Imaging Performance Testing: MR8, ETBH, HBSP

[Mustang]

Understanding and Mitigating Performance Variations Across Mechanical Hard Drive Tracks for Fair Benchmarking

We have recently observed a significant difference in access speeds for files stored at different locations on the same mechanical hard drive. The reason is likely that the hard drive platters rotate at a fixed speed, while the outer tracks have a longer physical circumference. Therefore, during the same time it takes for the platter to complete one rotation, the read/write head can read more data on the outer tracks. As a result, files stored on the outer tracks of the hard drive exhibit noticeably higher sequential read/write speeds compared to those stored on the inner tracks.

This difference is particularly evident when continuously accessing large files, such as during integrity checks of large disk image files. To ensure fairness in performance comparisons (whether for backup or integrity verification), we recommend using a dedicated partition to store the backup image to be tested. Additionally, ensure that during testing, the partition contains only the image file generated by the software being evaluated (the partition can be formatted before generating the backup image). This approach helps ensure that the image file being tested is located as close as possible to the beginning of the partition.

I completely agree with this explanation. I can see backup speed slow down on a mechanical hard drive toward the end. As you know my tests were done on a RAID 0 array with 2 NVMe SSD's. I can tell you that a Macrium Reflect X backup to a mechanical hard drive takes about 15 minutes vs. 1 minute to my RASID 0 array. I haven't tested Hasleo to a mechanical hard drive recently, but past test showed very similar times between Hasleo and Macrium Reflect 7. I found that Macrium Reflect X is actually a little slower than Macrium Reflect 7 when going to a mechanical Hard drive.