Torture Testing8610 Tipping Point
You didn’t think I would let them get off that easily, did you? There are other users (many of which wrote in asking for HDDErase, prompting us to post the older version to the site) who managed to get their X25’s past the tipping point, so it is reasonable to assume a user may be flashing their X25 while it is in that state. I (reluctantly) rolled my X25 back to 8610, wiped it, and ran my tweaked tool on it for a few hours (the same one I passed to Intel so they could replicate my initial discoveries in-house). With my drive well past the point of no return, I flashed it back to 8820.
The 8820 firmware worked aggressively to defragment on the fly and stuck surprisingly close to 80 MB/sec while doing so.
The new firmware continued to impress the heck out of me.
8820 Tipping Point Attempt
After the X25 shrugged off its fragmentation from the last test, I turned right around and hit it with my tool again. My goal was to see how the newer firmware would fare against the very same torture test.
Fail. For me at least.
While my test leans towards smaller writes, it still contains a mix of larger sized writes. It turns out the 8820 firmware was so aggressive that any fragmentation caused by the smaller combined writes was immediately cleaned up by a subsequent larger write covering that same area.