Why doesn't AT do 'real world' endurance testing for SSDs?

[pitz]

SSDs do a lot of behind-the-scenes stuff, even at idle, which isn't captured with the 'extreme' sort of accelerated SSD testing that is done by other websites. Additionally, there's the whole question of what really constitutes a 'typical' SSD workload.

So I really don't see the current benchmarks as presented by Anandtech to be an issue. As for writes, a year and a half into using an Intel 240gb 330, I have roughly 7.5Tb of NAND writes, and 6Tb of host writes and a little over 5.5gb of host reads (yes, more writes than reads!!! due to my heavy use of the hibernate feature in Win7!). Various websites have shown that the Intel 330's in smaller sizes can easily do 100Tb+. So not a worry in the world that I will experience NAND failure before the disk is damaged through other causes or simply becomes surplus to my needs.

 

[Elixer]

In the TechReport past article, I noticed they do checksums, and they had failures of that, which would have gone undetected by most people.

This still leads me to believe that there is some kind of firmware issue (from multiple OEMs), that can't detect certain kinds of NAND errors, something that ECC can't detect.

Right now, most people getting errors, either find out by either windows throwing up errors / warnings in the event logs, or, the system starts to become flaky, so, people just reinstall the OS.