Sign in with
Sign up | Sign in

Testing and Performance Part III

Seagate: Can There Be Only One?
By
Brought to you by What's this

Now let’s look at random operations:


Now we start to see the results we’d been expecting all along. The only clear laggard in random reads is the WD Green. The Seagate Barracuda and WD Black stay neck and neck through most of the test, with the Barracuda only pulling into the lead with the 256 KB to 1MB tests.

With random writes, we have a drastic change. The WD Black blows everything else away at 512 B, 1 K, and 2K. With 4K, the WD Blue suddenly leaps to the fore and stays there all the way out to 1 MB. Curiously, though, this is the one result where the WD Black and Blue models show undeniably superiority. Why don’t these numbers match our standalone 4K random write numbers from earlier? That’s a good question, and it reveals some of the ambiguity inherent in benchmarking, even within the same application. There are clearly some workload differences between our script’s routine and those found in the pure 100% settings used in our standalone runs. Given a choice between the two data sets, though, we’d lean toward the script’s as being more thorough and informative.

PCMark 7

Many readers will be most interested in the PCMark 7 results as they’re the only ones that integrate real-world applications into the testing process. However, we want to begin here by saying that PCMark 7 can report both system storage (our SSD C: drive, in this case) as well as the secondary drive under test as part of the same test run. Normally, the system drive gets omitted since it’s not the subject of the benchmarking investigation, but we’re showing these results below so that you can see the amount of variance from one run to the next.

There should be no influence from the secondary drive on results for the primary system drive’s performance, so we’re left to conclude that there is at least a 1.8% margin of error (the gap from 5323 to 5423) between PCMark 7 runs for overall score. In Gaming, we see this gap spread to 2.8%; in Starting Applications, it reaches 6.4%.

Now let’s look at the test results for the hard drives:

With PCMark 7’s error margin in mind, the 10.9% overall score spread we see between the highest (WD Blue) and lowest (WD Green) seems statistically significant. Seagate clocks in just a pinch behind with an 8.4% advantage over the Green. For the sake of this Overall Score graph, we raised Y-axis minimum value in order to accentuate the differences between results.

Moving into specific storage test results, the confusing aspect of this data is how it seems to disagree with Iometer. For example, we know that Iometer reports Seagate excelling at sequential operations, but PCMark 7 shows Seagate trailing in video editing and adding music, both of which are highly sequential tasks. To add a bit of perspective, though, Seagate trails the front-placed WD Blue in Adding Music by only 0.04 MB/s. Over an hour of non-stop music importing, this might amount to 10 songs, or a few additional seconds. Perhaps more than anything else, these PCMark 7 graphs indicate just how close our four drives are in performance. The Seagate Barracuda and WD Blue may emerge as the benchmark suite’s “winners,” but in practical terms, the variances seen here are minor.