CORE International, an AIWA Co.

CORETEST.TXT
Copyright 1994 CORE International. All Rights Reserved.

                 ͻ
                  COREtest(c) - Disk Performance Test Program  
                                    Rev. 3.03                  
                 ͼ

WELCOME TO COREtest, a fully functional disk performance testing tool.

COREtest 3.03 is the latest version of this widely used performance test
program.  With over one million copies in use throughout the world,
COREtest now offers enhanced performance testing for today's current disk
technologies.

This document contains important information not included in any other
documentation.

This file contains the following sections:
         License Information
         Limited Warranty and Disclaimer of Warranty
	 Acknowledgement
         Introduction
	 Command Line Parameters
	 Running COREtest
         About COREtest
	 Caching Information
	 Ranking Screen
	 Definitions
         Special Notes
         Questions & Answers

LICENSE INFORMATION
CORE International extends an end user license for use of COREtest.
Copies may be made; however, commercial distribution, from other than
CORE International and/or its agents, distributors, and resellers is
strictly forbidden.

LIMITED WARRANTY AND DISCLAIMER OF WARRANTY

THIS SOFTWARE AND ACCOMPANYING WRITTEN MATERIALS (INCLUDING
INSTRUCTIONS FOR USE) ARE PROVIDES "AS IS" WITHOUT WARRANTY OF ANY KIND.
FURTHER, CORE International, INC.(CORE) DOES NOT WARRANT, GUARANTEE, OR
MAKE ANY REPRESENTATIONS REGARDING THE USE, OR THE RESULTS OF USE, OF THE
SOFTWARE OR WRITTEN MATERIALS IN TERMS OF CORRECTNESS, ACCURACY,
RELIABILITY, CURRENTNESS, OR OTHERWISE.  THE ENTIRE RISK AS TO THE RESULTS
AND PERFORMANCE OF THE SOFTWARE IS ASSUMED BY YOU.  IF THE SOFTWARE OR
WRITTEN MATERIALS ARE DEFECTIVE YOU, AND NOT CORE OR ITS DEALERS,
DISTRIBUTORS, AGENTS, OR EMPLOYEES, ASSUME THE ENTIRE COST OF ALL NECESSARY
SERVICING, REPAIR, OR CORRECTION.

THE ABOVE IS THE ONLY WARRANTY OF ANY KIND, EITHER EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, THAT IS MADE BY CORE,
ON THIS CORE PRODUCT.  NO ORAL OR WRITTEN INFORMATION OR ADVICE GIVEN BY
CORE, ITS DEALERS, DISTRIBUTORS, AGENTS OR EMPLOYEES SHALL CREATE A
WARRANTY OR IN ANY WAY INCREASE THE SCOPE OF THIS WARRANTY AND YOU MAY NOT
RELY ON ANY SUCH INFORMATION OR ADVICE.  YOU MAY HAVE OTHER RIGHTS WHICH
VARY FROM STATE TO STATE.

NEITHER CORE NOR ANYONE ELSE WHO HAS BEEN INVOLVED IN THE
CREATION, PRODUCTION OR DELIVERY OF THIS PRODUCT SHALL BE LIABLE FOR ANY
DIRECT, INDIRECT, CONSEQUENTIAL OR INCIDENTAL DAMAGES (INCLUDING DAMAGES
FOR LOSS OF BUSINESS PROFITS, BUSINESS INTERRUPTION, LOSS OF BUSINESS
INFORMATION, AND THE LIKE) ARISING OUT OF THE USE OR INABILITY TO USE SUCH
PRODUCT EVEN IF CORE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.

This Limited Warranty shall be governed and construed in
accordance with the laws of the State of Florida.


ACKNOWLEDGEMENT

BY USING THIS VERSION OF COREtest YOU ACKNOWLEDGE THAT YOU HAVE READ THIS
LIMITED WARRANTY, UNDERSTAND IT, AND AGREE TO BE BOUND BY ITS' TERMS AND
CONDITIONS.  YOU ALSO AGREE THAT THE LIMITED WARRANTY IS THE COMPLETE AND
EXCLUSIVE STATEMENT OF AGREEMENT BETWEEN THE PARTIES AND SUPERSEDE ALL
PROPOSALS OR PRIOR AGREEMENTS, ORAL OR WRITTEN, AND ANY OTHER
COMMUNICATIONS BETWEEN THE PARTIES RELATING TO THE SUBJECT MATTER OF THE
THE LIMITED WARRANTY.

 
INTRODUCTION
    COREtest is a simple disk test tool which allows a user to
    quickly evaluate the performance of disk drives and attached controllers.
    COREtest is not meant to be exhaustive in performance testing yet its
    functional value should be weighed against the higher cost programs which
    may require additional hardware.

    COREtest is a DOS based program; it should NOT be run from inside a virtual
    DOS session of MS Windows or OS/2.

RUNNING CORETEST
    Run CORETEST from the DOS prompt.  Run CORETEST /? to see a listing of
    available command line options.
       For example:  C:>CORETEST or C:>CORETEST /?

COMMAND LINE PARAMETERS

Syntax of the command line parameters for COREtest is as follows:

 CORETEST [?] [HELP] [/B:xx] [/C:xxx] [/D:n[="xxx"]] [/P:xx] [/T:xxx] [/?]

 Format       Description
        
 /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).

ABOUT CORETEST
    This program will measure the performance of one or two hard disk drives.
    The evaluation is made using five tests: Read Buffered, Read Sequential,
    Read Random, 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 three data transfer tests measure the performance of the computer,
    controller, and drive functioning as a combined unit.  The block size is
    the same for each test and is calculated automatically 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 "Read Buffered" test reads the same
    cylinder continuously.  For drives with caches, this test should indicate
    the absolute maximum transfer or "burst" rate possible, since the data
    will always be transferred from the cache.  The "Read Sequential" and
    "Read Random" tests will read a sequential and random set of sectors,
    which are the two types of I/O typically performed in real-world
    applications.

