═══ 1. How to Use the BenchTech Reference ═══ The BenchTech Reference provides technical information about the BenchTech for OS/2 product. Much of the information presented here can also be found in the on-line help of the BenchTech for OS/2 program. How to Use the Contents When the Contents window first appears, some topics have a plus (+) sign beside them. The plus sign indicates that additional topics are available. To expand the Contents if you are using a mouse, click on the plus sign. If you are using the keyboard, use the Up or Down Arrow key to highlight the topic and press the plus (+) key. For example, BenchTech Test Suite has a plus sign beside it. To see additional topics for that heading, click on the plus sign or highlight that topic and press the plus (+) key. To view a topic, double-click on the topic (or press the Up or Down Arrow key to highlight the topic and press the Enter key). How to Obtain Additional Information After you select a topic the information for that topic appears in a window. Highlighted words or phrases indicate that additional information is available. You will notice that certain words and phrases are highlighted in green letters, or in white letters on a black background. These are called hypertext terms. If you are using a mouse, double-click on the highlighted word. If you are using a keyboard, press the Tab key to move to the highlighted word, and then press the Enter key. Additional information appears in a window. How to Use Action Bar Choices Several choices are available for managing information presented in the BenchTech Reference. There are three pull-down menus on the action bar: the Services menu, the Options menu, and the Help menu. The actions that are selectable on the Services menu operate on the active window currently displayed on the screen. These actions include the following: Bookmark Allows you to set a placeholder so you can retrieve information of interest to you. When you place a bookmark on a topic, it is added to a list of bookmarks you have previously set. You can view the list, and you can remove one or all bookmarks from the list. If you have not set any bookmarks, the list is empty. To set a bookmark, do the following: 1. Select a topic from the Contents. 2. When that topic appears, choose the Bookmark option from the Services pull-down. 3. If you want to change the name used for the bookmark, type the new name in the field. 4. Click on the Place radio button (or press the Up or Down Arrow key to select it). 5. Click on OK (or select it and press Enter). The bookmark is then added to the bookmark list. Search Allows you to find occurrences of a word or phrase in the current topic, selected topics, or all topics. You can specify a word or phrase to be searched. You can also limit the search to a set of topics by first marking the topics in the Contents list. To search for a word or phrase in all topics, do the following: 1. Choose the Search option from the Services pull-down. 2. Type the word or words to be searched for. 3. Click on All sections (or press the Up or Down Arrow keys to select it). 4. Click on Search (or select it and press Enter) to begin the search. 5. The list of topics where the word or phrase appears is displayed. Print Allows you to print one or more topics. You can also print a set of topics by first marking the topics in the Contents list. To print the document Contents list, do the following: 1. Choose Print from the Services pull-down. 2. Click on Contents (or press the Up or Down Arrow key to select it). 3. Click on Print (or select it and press Enter). 4. The Contents list is printed on your printer. Copy Allows you to copy a topic that you are viewing to the System Clipboard or to a file that you can edit. You can copy a topic in two ways: 1. Copy copies the topic that you are viewing into the System Clipboard. 2. Copy to file copies the topic that you are viewing into a temporary file named TEXT.TMP. You can later edit that file by using any editor. You will find TEXT.TMP in the directory where your viewable document resides. To copy a topic to a file, do the following: 1. Expand the Contents list and select a topic. 2. When the topic appears, choose Copy to file from the Services pull-down. 3. The system puts the text pertaining to that topic into the temporary file named TEXT.TMP. For information on one of the other choices in the Services pull-down, highlight the choice and press the F1 key. The Options menu allows you to change the way your Contents list is displayed. To expand the Contents and show all levels for all topics, choose Expand all from the Options pull-down. You can also press the Ctrl and * keys together. For information on one of the other choices in the Options pull-down, highlight the choice and press the F1 key. The Help menu allows you to select different types of help information. You can also press the F1 key for help information about the Information Presentation Facility (IPF). ═══ 2. Copyright/Trademark information ═══ BenchTech for OS/2 Copyright (c) 1992, Synetik Systems. All Rights Reserved. This program is protected under the copyright laws of the United States and other countries. Unauthorized reproduction of this program is subject to civil and criminal penalties. This version of the BenchTech for OS/2 Reference may be freely distributed. BenchTech, Synetik, and the "racing" apple and orange symbol are trademarks of Synetik Systems. Other trademarks that may be used in this document and in the BenchTech program: OS/2 is a trademark of the IBM Corporation. Microsoft is a trademark of Microsoft Corporation. DeScribe is a trademark of DeScribe, Inc. 80386, 80387, and 80486 are trademarks of the Intel Corporation. ═══ 3. Read Me ═══ This version of the BenchTech for OS/2 Reference is a intended for distribution by Synetik Systems to convey technical information about the BenchTech for OS/2 program. You may distribute this version of the BenchTech for OS/2 Reference, and only this version, amoung co-workers, customers and friends. The version of the BenchTech for OS/2 reference that comes with the BenchTech for OS/2 program, as well as the BenchTech for OS/2 program and related program files and documentation may NOT be distributed without written permission of Synetik Systems. ═══ 4. Removing BenchTech for OS/2 ═══ The BenchTech program does not write to your system configuration files, like CONFIG.SYS or OS2.INI, so removing the program from your system is easy. The following are step by step instructions for removing BenchTech for OS/2. These instructions assume that you installed BenchTech to C:\BnchTech and that you didn't move the BenchTech folder that was placed on your desktop. Adjust accordingly if you have moved the BenchTech folder or installed BenchTech in another directory. 