Intel SSD 750 PCIe SSD Review: NVMe for the Client
by Kristian Vättö on April 2, 2015 12:00 PM ESTAnandTech Storage Bench - The Destroyer
The Destroyer has been an essential part of our SSD test suite for nearly two years now. It was crafted to provide a benchmark for very IO intensive workloads, which is where you most often notice the difference between drives. It's not necessarily the most relevant test to an average user, but for anyone with a heavier IO workload The Destroyer should do a good job at characterizing performance.
| AnandTech Storage Bench - The Destroyer | ||||||||||||
| Workload | Description | Applications Used | ||||||||||
| Photo Sync/Editing | Import images, edit, export | Adobe Photoshop CS6, Adobe Lightroom 4, Dropbox | ||||||||||
| Gaming | Download/install games, play games | Steam, Deus Ex, Skyrim, Starcraft 2, BioShock Infinite | ||||||||||
| Virtualization | Run/manage VM, use general apps inside VM | VirtualBox | ||||||||||
| General Productivity | Browse the web, manage local email, copy files, encrypt/decrypt files, backup system, download content, virus/malware scan | Chrome, IE10, Outlook, Windows 8, AxCrypt, uTorrent, AdAware | ||||||||||
| Video Playback | Copy and watch movies | Windows 8 | ||||||||||
| Application Development | Compile projects, check out code, download code samples | Visual Studio 2012 | ||||||||||
The table above describes the workloads of The Destroyer in a bit more detail. Most of the workloads are run independently in the trace, but obviously there are various operations (such as backups) in the background.
| AnandTech Storage Bench - The Destroyer - Specs | ||||||||||||
| Reads | 38.83 million | |||||||||||
| Writes | 10.98 million | |||||||||||
| Total IO Operations | 49.8 million | |||||||||||
| Total GB Read | 1583.02 GB | |||||||||||
| Total GB Written | 875.62 GB | |||||||||||
| Average Queue Depth | ~5.5 | |||||||||||
| Focus | Worst case multitasking, IO consistency | |||||||||||
The name Destroyer comes from the sheer fact that the trace contains nearly 50 million IO operations. That's enough IO operations to effectively put the drive into steady-state and give an idea of the performance in worst case multitasking scenarios. About 67% of the IOs are sequential in nature with the rest ranging from pseudo-random to fully random.
| AnandTech Storage Bench - The Destroyer - IO Breakdown | |||||||||||
| IO Size | <4KB | 4KB | 8KB | 16KB | 32KB | 64KB | 128KB | ||||
| % of Total | 6.0% | 26.2% | 3.1% | 2.4% | 1.7% | 38.4% | 18.0% | ||||
I've included a breakdown of the IOs in the table above, which accounts for 95.8% of total IOs in the trace. The leftover IO sizes are relatively rare in between sizes that don't have a significant (>1%) share on their own. Over a half of the transfers are large IOs with one fourth being 4KB in size.
| AnandTech Storage Bench - The Destroyer - QD Breakdown | ||||||||||||
| Queue Depth | 1 | 2 | 3 | 4-5 | 6-10 | 11-20 | 21-32 | >32 | ||||
| % of Total | 50.0% | 21.9% | 4.1% | 5.7% | 8.8% | 6.0% | 2.1% | 1.4% | ||||
Despite the average queue depth of 5.5, a half of the IOs happen at queue depth of one and scenarios where the queue depths is higher than 10 are rather infrequent.
The two key metrics I'm reporting haven't changed and I'll continue to report both data rate and latency because the two have slightly different focuses. Data rate measures the speed of the data transfer, so it emphasizes large IOs that simply account for a much larger share when looking at the total amount of data. Latency, on the other hand, ignores the IO size, so all IOs are given the same weight in the calculation. Both metrics are useful, although in terms of system responsiveness I think the latency is more critical. As a result, I'm also reporting two new stats that provide us a very good insight to high latency IOs by reporting the share of >10ms and >100ms IOs as a percentage of the total.
