We had trouble with a serial-ATA raid controller in a windows server. Write performance was exceptionally poor, so we contacted the manufacturer.

We now received the following second followup:

Sorry for late reply.
We have test the AEC-6897 R5 performance.
The test result is simular your description.
Sorry for your inconvenience.

Well, thank you for acknowledging that write performance in a 7+1 SATA RAID5 is just 2.5 MB/s, after we had been stating that switching to RAID 0+1 is not acceptable for us. Now, do you have a timeframe when a working bios for your controller will be available?

However, when you produce a RAID 5 controller for 8 drives, don’t you think the 7+1 setup is, well, an obvious choice, that should have been tested for write speed before? :-(