1. Close the BenchTech Program and the BenchTech Reference 2. Drag the BenchTech folder that is on your desktop to the shredder. 3. Open the C: drive by double clicking. 4. Page down until you find the BnchTech directory. 5. Drag the BnchTech directory (folder) to the shredder. ═══ 5. Support ═══ BenchTech for OS/2 comes with one year of toll-free technical support. Support mechanisms are explained in the BenchTech for OS/2 Users guide and the BenchTech for OS/2 Reference. ═══ 6. Ordering Information ═══ For a limited time, BenchTech for OS/2 is available for $79.95 directly from Synetik Systems. To order: Call (800) 598-1718 in the continental United States or Call (303) 241-1718 outside the continental United States or Fax (303) 243-4516 You may also request a full-color brochure or get information about: Site licenses Network licenses Custom version of the program Other Synetik Systems programs Custom programming Thank you for your interest in BenchTech for OS/2. ═══ 7. BenchTech General Concepts ═══ BenchTech is a suite of 32 bit benchmarks and associated tools designed specifically for OS/2 version 2.0 and beyond. OS/2 2.0 performance can vary greatly between one computer design and another. Even though all systems that run OS/2 are based on the Intel (r) 80386 architecture, the observed performance is also very dependant on the cache design, if the system uses a cache, the number of wait states required, the video system, the disk drive and many other factors. BenchTech consists of several benchmarks or "tests" as they are often referred to in this and other BenchTech documentation. Two of the tests, the BenchTech CPU Test and the BenchTech Composite Test, are special, designed to be used to compare different computer systems. You may see results of these tests advertised by computer manufacturers. When making comparisons between different computer systems, make sure that the guidelines for measurements using the BenchTech CPU Test and BenchTech Composite Test are followed. These guidelines are specified in the BenchTech users guide. Double click on the highlighted text below for a description of these tests. 1. BenchTech CPU Test 2. BenchTech Composite Test For an indication of how long each test may take, as well as a listing of the units that each test's results are reported in, see Test Duration and Units. The user interface for BenchTech is designed around a notebook metaphor using OS/2's notebook control. The idea is to organize the tests in such a manner that they are easy to access but not overwhelming to the user. The first page of the notebook contains a window that displays the results from all the tests. (You might think of this page as the cover of the notebook, where you've written down a summary of your important results.) Results Page The next five pages of the notebook contain checkboxes for all tests except BenchTech CPU and BenchTech Composite. Some of the tests also have optional parameters that are adjustable. CPU Page Disk Page Video Page Application Page User Page The next two pages contain controls for setting various options for the way that the BenchTech tests are performed and recorded. Options Page Logging Options Page The last page of the notebook contains the BenchTech macro feature. This feature allows the user to write scripts (macros) of BenchTech commands to automate testing. Macro Page ═══ 8. BenchTech CPU ═══ The BenchTech CPU Test is designed to examine the performance of your computer system's CPU by performing a variety of CPU related tasks, including Integer Math, Floating Point Math, and Data Movement. The results are reported in BenchTech CPU units per second. The higher the score, the better the relative performance. Note: This test will take advantage of a math co-processor such as the Intel 80387, if one is installed on your system. Such a co-processor will greatly reduce the time spent doing certain mathematical functions. However, even if you had an infinitely fast co-processor, the test still reflects the other functions which do not take advantage of this chip. For example, suppose that, on a machine without a co-processor, the test takes 100 seconds overall, 20 seconds of which are devoted to math functions which use the co-processor. Running it on a machine that has a co-processor that performs these function at 100 times or even 1000 times the speed of the system without the co-processor would yield an overall advantage of only 20%. ═══ 9. BenchTech Composite ═══ The BenchTech Composite Test looks at the performance of your computer system at a higher level than most benchmarks. The idea is to get a realistic estimate of your computer's relative performance when running real OS/2 application code. This test uses the Graphics Programming Interface built into OS/2, just like real OS/2 applications do. The BenchTech Composite Test is somewhat unique in that it uses multiple "threads" to perform its various functions. Threads are an integral part of OS/2 multi-tasking. Switching between threads can be an important component of application performance. BenchTech switches between threads using OS/2's event semaphore and message passing mechanisms. Message passing and semaphores are not a dominant part of this test, but they are an important factor. This test looks at three parts of your computer system: 1. The Hard Disk Drive 2. The Video System 3. The CPU This test uses the disk drive specified on the Disk Page. If there is insufficient space available (this test requires approximately 3.5 Megabytes) on the specified disk, the test will fail. Since this test uses the OS/2 GPI for drawing to your video display, the results of this test are dependent on: 1. Your video adapter 2. The display driver Care needs to be taken when comparing the results of this test from system to system. See "Guidelines for Advertising using the