In terms of throughput, the SSD 750 is actually marginally slower than the SM951, although when you look at latency the SD 750 wins by a large margin. The difference in these scores is explained by Intel's focus on random performance as Intel specifically optimized the firmware for high random IO performance, which does have some impact on the sequential performance. As I've explained above, data rate has more emphasis on large IO size transfers, whereas latency treats all IOs the same regardless of their size.
The number of high latency IOs is also excellent and in fact the best we have tested. The SSD 750 is without a doubt a very consistent drive.
Facebook
Twitter
RSS
132 Comments
View All Comments
KAlmquist - Thursday, April 2, 2015 - link
It's too bad that Anandtech didn't benchmark the 400 GB model, since that's the one most people are going to be most interested in buying. I assume that it's a case of Intel not making the 400 GB model available for review, rather than Anandtech deciding not to review it.jwilliams4200 - Thursday, April 2, 2015 - link
Agreed, the 400 GB model is more interesting to consumers.Also, I hope that if Anandtech does test the 400GB model, that they re-run the tests of the comparison SSDs so that the competitors are overprovisioned to 400GB usable capacity (from 512GB or whatever nominal capacity). That is the only reasonable way to compare, since anyone who wants high sustained performance and is willing to try a drive with only 400GB to achieve it would obviously be willing to overprovision, for example, a 512GB Samsung 850 Pro to only 400GB usable to achieve higher sustained performance.
Kristian Vättö - Thursday, April 2, 2015 - link
That is something that I've had on my mind for a while now and I even have a way to do it now (the Storage Bench traces are a bit tricky since they are run on a raw drive, but thankfully I found an hdparm command for limiting the far LBA count). The only issue is time because it takes roughly two days to test one drive through the 2015 suite, so I may be include a drive or two as comparison points but I definitely can't test all drives with added OP.Kristian Vättö - Thursday, April 2, 2015 - link
Not far LBA count, but raw LBA count, obviously :)Stahn Aileron - Friday, April 3, 2015 - link
Honestly, I'd rather have AnandTech test drives and components as-is ("stock" from the manufacturer) and publish those results rather than spend time doing tests on non-standard, customized configurations. Let the customers do that if they truly need that type of set-up or leave it to integrators/specialists.As far as I know, most customers of a product just want to use it immediately right of the box, no mucking with special settings. Most products are advertised that way as well.
Really, just test the product(s) as advertised/intended by the manufacturer first and foremost to see if it matches their claims and properly serves the target userbase. Specialty cases should only be done if that is actively advertised as a feature, there is truly high interest, something makes you curious, and/or you have the time.
jwilliams4200 - Friday, April 3, 2015 - link
If this were a review site for the totally clueless, then you might have a point. But anandtech has always catered to enthusiasts and those who either already know a lot about how computer equipment works, or who want to learn.The target audience for this site would certainly consider something as simple as overprovisioning an SSD if it could significantly increase performance and/or achieve similar performance at lower cost relative to another product. So it makes sense to test SSDs configured for similar capacity or performance rather than just "stock" configuration. Anyone can take an SSD and run a few benchmarks. It takes a site as good as anandtech to go more in-depth and consider how SSDs are actually likely to be used and then present useful tests to its readers.
Kristian Vättö - Thursday, April 2, 2015 - link
That is correct. I always ask for all capacities, but in this case Intel decided to sample all media with only 1.2TB samples. I've asked for a 400GB, though, and will review it as soon as I get it.Mr Alpha - Thursday, April 2, 2015 - link
Has anyone managed to find this mythological list of compatible motherboards?Kristian Vättö - Thursday, April 2, 2015 - link
I just asked Intel and will provide a link as soon as I get one. Looks like it's not up yet as they didn't have an answer right away.tstones - Thursday, April 2, 2015 - link
Older chipsets like z77 and z87 will support NVMe?