Benchmark - SYSmark 2014 SE
80GB Windows 10 disk image
SYSmark is used by PC manufacturers to benchmark their PCs against a known and generational updated reference score, using actual applications such as Microsoft Office and Adobe Creative system instead of recording disk traces. THis is meant to present a real world case for performance, real apps used everyday by users run on hardware whose settings and behavior is tightly controlled by the sysmark program itself. OEMs also use its sister app, mobilemark to gauge battery life for their marketing materials. THere are some vendors who op posse sysmark because it includes some tests that they think are not daily route such as OCR but those vendors are the ones who have a performance deficient in some of these tests hence protecting their reputation.
The quirk with the latest version of Sysmark is the reference score already uses a recent 6th Skylake i3-6100 CPU and a modern 256GB Samsung SSD, so runs comparing against this reference score of 1000 will demonstrate slight differences or lack thereof between SSDs. Additionally we are comparing an older quad core i5 versus a newer dual core i3 with the former using an addin GTX 950 GPU and the later Intel onboard HD Graphics.
Despite this 'quirk', the reference score using a modern SSD does serve us a comparison purpose going forward and especially if NVME or other future storage is tested. This method will give us a real world comparison using real world apps and tasks.
From our sysmark tests we can see the influence of the differences in the host system for some of the tests ie Office productivity where higher clock speed matters, media creation where more CPU cores matters and secondary disk throughput. What we want to focus on is responsiveness and Overall score. All the tested SSDs provide good responsiveness, answering the question 'what is the difference between a Samsung and crucial SSD' from a actual usability point of view.
The ultimate test would be to benchmark using the same hardware as the reference score therefore eliminating differences in the CPU platform when looking at the affect of a particular SSD. However, SYSMark was meant as a total system test not a storage subsystem test.
SYSmark will fail to work on a disk image that already contains user applications such as MS Office and Adobe Creative Cloud. SYSMark includes frozen versions of these software which are encrypted and cannot be patched. This is to ensure 100% repeatability. In theory I could have still setup the clean 80GB disk image and threw in another 290GB of data to fill it up but that would mean several things, I introduce a third scenario and the Intel drive could not be tested.