BenchTech CPU and BenchTech Composite Tests" in your users guide. Note: The first release of IBM OS/2 2.0 did not contain the 32 bit graphics engine. IBM has made it known that this feature will be added to OS/2 sometime in 1992. Care needs to be taken when comparing the results of this test since this upgrade may increase the relative performance of the video system and hence the score of this test. Hopefully, IBM will update the OS/2 version number (i.e. 2.01) so that the difference will be seen in the BenchTech header. Note: This test will take advantage of a math co-processor such as the Intel 80387, if one is installed on your system. Such a co-processor will greatly reduce the time spent doing certain mathematical functions. However, even if you had an infinitely fast co-processor, the test still reflects the other functions which do not take advantage of this chip. For example, suppose that, on a machine without a co-processor, the test takes 100 seconds overall, 20 seconds of which are devoted to math functions which use the co-processor. Running it on a machine that has a co-processor that performs these function at 100 times or even 1000 times the speed of the system without the co-processor would yield an overall advantage of only 20%. Since there are more functions in this test that do not use the co-processor than in the BenchTech CPU test, the overall effect of the co-processor is less dramatic than in the BenchTech CPU Test. ═══ 10. BenchTech Test Suite ═══ The BenchTech test suite consists of twenty-one benchmark tests, three application-oriented tests, and one user-specified test. For information about the individual tests, double click on any of the following highlighted text. 1. Floating Point Math 2. Integer Math 3. Data Movement 4. Memory Access 5. Sieve of Eratosthenes 6. CPU Cycles 7. Dhrystone 8. Disk Sequential Write 9. Disk Sequential Read 10. Disk Random Access Read/Write 11. Diskette Sequential Write 12. Diskette Sequential Read 13. Diskette Random Access Read/Write 14. Metafile Display 15. Bitmap Bit-Block Transfer (BitBlt) 16. Bitmap Scaling 17. Text Display 18. Simple Vectors 19. Complex Vectors 20. Filled Vectors 21. Towers of Hanoi 22. Application Load 23. Excel 24. DeScribe 25. User Tests ═══ 11. Description of Pages ═══ This section describes each of the pages of the notebook in some detail. Double click on any of the following highlighted text for details about the corresponding page. 1. Results Page 2. CPU Page 3. Disk Page 4. Video Page 5. Application Page 6. User Page 7. Options Page 8. Logging Options Page 9. Macro Page Each of the pages in the notebook has the name of the page displayed on an associated tab. To move to a page, pick on the corresponding tab. Alternately, use the menu option Pages to jump to a specific page. Each of the Pages menu options also has an associated keyboard shortcut. ═══ 11.1. Results Page ═══ The first page in the notebook, the Results Page, contains an output window that displays results from all tests that run under BenchTech. This output window is read only, but you may select text and copy it to the clipboard. The results can then be pasted into other OS/2 applications. For detail on cut and paste, see Edit Menu Option. For an indication of how long each test may take, as well as a listing of the units that each test's results are reported in, see Test Duration and Units. There are also special output windows that show the latest results of the BenchTech CPU Test and BenchTech Composite Test. Five buttons related to the starting and stopping of BenchTech tests are located on the Results Page. These buttons are also located on the menu bar for users who prefer the pull-down menus. These buttons are: 1. Start CPU 2. Start Composite 3. Start Selected 4. Start Macro 5. Halt Tests The window labeled Status contains an indication of the status of BenchTech at any given time. For example, when a test is running the status window displays Running, and when a test is being halted, the indicator displays Halting. The results page also contains buttons to Clear Results and Print Results, which Clear and Print the results window, respectively. ═══ 11.1.1. Start BenchTech CPU ═══ The Start BenchTech CPU button begins the execution of the BenchTech CPU Test. For a description of this test, see BenchTech CPU Test. Note: If the Pause option is greater than zero, BenchTech will pause for approximately the specified period before starting the test. Similarly, the Repeat option is used to determine how many times to execute the test. For a description of these options, see Pause and Repeat, in the description of the Options Page. ═══ 11.1.2. Start BenchTech Composite ═══ The Start BenchTech Composite button begins the execution of the BenchTech Composite Test. For a description of this test see BenchTech Composite Test. Note: If the Pause option is greater than zero, BenchTech will pause for approximately the specified period before starting the test. Similarly, the Repeat option is used to determine how many times to execute the test. For a description of these options, see Pause and Repeat, in the description of the Options Page. ═══ 11.1.3. Start Selected Tests ═══ The Start Selected button begins the execution of any tests that have been selected on the CPU Page, the Disk Page, the Video Page, the Application Page or the User Page. The tests will be executed in the order that they appear on the notebook pages. Note: If the Pause option is greater than zero, BenchTech will pause for approximately the specified period before starting the test. Similarly, the Repeat option is used to determine how many times to execute the test. For a description of these options, see Pause and Repeat, in the description of the Options Page. ═══ 11.1.4. Start Selected Macro ═══ The Start Selected Macro button will begin the execution of the macro that has been selected on the Macro Page. If no macro has been selected and execution is requested, the macro handler will produce a non-zero return code in the macro results window. ═══ 11.1.5. Halt Tests ═══ To stop a test, select the Halt Tests button. The currently running test should stop within a second or two, although some tests may take longer. Halting a test will also stop the execution of any selected tests that are waiting to run. This option can also