Contents of the CORETEST.DOC file
COREtest - Disk Performance Test Program v2.91
This program will measure the performance of one or two hard disk
drives. The evaluation is made using three tests: data transfer,
random seek, and track-to-track seek. The program will not destroy any
data on the disk and may be run as many times as desired.
The data transfer test measures the performance of the computer, disk
controller, and drive functioning as a combined unit. This is
accomplished by reading repeatedly a specific block of data and timing
each transfer. The resulting effective transfer rate is produced from
the combined relationship of the computer bus structure (8, 16, or 32
bits), the transfer rate of the controller with the drive, the
interleave factor for which the drive was formatted, and the size of the
data block transferred.
The block size is automatically calculated or may be explicitly set from
1 to 64K in 1K increments. If automatically calculated, the block size
will be set to 64K or the amount of data stored on one cylinder,
whichever is less. The transfer rate is displayed in units of kilobytes
per second (KB/sec). This indicates how much data will be moved from
the disk to memory in one second.
The next two tests measure the time required to position the read/write
heads to a specific location where data is stored. The first of these
tests will measure what is referred to as the "average access time" by
moving the heads across one-third of the cylinders, simulating true
random access. The final test will measure the time to move the heads
from one track to an adjacent track, measuring the "track-to-track
access time." Both these results are reported in milliseconds (ms =
thousandths of a second). The smaller the access time, the faster the
data may be located and transferred to the computer.
After the tests are complete, a "performance index" is calculated for
the tested system. This number measures the overall performance of a
given computer, the disk controller, and its hard disk. It may be used
to evaluate the performance of different computer systems on a
comparable scale. For example, a standard IBM 10MB XT system will
typically report an index of one while an IBM 20MB AT will yield an
index of approximately 2.4. The higher the index, the more powerful the
computer system. It can generally be said that a computer with a
performance index of two is twice as fast as a computer with an index of
If COREtest detects that a cache program is installed, it will display
a message indicating same. The presence of a cache program can mis-
represent the actual drive performance. When comparing the performance
of disk drives, you should do so without any cache programs installed.
This will give a true representation of the actual drive performance.
The final screen displays where your drive(s) rank compared to others
in the industry.
Various options are available to change the test parameters to better
compare performance across a wide variety of applications. These
options must be placed on the command line when program execution is
started. The format of the command line and the option descriptions are
CORETEST [?] [HELP] [/B:xx] [/C:xxx] [/D:n[="xxx"]] [/P:xx] [/T:xxx]
/B:xx Use a block size of xxK. The block size may be up to
64K. Default is the largest transfer possible without
moving the disk heads.
/C:xxx Set maximum seek stroke to xxx. This is used to set drives to
equal cylinder comparisons. If not specified, then the full
cylinder range of the drive will be used (default).
/D:n[="xxx"] Test disk number n (n = 0-7). This parameter may appear
up to two times. If a string is enclosed in quotes, then this
string will be used to identify the drive on the display. If
omitted, up to the first two drives will be tested.
/P:xx Pause for xx seconds when test complete (for batch files).
/T:xxx Perform transfer test for xxx seconds (default = 4 seconds).