PCMark Vantage - Drives with Data Testing
I would like to thank my friends for informing me that other review websites have 'stepped-up' their SSD review process (and in some cases, copied ours - they said it, not me). My answer to increased competition is with increased drives to deliver a more accurate review.
As you know, I will not publish results from tests not validated or publically available. For the last month I've spent countless hours building, modifying and validating a test that will accurately display real world performance of an SSD as it is used in your PC. The final version of this test was completed five days ago and then went through an exhaustive series of tests on both HDDs and SSDs. An article will go live soon that details how you can run this same test, with the same files at your convenience. All files are publically available except for Windows 7. You should have that one anyway.
SSDs perform differently when used for a period of time and when data is already present on the drive. The purpose of the Drives with Data testing is to show how a drive performs in these 'dirty' states. SSDs also need time to recover, either with TRIM or onboard garbage collection methods.
- Drives with Data Testing - 25%, 50%, 75% Full States
- Files needed for 60 (64GB), 120 (128GB), 240 (256GB)
- 60GB Fill - 15GB, 30GB, 45GB
- 120GB Fill - 30GB, 60GB, 90GB
- 240GB Fill - 60GB, 120GB, 160GB
- Empty but Dirty - a test run just after the fill tests and shows if a drive needs time to recover or if performance is instantly restored.
All test files must be publically available and be a collection of both compressed and uncompressed data.
HDD1 - Windows Defender
HDD2 - Gaming
HDD3 - Windows Photo Gallery
HDD4 - Vista Startup
HDD5 - Windows Movie Maker
HDD6 - Windows Media Center
HDD7 - Windows Media Player
HDD8 - Application Loading
No one said that the chart had to be easy to read. I'm working on a better way to display the data collected.
We are going to avoid the temptation to start at the top, and instead go from the bottom up. The Corsair Performance 3 is also a large reason why we brought this method of testing to light to begin with. The Corsair Performance 3 with the Marvell BKK2 controller revision was designed to keep a steady state of performance even when the drive was filled with data up to at least 75%. The Empty / Dirty is relatively unchanged, but since performance wasn't lost in the tests with data we weren't surprised.
Moving up on the chart to the OCZ Technology Vertex 3, we see the Vertex 3 lose a lot of performance between the results on the previous page and the 25% full state on this page. More performance is lost between the 25% and 50% state, but after 50% performance levels off. When the Vertex 3's contents are deleted performance isn't immediately recovered and the 75% results look a lot like those of the Empty / Dirty state.
Now for the drive we are previewing today, the Crucial m4. Just like with the Vertex 3, the m4 lost a lot of speed with 25% of its capacity taken. The curve kept going down with the 50% tests, but for the most part the 50% and 75% tests showed a steady state of performance. The part that is very impressive for the m4 is how quickly it recovers once the data is cleared. All tests are run back to back in rapid succession and are not given a chance to recover. Once the 25% test is completed more data is sent to the drive and the next test is started.
This has prompted a new string of tests which I'm currently validating now where a 10 minute break is given between each test for the data to 'settle'. Timing constraints will leave those results to be published in the testing methodology article that will go live soon. Testing is being run on drives dating back to the Indilinx Barefoot.
There is a very wide gap in performance between the Performance 3, Vertex 3 and m4 in the dirtiest of states; 75% full with a rapid growth of data without the drives being given the opportunity to clean themselves.