   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."  This is done by moving the heads across
   one-third of the cylinders and thus 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 of 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 1.2 while an IBM 20MB AT will
   yield an index of approximately 2.5. The higher the index, the more
   powerful the computer system. Generally, a computer with a performance
   index of two is twice as fast as a computer with an index of one.

   Caching Information
   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.

   Ranking screen
   The final screen displays where your drive(s) rank compared to others
   in the industry.

DEFINITIONS
  
   The following definitions relate to the operation of the COREtest
   program. Some definitions are strictly connoted to ensure exact
   understanding of the intended environment.

   Average Access Times: The combination of the seek time of the head/arm
         across the data platter plus the track-to-track latency;
         normally listed in milliseconds or ms.

   Data Transfer Time: The amount of bytes transferred from the disk
         drive to the cpu; normally this metric is listed in MegaBytes per
         Second or MB/sec.

   Media Transfer Time: Normally listed in Megabits per Second or Mb/sec,
         this metric is the raw number of data bits, including overhead
         bits, transferred from the actual disk drive platter.  This number
         becomes the Data Transfer Time, after all overhead bits are
         removed.
	 
   Seek Times: Calculated as an average or half of a
         full stroke across the platter by the heads.
     
   SCSI: Small Computer System Interface

   IDE: Integrated Drive Electronics

   ATA: ANSI standard for IDE which means AT bus Attachment.

   ESDI: Enhanced System Disk Interface

   MFM: Multi-Frequency Modulation

   RLL: Run Length Limited

   Buffer: Normally, a small portion of ram which is used for temporary storage
           of data from a disk drive and/or controller.

   Cache: A buffer that contains data that is maintained in a valid state
          in anticipation of a future need for the same data.

   Read Buffered: Read data from the drives internal cache buffer without
         the drive having to retrieve data from its disk platters.  The
         first read would cause the drive to fill its internal cache buffer
         with data from the disk platters.  Subsequent reads of the same
         disk location will be satisfied from the drives cache buffer.

   Read Sequential: Each read to the drive requests the next sequential block
         of data from the drive.

   Read Random: Read requests are for blocks of data from random locations
         on the disk drive.

   Random Seek: Seeking to random locations on a disk drive.

   Track-to-Track Seek: Seeking from the current track location to the next
         adjacent track.

SPECIAL NOTES


"N/A" for Avg. Seek Time and Performance Index

    Some controllers, especially SCSI, don't generate seek requests to
    drives when CORETEST asks it to do so. The controller's BIOS will
    simply return a "GOOD" status to CORETEST, indicating that the seek
    has completed. On such controllers, CORETEST cannot determine the
    attached devices' seek performance, and will display "N/A," or
    "Not Available," for the seek time. CORETEST will also display "N/A"
    for the Performance Index, since an accurate Performance Index rating
    requires a valid seek time.


Cache Segment Sizes and the Read Buffer Test

    Some expermimentation with the /B parameter (block size) may be required
    to produce the fastest Read Buffered test result for a given
    configuration. Caching controllers and drives normally have
    pre-determined cache segment sizes. A cache segment is a single unit
    of cache used to hold cached data from one contiguous area on the drive.
    If the block size exceeds this segment size, the Read Buffered test will
    force the drive to read data from disk every I/O, producing a skewed
    Read Buffered result (remember that the Read Buffered test is supposed to
    measure the raw transfer speed of the disk interface, and assumes that
    the data being read from the drive will be in cache). Typical cache
    segment sizes are 32kb, 64kb, and above. If the Read Buffered test
    results seems low (within 20% of Read Sequential test result), run
    CORETEST with a smaller block size.

QUESTION &  ANSWERS

  Q. Are the results of CORETEST 3.03 comparable to previous versions?
  A. Yes, CORETEST 3.03 will produce the same Seek, Read Buffered, and
     Performance Index results as previous versions. However, previous
     versions of CORETEST (2.92 and below) didn't have the Read Sequential
     and Read Random tests, so these results are unique to CORETEST 3.03.

  Q. The timer that indicates seconds elapsed runs "slow." Does this produce
     erroneous results?
  A. No. The timer indicates time that has elapsed for I/O operations, not
     real time. CORETEST times the actual I/O time, turning the timer off
     while it updates the display. Depending on the speed of your video
     card/BIOS, the test may take longer than the specified test time.
     (default test time is 4 seconds, use /T to override)

  Q. The results of the Read Sequential test are faster than the Read
     Buffered test. Isn't the Read Buffered test supposed to test the maximum
     burst rate of my configuration?
  A. If the block size for the test exceeds the controller's/drive's cache
     segment size, the Read Buffered test will produce much slower results.
     See the "Special Notes" section for more details.

  Q. I have tested two drives which have near Read Buffered and Seek time
     results, but one drive has a much faster Read Random result. But
     CORETEST generates the same performance index for both drives. What
     gives?
  A. Versions 2.92 and below of CORETEST didn't have the Read Sequential and
     Read Random tests. To remain backward compatible, CORETEST 3.03
     excludes these two tests when calculating the Performance Index. The
     result is a Performance Index that is aligned with previous versions
     of CORETEST.

  Q. I have tested the same drives listed in CORETEST's Industry Comparison
     chart, but I get results that are much faster or slower. Why is this?
  A. Most of the contemporary drives listed in the comparison chart are SCSI
     drives. All SCSI drives were tested using the Buslogic 747 EISA SCSI
     adapter. Different SCSI adapters and/or system configurations
     will produce varied results.