be used to stop a macro in progress. Note: The keyboard shortcut "Control + H" is also available for this command. This keyboard shortcut may be useful during video tests, where getting mouse input in quickly enough can be difficult. ═══ 11.1.6. Clear Results ═══ The Clear Results button will clear the contents of the results window. If the Header message type is set on the Logging Options Page, a new header will be written into the results window. Note: When results are cleared using this button, they are not logged even if the logging option is set to Append or write. For information on logging, see Logging Options Page. ═══ 11.1.7. Print Results ═══ The Print Results button will print the contents of the results window to your default OS/2 printer. Note: The results window contents are cleared when it contains approximately 30,000 characters. If the Logging Option is set to Append or Write, the contents are saved to a log file before the window is cleared. For information on logging, see Logging Options Page. ═══ 11.2. CPU Page ═══ The CPU Page allows you to select any combination of seven tests that relate to CPU and memory performance. To select a test for execution, pick on the corresponding checkbox. To select all seven, pick on the button labeled All CPU/Memory Tests. To execute the selected tests, select the Start Selected menu option under Start on the menu bar, or jump to the Results Page and click on the Start Selected button. For more information about the tests on this page, double click on any of the following highlighted text. 1. Floating Point Math 2. Integer Math 3. Data Movement 4. Memory Access 5. Sieve of Eratosthenes 6. CPU Cycles 7. Dhrystone ═══ 11.3. Disk Page ═══ The Disk Page allows you to select any combination of six tests that relate to disk drive and diskette drive performance. To select a test for execution, pick on the corresponding checkbox. To select all six, pick on the button labeled "All Disk/Diskette tests". To execute the selected tests, select the Start Selected menu option under Start on the menu bar, or jump to the Results Page and click on the Start Selected button. For more information about the tests on this page, double click on any of the following highlighted text. 1. Disk Sequential Write 2. Disk Sequential Read 3. Disk Random Access Read/Write 4. Diskette Sequential Write 5. Diskette Sequential Read 6. Diskette Random Access Read/Write ═══ 11.4. Video Page ═══ The Video Page allows you to select any combination of eight tests that relate to video system performance. To select a test for execution, pick on the corresponding checkbox. To select all eight, pick on the button labeled "All Video tests". To execute the selected tests, select the Start Selected menu option under Start on the menu bar, or jump to the Results Page and click on the Start Selected button. For more information about the tests on this page, double click on any of the following highlighted text. 1. Metafile Display 2. Bitmap Bit-Block Transfer (BitBlt) 3. Bitmap Scaling 4. Text Display 5. Simple Vectors 6. Complex Vectors 7. Filled Vectors 8. Towers of Hanoi Note: The first release of IBM OS/2 2.0 did not contain the 32 bit graphics engine. IBM has made it known that this feature will be added to OS/2 sometime in 1992. Care needs to be taken when comparing the results of this test since this upgrade may increase the relative performance of the video system and hence the results of these tests. Hopefully, IBM will update the OS/2 version number (i.e. 2.01) so that the difference will be seen in the BenchTech header. ═══ 11.5. Applications Page ═══ The Application Page allows you to select any combination of three tests that demonstrate real OS/2 application performance. To select a test for execution, pick on the corresponding checkbox. To execute the selected tests, select the Start Selected menu option under Start on the menu bar, or jump to the Results Page and click on the Start Selected button. Note: The Excel test uses Microsoft (r) Excel for OS/2 and the DeScribe (r) test uses the DeScribe Word Processor. These packages are required to run these tests and are not included with BenchTech. Note: Microsoft is a registered trademark of the Microsoft Corporation. DeScribe is a registered trademark of DeScribe, Inc. For more information about the tests on this page, double click on any of the following highlighted text. 1. Application Load 2. Excel 3. Describe ═══ 11.6. User Page ═══ The User Page allows you to specify your own tests to run under BenchTech. The test can consist of OS/2, OS/2 PM, DOS, or Win-OS/2 programs. To select a test for execution, pick on the corresponding checkbox. To execute the selected tests, select the Start Selected menu option under Start on the menu bar, or jump to the Results Page and click on the Start Selected button. For more information about the User Test, double click on the following highlighted text. 1. User Test ═══ 11.7. Options Page ═══ The Options Page allows you to set certain options related to how the BenchTech tests are executed. The options are: 1. Pause 2. Repeat (Number of times to execute tests) 3. Keep/Deselect (After Tests Complete) 4. System Disk Cache 5. Graphics Window Size ═══ 11.7.1. Pause ═══ The Pause option allows you to specify how long BenchTech pauses before each test begins execution. You may use the spin button to "spin" to a pause value or type over the current value to set your own. ═══ 11.7.2. Repeat ═══ The Repeat option allows you to specify how many times each test is to be executed. You may use the spin button to "spin" to a repeat value or type over the current value to set your own. ═══ 11.7.3. Keep/Deselect (After Tests Complete) ═══ If the Keep Tests Selected option is selected, when BenchTech has completed execution of the tests, the selected tests will remain selected. If, however, the Deselect All Tests option is selected, the tests will be deselected after execution. ═══ 11.7.4. System Disk Cache ═══ The System Disk Cache options apply only to the Disk tests. If the System Disk Cache Option is set to Use, BenchTech will use the system disk cache (as specified in your OS/2 CONFIG.SYS file). If this option is set to Bypass, BenchTech will not use the system disk cache. The BenchTech Composite Test bypasses the system cache. Note: BenchTech has no control over a disk cache implemented in hardware, such as a caching SCSI controller board. ═══ 11.7.5. Graphics Window Size ═══ The Graphics Window Size options apply only to the Video tests. This option is used to determine how the graphics window is to be sized. 