Next we switched the MSI P4N Diamond for an MSI 945G Neo which uses the Intel 945G chipset with ICH7R Southbridge. The MSI 945G Neo has four SATA ports on the ICH7R RAID controller with no other SATA ports available. However, MSI includes a VIA VT6410 IDE controller so we switched from the SATA WD Raptor that we used in our other tests to an IDE WD1200JB drive.
We ran the same tests on the ICH7R controller as we did on nForce4, using a single Seagate, a pair of drives in a striped array, a pair of drives in a mirrored array and three drives in RAID 5.
The single drive had near-identical results in HD Tach to those we saw with the nForce4 but in our real world file transfer test the Intel controller was significantly slower, and the same was true in PCMark05. As you can see from the results in the thumbnails, a striped array also runs slower on IC7R than it does on nForce4 for Intel as does a mirrored array.
With RAID 5, ICH7R closes the gap on nForce4 for Intel but thatâ€™s surely because the nVidia RAID 5 is so terribly slow.
As for setting up the RAID, we liked the Intel Matrix Storage Manager software and found it fairly intuitive. Creating a RAID array in Windows was quick and simple but we found the nested view of the array unnecessarily complicated. The Matrix Console also enables you to migrate from a single drive to a RAID array or from a two drive array to an array with more drives. When migrating to RAID 5, youâ€™d better believe the warning that the process can take up to five hours (it actually took four and half hours).
Hereâ€™s the process viewed through the Matrix Console where you can see the striped array plus a third drive.
Naturally we pulled out a SATA cable to see how the RAID 5 array handled a shock to the system. The software pops up a warning that the RAID array is damaged, and tells us that we need to switch to the Advanced View to manage the array.
We make the switch and lo, one of our drives was missing. But then we knew that and once the drive was reconnected the â€˜problemâ€™ was cured.