Promise FastTrack TX2000
6. SCSI Mechanic Tests
Review Pages
Promise FastTrack TX2000 - Page 6
SCSI Mechanic Linear Reading/Writing Tests
We performed the SCSI Mechanic tests as described in our test methology guide. The test results are illustrated in the following graph. Note that all numbers are Kb/s:
The above graph shows how the Promise TX2000 performs with two disk drives under various modes (Stip, Mirror) and with different Strip size. At the "Same Sector" test, there is difference from the normal ATA100 connection interface. Under RAID 0,1 ATA133 connection the Maxtor drives reached almost 90Mb/s. There are slight differences between the various "Strip" sizes. The best result came from Raid0 - 512Kb Strip size.
At the "Sequential" test, the best result comes from "Strip 8kb" with 53364 Kb/s. The second best performance comes from "Strip 128Kb" with 48788Kb/s and the third best from "Strip 64 Kb" with 47888Kb/s. As it seems the larger Strip sizes of 512Kb and 1024Kb decrease performance down to 34 and 29Mb respectively. The RAID-1 ATA133 format has the same performance as the normal ATA100 interface.
At the "ButterFly" the highest performance comes from "Strip 64Kb" with 15290Kb/s. The second best performance comes from "Strip 8Kb" and the third best from "Strip 128Kb". Again the lowest performance comes from "Strip 1024Kb". Last at the "Random" test, the highest performance comes from "Strip 1024Kb" with 9870Kb/s and the second best from "Strip 512Kb" with 9790Kb/s. The "Strip 8Kb" comes last with 8982Kb/s, which shows that larger Strip blocks does affect "Random" reading.
At the linear writing tests, things change dramatically. The "Strip 1024Kb" block size does affect performance and it has the highest writing scores. The "Strip 512Kb" block size setting has the second best result. The RAID-1 has the lowest "ButterFly" and "Random" I/O tests.
Review Pages