HPT372 2.344 slow drive issue
I have a Soyo P4I875P motherboard that has a HPT372 chip with bios 2.344 The hard drive on the controller is a Maxtor DiamondMax Plus 9 200GB. I have tested the drive on ICH5 and it does around 48-52MB/s but on HPT372 it does ~ 9MB/s. This still shows up on 2.6.0-test4-bk2 and did on 2.4.x but I haven't tested on 2.4 for several weeks.
hdparm -Tt /dev/hdi/dev/hdi:
Timing buffer-cache reads: 4128 MB in 2.00 seconds = 2063.28 MB/sec
Timing buffered disk reads: 156 MB in 3.00 seconds = 51.99 MB/sec
bonnie++
calc,2G,9054,32,8713,3,7028,1,19993,59,27927,2,144.0,0,16,3190,99,+++++,+++,+++++,+++,2980,91,+++++,+++,7685,85
Thanks,
Chris Cheney
Quoted 5 lines · Chris Cheney
I have a Soyo P4I875P motherboard that has a HPT372 chip with bios 2.344 The hard drive on the controller is a Maxtor DiamondMax Plus 9 200GB. I have tested the drive on ICH5 and it does around 48-52MB/s but on HPT372 it does ~ 9MB/s. This still shows up on 2.6.0-test4-bk2 and did on 2.4.x but I haven't tested on 2.4 for several weeks.
I reported a similar issue some weeks ago (Abit K7D-RAID, integrated HPT372 IDE controller), I played with 2.4.22-pre10 for a while. With the opensource HPT driver, I reached good speed sometimes (two ATA drives were connected to the HPT, one drive on each channel), but not reliable - after a reboot, the md synchronisation between the two drives went down to 1MB/s... The HPT driver uses the drives as SCSI drives, so it's a little painful to switch between the kernel and the HPT driver (initrd, mdadm, and so on...). Finally I gave up, and disabled the HPT controller in the system BIOS. Maybe the problem is a simple IDE autotune problem, I don't know.
Geller Sandor <[email protected]>