Does anyone know how the sysspeed hdd test works?

[Glen M]

I recently did some testing of different HDD solutions for the Amiga and the IDE HDD came out on top (and by a long way) but it was put to me that this is a false reading caused by the cache on the drive. Looking at the datasheet there is 8mb of cache on the hdd tested. Now to my mind the comment is a null point as the cache is doing its job and speeding up access but nevertheless I'm wondering if there is a way to get around the it.

Is the raw read result the one I want to be comparing?

I assume the create, write, read file tests do just that and the software times how long it takes giving you the MB/s. How big is the file written? I assume its about 1MB if even.

Would there be any merit in testing time in the likes of dopus to copy a large file from drawer A to B?