Quote from: Nytebyrd on October 10, 2020, 07:54:08
The way the test ssds gives a very inaccurate perception of their longevity. There isn't currently a better way, but most ssds far out-live their expected write count.
They tend to test them by repeatedly writing until they fail. That's an inaccurate representation of actual useage. You're not going to continually write and rewrite 600tb of data on an ssd in a single sitting. This tends to produce excess heat which is typically what kills electronics in general. Worst case data at least tends to be useful though.
The spacing of the memory modules and controller is far greater than you'd tend to see on a typical m.2. This should allow for more stable thermals even under periods of extended writing.
As someone who uses liquid metal, I'd be far more worried about its long term application. Typically you'll see a galvanic reaction with copper and nickel. You'll see the hard metal seeming to absorb the liquid and leave behind a stain and crusty residue the first couple of applications. Which means the liquid metal can have a very short life span of a few months to a year initially. I'm a fan of how they decided to prevent it from bleeding, but nothing was said about what was done to extend its usefullness past a year, let alone 6-7 years. And few people are comfortable applying it themselves, for a good reason.
Finally, someone who actually knows what they hell they are talking about. People don't understand that MTTF means that, the mean time between failure in the most extreme cases. People also forget that the controllers on even old SSDs do a lot of magic in the background to make certain that the NAND modules age evenly. In the event that an individual module fails or goes bad/sees heavy errors, the SSD controller can mark that unit unusable and avoid writing to it so it will continue to function normally. The DDR memory will likely fail before the SSD does on the PS5 even in extreme cases.