Is it possible for kdiskmark to give inaccurate readings? I tested my array with your tool and it reported about 200 MB/s with a 6GB test file but using ddrescue to write zeros to the same FS on the same drive is averaging 600 MB/s the load on the system is minimal the drive is under the same external load (seeding linux ISOs at a few hundred kb/s) in both tests
the numbers are repeatable.
im sure none of the logic is to blame here but is it possible that there is a bottleneck in FIO? FS is btrfs the results are accurate when i do the same test with my internal NVME drive
Is it possible for kdiskmark to give inaccurate readings? I tested my array with your tool and it reported about 200 MB/s with a 6GB test file but using ddrescue to write zeros to the same FS on the same drive is averaging 600 MB/s the load on the system is minimal the drive is under the same external load (seeding linux ISOs at a few hundred kb/s) in both tests
the numbers are repeatable.
im sure none of the logic is to blame here but is it possible that there is a bottleneck in FIO? FS is btrfs the results are accurate when i do the same test with my internal NVME drive