Page 10: HD Tach and Timed Data Transfer Tests

HD Tach V2.70 Source: TCD Labs Inc.

HD Tach is a physical performance hard drive test for Windows 95/98/ME and Windows NT/2000. In Windows 9X/ME it uses a special kernel mode VXD to get maximum accuracy by bypassing the file system.
A similar mechanism is used in Windows NT/2000/XP. HD Tach reads from areas all over the hard drive and reports an average speed.
It also logs the read speeds to a text file that you can load into a spreadsheet and graph to visually read the results of the test.
Hard Drive Tach 2.70 - Benchmark Results
Physical Drive Size Access Time Read Bust Speed Read Speed Max Read Speed Min Read Spin Avg CPU Ultilization
No RAID, Single HDD 120GB 13.8ms 83.0 45.6 22.7 37.6 26.8%
Hardware RAID 0 240GB 13.5ms 83.1 71.4 24.9 40.4 23.6%
(Software RAID 0)* - - - - - - -
Hardware RAID 1 120GB 13.8ms 82.9 32.5 13.2 20.4 19.7%
Units: MB/s MB/s MB/s MB/s
*Unfortunately, as it can only 'see' hardware RAID arrays and not Windows created software RAID arrays, we were unable to use HD Tach to benchmark our software stripe. The traditional drive speed-testing benchmark show us that Hardware RAID 0 offers the fastest data read times, while Hardware RAID 1 offers the best CPU utilization.
Timed Data Transfers Source: n/a

While not the most scientific of methods, this certainly illustrates real world file copying performance. We hauled the 'I386' directory from a Windows Server CD onto our system drive, then proceeded to copy this 488 MB chunk of many, many files to and from each of our drive configurations several times, averaging the scores. Timing the transfers revealed some interesting things.
Time Data Transfers
Physical Drive Size Upstream Transfer Downstream Transfer
No RAID, Single HDD 120GB 43s 39s
Hardware RAID 0 240GB 33s 34s
Software RAID 0 240GB 33s 35s
Hardware RAID 1 120GB 49s 51s
Units: seconds seconds
Again the performance advantage RAID 0 gives us is clear.
RAID Test conclusions:
Both hardware and software RAID 0 should offer a significant increase in overall hard disk performance to any system. The tradeoff between the two is in expense (if your system does not have a hardware RAID controller built in) vs. the slightly increased load on the CPU that software RAID imposes.
We took screen shots of the task manager CPU usage graph while we were running our file copying tests on both RAID 0 configurations:
Hardware stripe: Software stripe:

Not a huge difference, but it's there. Overall though, either implementation will serve you well.