testing disk drive speeds
the journal of Michael Werneburg
twenty-eight years and a million words
Kokubunji, 2024.10.17
Today I added a second M.2 drive to my aging PC to see if I could coax more performance out of the thing while editing a lot of photos at once. I then tested the speed versus a number of other drives on my system. These are the results in order of ascending performance.
I use this as a "go disk", that is if I wake to a fire I will grab this out of its half-way-enclosed external holder and run out the door. The performance at 70 Mbps is downright terrible due to the external USB connection and old SATA II interface. It is only good for backups and even then only just.
The performance here is around 2.5x better than the external drive, and meets my expectation. These old workhorse drives are great for reliably storing a lot of data, but the are very slow for internal drives. I like them because they fail differently to solid state drives and are effectively a different type of medium.
Once again the throughput on this drive triples over the previous class of device. Unlike the previous models, this is not a spinning platter disk but solid state: it's all electronics. The throughput is at the SATA II limit, which is good for this kind of device.
This is the drive I just added. Theoretically, on a new PC I could get 32Gbps from this system. But alas, my old motherboard doesn't support this class of drive directly and I had to mount it on a PCIE card. Still, the performance is 6x what my next-fastest drive offers.
This is my older M.2 drive, which is a lowly SATA II / B-Key unit. The performance is quite good, slightly exceeding the new drive.
RAM drive (16GB of memory)
This is a bit of a cheat. I made a pseudo-volume from some of the RAM on my system. Nothing written to this "drive" will survive a reboot. It's interesting that the speed is greater than that of the M.2 drives, but not by much.