SATA SSD LABS TEST
Edward Chester takes a look at the latest SATA solid state drives to find which ones offer the best bang per buck
How we test
SATA drives might not represent the glamorous end of the SSD market, but they’re fast enough to cope with large file transfers, while offering quick Windows boots and fast game loading times. Some games still take a while to load, but the bottleneck is seldom storage, unlike in the days of hard drives.
What’s more, there are a couple of exciting new additions to the world of SATA SSDs. Both Samsung and Micron have released brand-new models, and while performance is still limited by the SATA interface, capacities are up and prices are set to trickle downwards. While NVMe M.2 drives can be much quicker in some tests, they also tend to be around 40 per cent more expensive. Unless you have a specific need for that speed, it’s seldom worth it.
You can also get SATA SSDs in an M.2 form factor, so you can still get a value option without compromising your tiny build. For this test, though, we’ve stuck to traditional 2.5in drives. Testing was carried out with an Asus Z-170A motherboard, an Intel Core i5-6600K, 16GB of Corsair Vengeance LPX 2666MHz DDR4 RAM and a Sapphire Vapor-X R9 290 graphics card.
We start with two stalwarts of synthetic SSD benchmarking, CrystalDiskMark and AS SSD. These tests are run at default settings, other than changing the workload size to 5/4GB, to give the drives a bit more of a sustained run. These benchmarks give a solid indication of peak sequential read and write speeds, as well as random read and write speeds at several queue depths. Each test is run three times and the average is recorded.
Next up is PCMark 8’s secondary storage test, which runs a series of real app storage traces to gauge performance in difference application scenarios. The traces include loading games, running both a heavy and light workload in Photoshop, and running lighter traces in Word and Excel.
Our final test is IOmeter, which puts each drive under a sustained random read and write workload. We can then check that the drive’s claimed random IOPS figures hold up. We used to also test boot speed using Bootracer, but with these drives all being limited to SATA, there’s now no practical difference in Windows load times.