1. "Fixed Size" - The size of the window is constant in pixels. 2. "Proportional Display" - The size of the window is proportionate to the size of your video display. On a standard VGA screen this option produces the same size window as the fixed size option. On a larger display (i.e. 800 x 600 or 1024 x 768, etc), the window is proportionately larger. 3. "Semi-Proportional" - This option essentially is a compromise between Fixed and Proportional Display. The reason for this option is that as the display resolution becomes greater, people tend to make window sizes that are larger than their VGA counterparts, but not proportionately larger. BenchTech uses the same scheme to enlarge the size of the notebook as the screen size gets larger. On a VGA screen, the notebook covers nearly the entire screen. On a 1024 x 768 display, we enlarge the notebook, but by a smaller factor. The result is a notebook size that is readable, but doesn't occupy too much display area. 4. "Full Screen" - This option results in a window size equal to the size of your display screen. Note: The BenchTech Composite Test uses a fixed size window and is not affected by this option. ═══ 11.8. Logging Options Page ═══ The Logging Options Page allows you to set options concerning the logging of BenchTech results. Definition of Logging The options are: 1. Output Message Types 2. Log File Option 3. Change Log File Name ═══ 11.8.1. Output Message Types ═══ BenchTech produces four types of output messages that are written to the results window. They are: 1. Results 2. Errors 3. Informational Messages 4. Header Information Examples of each type of message are given below. A particular message is written to the results window only if its corresponding type was enabled in the Output Message Type checkbox list. Example of results message: 20:27:22 Towers of Hanoi: Time to solve 10.22 seconds or 20:31:43 BenchTech CPU: Score = 25.6 Example of error messages: Disk Write: was Halted or Unable to create file, may not be enough space on disk Disk Write: ended in Error Example of Informational messages 20:27:12 Towers of Hanoi: Test starting 20:27:22 Towers of Hanoi: (10 disks, size 12, total disks moved 1023) Example of header messages: BenchTech Output Log Created Wed May 20 20:13:46 1992 BenchTech Version 1.00 OS/2 Version 2.00 ═══ 11.8.2. Log File Option ═══ The Log File Option allows the user to specify how or if results are logged to a log file. Select Write to write log data over any previous log file data. Select Append to append log data to a log file. Select Off to disable writing log data to a log file. Log data are generally written to the log file when BenchTech is ended. When you close the BenchTech program, if this option is set to Write or Append the log data will be written to the log file. In addition, if the results window contents reach approximately 30,000 characters, the contents of the windows are logged according to this option and the results window is cleared. To insure that none of your results are lost, set this to Append before running any tests. ═══ 11.8.3. Change Log File Name ═══ The Change... button, under Current Log File, allows you to specify the name of the file where you would like logging to take place. Clicking on this button brings up the standard OS/2 file dialog box, where you make your log file selection. ═══ 11.9. Macro Page ═══ The Macro Page contains controls that allow you to select, perform simple editing on, and execute a BenchTech macro. This page also contains windows that show the return code from the macro. Definition of Macro BenchTech macros are based on the Rexx procedure language. Rexx is an easy to use, structured programming language that is included with the OS/2 operating system. Rexx is also available on several other IBM and non-IBM platforms. Several BenchTech specific functions are available in addition to the standard Rexx instruction set. These function only operate when the BenchTech macro is executed under the control of BenchTech. The Macro Page allows you to: 1. Select a Macro for Execution 2. Perform Simple Editing on the Macro 3. Start the Selected Macro The additional Rexx functions provided by BenchTech are described in the BenchTech Macro Functions section. ═══ 11.9.1. Select a Macro for Execution ═══ The Select button on the Macro Page allows you to specify the macro to be executed. Clicking on this button brings up the standard OS/2 file dialog box, where you make your macro file selection. ═══ 11.9.2. Edit the Selected Macro ═══ To edit the selected macro, use your mouse to make the macro edit area current. Edit the file as you would with the OS/2 system editor. To save the file, select Save. To save the file under a different, select Save As.... To cancel the edit session, pick Cancel.. Note: You cannot execute a macro when it has been edited but not saved. To execute a macro once it has been changed, you must select Cancel, Save As, or Save. Note: You don't have to use the editor built into BenchTech. Use whatever editor you prefer. ═══ 11.9.3. Start the Selected Macro ═══ The Start Selected Macro button on the Macro Page will begin the execution of the selected macro. This button performs the same function as the Start Macro menu option under Start on the menu bar or the Start Selected button on the Results Page. ═══ 11.9.4. BenchTech Macro functions ═══ The Rexx functions that are provided by BenchTech are as follows: 1. BT_OutputMessage 2. BT_ClearResults 3. BT_PrintResults 4. BT_StartSelected 5. BT_StartCPU 6. BT_StartComposite 7. BT_SelectTest 8. BT_SetOptions 9. BT_TurnToPage ═══ 11.9.5. BT_OutputMessage ═══ BT_OutputMessage Expected number of parameters: 1 Parameter 1: string to be displayed This function will echo the parameter to the results display. Example: BT_OutputMessage('Hello...'); will output the words "Hello..." on the results display. ═══ 11.9.6. BT_ClearResults ═══ BT_ClearResults Expected number of parameters: 0 This function will clear the results window in the same manner as the Clear Results button on the Results Page. Example: