Performance Over Time and TRIM

We verified the new Vector 150 responded corectly to TRIM requests, and did so within a reasonable turnaround, so nothing to report there. In order to test long-term performance scenarios where the OS is not sending TRIM commands to cover large blocks of data being randomly accessed in-place (such as VM HDD image files), I hit the Vector 150 with random accesses and with no TRIM issued. I then ran a couple of HDTach passes. The way this works is the first pass will show a relative slow-down due to the drive pushing though the fragmented areas of flash. The second pass should then (in theory) show the speeds returning to rated specs across the flash area. Here goes:

First pass:

Second pass:

The Vector 150 performed as expected here. No surprises. We are looking more for a nice flat line on the second pass moreso than ultimate speed attained. Some SSDs don't agree with the write method HDTach uses, and as a result, actually slow down a bit on the second pass. This does not reflect real world performance of the SSD.

« PreviousNext »