Why go through all this work just to wreck SSDs?
I'd say for the perfect mixture of curiousity and insanity.
I spent hours on Google searching for information on NAND wear, write amplification, over provisioning, wear leveling, TRIM, garbage collection, block sizes, die thickness and P/E cycles. But when it came to the questions about endurance, reliability and the expected life-time of SSD, it was mostly personal opinions.
I really liked the user initiative over at Xtremesystems.org. They started testing the write endurance of older SSDs by writing data to them until they failed. This is my way of picking up where they left off and add a litte something to the mix.
My initial intention was to test two specific SSDs. And that is what got this whole thing started.
- Intel SSD 520-Series with 25nm MLC NAND is my most important unit to test. It's often used in servers.
- SanDisk Ultra Plus is a good entry level SSD with 19nm MLC NAND. It's good for caching static data.
Next I would like to test these SSDs.
- Samsung 840 EVO (19nm) TLC NAND vs. Samsung 840 (21nm) TLC NAND
- Intel SSD 530-Series - a worthy replacement for the old 520?
- Seagate SSD 600
Please feel free to make a tiny donation, every cent is appreciated and will be spent on SSDs.
Which SSD would you like to see tested next?
Christian Löfstedt Google+
Geeks and nerds alike are most likely to be the ones to consider making donations for upcoming SSD tests. Feel free to prove me wrong :-)
Is this really normal?
No. The tests preformed here must not be confused with normal use. They are designed to stress SSDs and simulate a really busy environment.
Normal workstation use would be more like 10-20GB written daily.
Why Test Smaller SSDs?
They use fewer NAND circuits and will wear out faster. Once we know the write endurance of these units, it'll be easy to estimate the equivalent numbers for the larger units.