BT_ClearResults(); will clear the results display. ═══ 11.9.7. BT_PrintResults ═══ BT_PrintResults Expected number of parameters: 0 This function will print the results window in the same manner as the Print Results button on the Results Page. Example: BT_PrintResults(); will print the results display to your default OS/2 printer. ═══ 11.9.8. BT_StartSelected ═══ BT_StartSelected Expected number of parameters: 0 This function will begin execution of the selected tests in the same manner as manually selecting the Start Selected button on the Results Page. Example: BT_SelectTest('Dhrystone','TRUE'); BT_StartSelected(); will start the Dhrystone test, as well as any other currently selected tests. ═══ 11.9.9. BT_StartCPU ═══ BT_StartCPU Expected number of parameters: 0 This function will begin execution of the BenchTech CPU Test in the same manner as manually selecting the Start CPU button on the Results Page. Example: BT_StartCPU(); will start the BenchTech CPU Test. ═══ 11.9.10. BT_StartComposite ═══ BT_StartComposite Expected number of parameters: 0 This function will begin execution of the BenchTech Composite Test in the same manner as manually selecting the Start Composite button on the Results Page. Example: BT_StartComposite(); will start the BenchTech Composite Test. ═══ 11.9.11. BT_SelectTest ═══ BT_SelectTest Expected number of parameters: 2 Parameter 1: Name of test Parameter 2: TRUE (select) or FALSE (deselect) This function will select or deselect the specified test. Valid test names for parameter 1 are: FLOATINGPOINT INTEGER DATAMOVEMENT MEMORYACCESS SIEVE CYCLES DHRYSTONE DISKWRITE DISKREAD DISKRANDOM DISKETTEWRITE DISKETTEREAD DISKETTERANDOM METAFILE BITMAPDONTSTRETCH BITMAPSTRETCH TEXTDISPLAY SIMPLEVECTOR COMPLEXVECTOR FILLVECTOR HANOI APPLICATIONLOAD EXCEL DESCRIBE USER The names may be specified in any combination of upper and lower case. Example: BT_SelectTest('Metafile','TRUE'); BT_SelectTest('Excel','false'); will select the Metafile Test, and deselect the Excel test. ═══ 11.9.12. BT_SetOptions ═══ BT_SelectTest Expected number of parameters: 1, 2 or 3 Parameter 1: Name of test or option Parameter 2: Options setting 1 Parameter 3: Options setting 2 This function will begin set the specified option(s). Valid values for parameter 1 are: SIEVE DHRYSTONE HANOI EXCEL DESCRIBE USERSETUPCOMMAND USERSETUPPARAMETERS USERTIMEDCOMMAND USERTIMEDPARAMETERS USERCLEANUPCOMMAND USERCLEANUPPARAMETERS DISK DISKETTE PAUSE REPEAT DESELECT KEEPSELECTED USECACHE BYPASSCACHE OUTPUTRESULTS OUTPUTERRORS OUTPUTINFORMATION OUTPUTHEADER LOGAPPEND LOGWRITEOVER LOGOFF GRAPHICSFIXED GRAPHICSPROP GRAPHICSSEMI GRAPHICSFULL The names may be specified in any combination of upper and lower case. The Hanoi Test is the only option that takes 3 parameters. Example: BT_SetOption('Sieve','500'); BT_SetOption('DeScribe','D:\TEST\DESCRIBE'); BT_SetOption('UserTimedCommand','C:\TEST\TEST.EXE'); BT_SetOption('UserTimedParameters','/l 25'); BT_SetOption('Hanoi','10','12'); BT_SetOption('DISK','F:'); BT_SetOption('pause','2.0'); BT_SetOption('BypassCache'); BT_SetOption('OutputResults','TRUE'); BT_SetOption('LogAppend'); will 1. set the number of iteration for the Sieve Test to 500 2. set the pathname for the DeScribe Test to 'D:\TEST\DESCRIBE' 3. set the path and file name for the timed program in the User Test to 'C:\TEST\TEST.EXE' 4. set the parameters for the timed program in the User Test to '/l 25' 5. set the number of disk to 10 and the disk size to 12 for the Hanoi Test 6. set the disk to be tested for the disk tests to F: 7. set the pause value to 2.0 seconds 8. set the System Cache option to Bypass 9. set the Results output message type to on 10. set the Log File Option to Append ═══ 11.9.13. BT_TurnToPage ═══ BT_TurnToPage Expected number of parameters: 1 Parameter 1: Page to turn to This function will change the current notebook page to the specified page. Valid values for parameter 1 are: NEXT PREVIOUS RESULTS CPU VIDEO DISK APPLICATION USER OPTIONS LOGGING MACRO The names may be specified in any combination of upper and lower case. Example: BT_TurnToPage(LOGGING); will change the current notebook page to the Logging Options Page. ═══ 12. Menu Options ═══ This section describes BenchTech menu options. Double click on the selections below for a description of the corresponding menu option. 1. Edit 2. Start 3. Pages 4. Options 5. Help ═══ 12.1. Edit Menu Option ═══ The Edit menu option contains menu options related to the OS/2 clipboard. You can use these options to transfer text data between BenchTech and other OS/2 applications. Note: The Cut, Paste, and Clear options are only enabled on the macro edit window. Copy is always enabled, i.e. you can copy data from any window, including the results window, and paste it into another application. You can cut, copy, paste or clear data in the macro edit window. 1. Cut 2. Copy 3. Paste 4. Clear ═══ 12.1.1. Cut ═══ The Cut menu option allows you to "cut" text data from the macro edit window. "Cut" means delete the data from the window and place it in the clipboard. To "cut" text: 1. Select the text to be cut. 2. Select Edit. 3. Select Cut. The keyboard shortcut for this option is "Shift + Delete." ═══ 12.1.2. Copy ═══ The Copy menu option allows you to "copy" text data from any BenchTech window. "Copy" means to copy the selected text to the clipboard. To "copy" text: 1. Select the text to be copied. 2. Select Edit. 3. Select Copy. The keyboard shortcut for this option is "Control + Insert." ═══ 12.1.3. Paste ═══ The Paste menu option allows you to "paste" text data into the macro edit window. "Paste" means copy the data from clipboard and place it in the macro being edited. To insert text that you have cut or copied to the clipboard into the macro being edited: 1. Move the cursor to the point in the file where you want to insert the text. 2. Select Edit. 3. Select Paste. The keyboard shortcut for this option is "Shift + Insert." ═══ 12.1.4. Clear ═══ The Clear menu allows you to delete selected text without copying it to the clipboard. To delete text: 1. Select the text to be deleted. 2. Select Edit. 3. Select Clear. The keyboard shortcut for this option is "Delete." ═══ 12.2. Start Menu Option ═══ The Start menu option contains menu options to start and stop the BenchTech tests. 