each lane of v1 PCI-Express had 2.5GT/s so with 8b/10b encoding you end up with 2.5G/10 = 250MB/s. Quadruple that for four lanes and you end up with 1GB/s.

    v2 of PCI-Express is double that and v3 of PCI-Express is further double that and there is the 4GB number.
  • aggrokalle - Thursday, April 2, 2015 - link

    i'm interrested in this as how many nand-channels got the 1.2tb and 400gb version Kristian?
  • tspacie - Thursday, April 2, 2015 - link

    Was there an approximate release date?
  • gforce007 - Thursday, April 2, 2015 - link

    When will these be available for purchase? Also I have a m.2 slot on my motherboard (z10PE-D8 WS) Id rather utilize the 2.5 15mm form factor. I am a bit confused. I dont think that board has SFF-8639. Is there an adapter. Will that affect performance? I assume so and by how much?
  • knweiss - Thursday, April 2, 2015 - link

    The motherboard (host) end of the cable has a square-shaped SFF-8643(!) connector. E.g. ASUS ships an M.2 adapter card for the X99 Sabertooth that offers a suitable port. SFF-8639 is on the drive's end.
  • emn13 - Thursday, April 2, 2015 - link

    That endurance number is scarily low for a 1.2TB drive. 70GB a day for 5 years - thats about 128 TB of writes total, and that's just 100 drive writes! Put another way, at around 1GB/sec (which this drive can easily do), you'd reach those 100 drive writes in just 36 hours.

    Of course, that's an extremely intensive workload, but I sure hope this is just intel trying to avoid giving any warrantee rather than an every remotely realistic assessment of the drives capabilities.
  • p1esk - Thursday, April 2, 2015 - link

    This is a consumer drive. What's your use case where you write more than 70GB a day?
  • juhatus - Friday, April 3, 2015 - link

    Raw 4k video and its not even close to being enough.

    At 4K (4096 x 2160) it registers 1697 Mbps which equals 764 GB/hour of 4K video footage. A single camera large Hollywood production can often shoot 100 hours of footage. That’s 76 TB of 4K ProRes 4444 XQ footage.

    The upcoming David Fincher film GONE GIRL crept up on 500 hours of raw footage during its multi camera 6K RED Dragon production. That equates to roughly 315 TB of RED 6K (4:1) footage. Shit just got real for data management and post production workflows.

Log in

Don't have an account? Sign up now