1. Start CPU 2. Start Composite 3. Start Selected 4. Start Macro 5. Halt Tests ═══ 12.2.1. Start CPU ═══ The Start CPU menu option starts the BenchTech CPU test. This button performs the same function as the Start BenchTech CPU button on the Results Page. The keyboard shortcut for this option is "Control + P." ═══ 12.2.2. Start Composite ═══ The Start Composite menu option starts the BenchTech Composite test. This button performs the same function as the Start BenchTech Composite button on the Results Page. The keyboard shortcut for this option is "Control + C." ═══ 12.2.3. Start Selected ═══ The Start Selected menu option starts the currently selected tests. This button performs the same function as the Start Selected button on the Results Page. The keyboard shortcut for this option is "Control + S." ═══ 12.2.4. Start Macro ═══ The Start Macro menu option starts the currently selected macro. This button performs the same function as the Start Macro button on the Results Page. The keyboard shortcut for this option is "Control + M." ═══ 12.2.5. Halt ═══ The Halt menu option stops the current test or macro. This button performs the same function as the Halt Tests button on the Results Page. The keyboard shortcut for this option is "Control + H." ═══ 12.3. Pages Menu Option ═══ The Pages menu option presents a list of menu (and keyboard) shortcuts for jumping to pages in the notebook. 1. Next Page 2. Previous Page 3. Results Page 4. CPU Page 5. Disk Page 6. Video Page 7. Application Page 8. User Page 9. Options Page 10. Logging Options Page 11. Macro Page ═══ 12.3.1. Next Page ═══ Use this menu option to turn to the next page in the notebook. The keyboard shortcut for this option is "Alt + N." ═══ 12.3.2. Previous Page ═══ Use this menu option to turn to the previous page in the notebook. The keyboard shortcut for this option is "Alt + P." ═══ 12.3.3. Results Page ═══ Use this menu option to jump directly to the Results Page in the notebook. The keyboard shortcut for this option is "Alt + R." ═══ 12.3.4. CPU Page ═══ Use this menu option to jump directly to the CPU Page in the notebook. The keyboard shortcut for this option is "Alt + C." ═══ 12.3.5. Disk Page ═══ Use this menu option to jump directly to the Disk Page in the notebook. The keyboard shortcut for this option is "Alt + D." ═══ 12.3.6. Video Page ═══ Use this menu option to jump directly to the Video Page in the notebook. The keyboard shortcut for this option is "Alt + V." ═══ 12.3.7. Application Page ═══ Use this menu option to jump directly to the Applications Page in the notebook. The keyboard shortcut for this option is "Alt + A." ═══ 12.3.8. User Page ═══ Use this menu option to jump directly to the User Page in the notebook. The keyboard shortcut for this option is "Alt + U." ═══ 12.3.9. Options Page ═══ Use this menu option to jump directly to the Options Page in the notebook. The keyboard shortcut for this option is "Alt + O." ═══ 12.3.10. Logging Options Page ═══ Use this menu option to jump directly to the Logging Options Page in the notebook. The keyboard shortcut for this option is "Alt + L." ═══ 12.3.11. Macro Page ═══ Use this menu option to jump directly to the Macro Page in the notebook. The keyboard shortcut for this option is "Alt + M." ═══ 12.4. Options ═══ The Options menu selection presents the user the opportunity to save most of the BenchTech options. BenchTech uses the saved option the next time the program is started. In addition, there are options to revert back to the saved options or to revert back to the original BenchTech defaults. 1. Save Options 2. Revert to Saved 3. Revert to Defaults ═══ 12.4.1. Save Options ═══ Use the Save Options menu option to save user options so that they may be used the next time BenchTech is started, or so they can be recalled. The options are saved in a file called BT.INI in the directory that BenchTech was started from. If the file does not exist, it is created. The options that are saved are: 1. The number of iterations specified for the Sieve Test 2. The number of iterations specified for the Dhrystone Test 3. The number of disks specified for the Hanoi Test 4. The size of the disks specified for the Hanoi Test 5. All options on the Options 1 page 6. All Logging options on the Options 2 page 7. The Log File name. If the options have been saved (and BT.INI has not been erased), BenchTech will use the saved options the next time it is started. ═══ 12.4.2. Revert to Saved ═══ The Revert to Saved menu option allows you to change all options back to the values previously saved. See Save Options for information about how options are saved. ═══ 12.4.3. Revert to Defaults ═══ The Revert to Defaults menu option allows you to change all options back to the values originally supplied with BenchTech. See Save Options for information about what options are saved, and consequently, what options are restored using Revert to Defaults. ═══ 12.5. Help Menu Option ═══ The Help menu option presents help information about the BenchTech Program. 1. Help Index 2. General Help 3. Using Help 4. Product Information ═══ 12.5.1. Help Index help ═══ Use Help index to display an alphabetic list of references to the help topics in the active program. Note: To see all text of the index entries, maximize the Index window. You can select an index entry to display the help topic that is related to that index entry. ═══ 12.5.2. Help for General Help ═══ The General help panel is an overview of the features of BenchTech. To select one of the topics shown on the General Help panel, double click on the highlighted text. ═══ 12.5.3. Help for Using Help ═══ Use Using help to find out how to use the help facility. ═══ 12.5.4. BenchTech Product Information help ═══ This dialog contains copyright information about BenchTech. ═══ 13. Description of Tests ═══ This section describes each of the BenchTech tests. ═══ 13.1. Floating Point Math ═══ The Floating Point Math Test is designed to measure the performance of code that is typical of floating point intensive programs, such as certain spreadsheets and Computer Aided Drafting (CAD) programs. "Floating point" is a way of representing real numbers including digits to the right of a decimal point. Floating point numbers are not necessarily represented exactly, but the degree of precision maintained usually far exceeds what is necessary for a given application. For example the number 6.0 is a floating point number, as is 3.1415926. Floating point numbers are often written in scientific notation like 5.4321e-12 where the e means "exponent" and the digits to the right of the e are the base 10 exponent, i.e. 5.4321 x 10 raised to the -12th power. This test will utilize a standard math co-processor, like the Intel 80387, if one is configured on your system. It will not use specialized math processors. If your system does not have a math co-processor, this test will run, although it will take considerably longer to complete. See Test Duration and Units. ═══ 13.2. Integer Math ═══ The Integer Math Test performs integer mathematics using both 16 bit and 32 bit integers (emphasis is on the 32 bit calculations, however). The test does not utilize a math co-processor. ═══ 13.3. Data Movement ═══ Much of what a computer does is move data from place to place inside of its memory. Many benchmarks focus on the mathematical computing power of the system, but fail to take into account the cost of moving data. The Data Movement Test focuses on this aspect of performance by measuring how long it takes to move data within memory. Examples of situations where data movement plays an important role in computing power are sorting of data, any text processing, and graphics work. ═══ 13.4. Memory Access ═══ The Memory Access Test measures how long it takes to access your computers memory. This access is not sequential, making it more difficult for your computer's caching algorithm to have the data ready in cache. This test queries a predefined amount of memory within a certain address range and reports the results. It then increases the range size but does not increase the amount of memory to be queried. If access to every location in memory takes the same amount of time, the results will be equal for each range size. If, however, it takes longer on average to access memory that is farther away, it will take longer for the test to complete at each progressively larger range. For example, on a computer with no memory cache, the results for each range size will be virtually the same. On a computer with a cache, like a 486 33 Mhz clone system with 256K cache, the time to complete the memory access gets appreciably longer as the range gets larger. Given that most of today's high performance computer systems rely on cache memory, this test is especially useful to understand how the true performance of the CPU and memory will be. ═══ 13.5. Sieve of Eratosthenes ═══ The Sieve of Eratosthenes is BenchTech's version of a very simple public domain benchmark that counts prime numbers between 0 and some number (this test use 8190). The test measures the amount of time it takes to complete the count. Sieve is an example of a benchmark that is likely to fit nicely within cache and therefore should not be used to predict any kind of overall performance of the system. ═══ 13.6. CPU Cycles ═══ The CPU Cycles Test is a very simple benchmark that measures how fast your system is able to perform simple instructions. You may find the results of this test to be on the order of the rated MIPS of your machine. CPU Cycles is an example of a benchmark that is likely to fit nicely within cache and therefore should not be used to predict any kind of overall performance of the system. CPU Cycles is also not a very well behaved piece of code in the sense that it is very unaware that it is running is a multi-tasking operating system. Therefore, when halting CPU Cycles it may take up to a minute to finally halt the test. Note: If you need to stop this test immediately, close the program. ═══ 13.7. Dhrystone ═══ The Dhrystone is BenchTech's version of the Dhrystone public domain benchmark. The only modifications that we have made relate to: 1. Allowing BenchTech to halt the benchmark 2. Making the number of loops a modifiable value 3. Displaying the results in the results window instead of standard output ═══ 13.8. Disk Write ═══ The Disk Write Test measures the amount of time it takes to do a sequential write to your hard disk drive. This test requires approximately 3.5 Megabytes of free space on your disk drive. Results for this test are expressed in Megabytes per second. Definition of Megabyte. Note: You have the option of either using or bypassing the OS/2 disk cache specified in your CONFIG.SYS. For more detail on how to set this option, see System Cache. Results obtained using the System Cache tend to vary, particularly for large caches. One effect that can be noticed when using this test is the so-called "lazy-write" of the operating system. If you run this test bypassing the cache, you'll notice that the disk drive light will go out when the test ends. If you run this test using the disk cache, the test will end before the disk drive light goes off. Because of this effect, we recommend that you set you test up so that the lazy write does not occur during the subsequent test, either by bypassing the cache or using a large pause value. ═══ 13.9. Disk Read ═══ The Disk Read Test measures the amount of time it takes to do a sequential read from your hard disk drive. This test requires approximately 1.2 Megabytes of free space on your disk drive. Results for this test are expressed in Megabytes per second. Definition of Megabyte. Note: You have the option of either using or bypassing the OS/2 disk cache specified in your CONFIG.SYS. For more detail on how to set this option, see System Cache. Results obtained using the System Cache tend to vary, particularly for large caches. ═══ 13.10. Disk Random Access ═══ The Disk Random Access Test measures the amount of time it takes to do a series of read/writes from your hard disk drive. These operations are performed in a manner that is not sequential. This test requires approximately 1.2 Megabytes of free space on your disk drive. Results for this test are expressed in Megabytes per second. Definition of Megabyte. Note: You have the option of either using or bypassing the OS/2 disk cache specified in your CONFIG.SYS. For more detail on how to set this option, see System Cache. Results obtained using the System Cache tend to vary, particularly for large caches. ═══ 13.11. Diskette Write ═══ The Diskette Write Test is the same as the Disk Write Test, except that less data is written, and the target drive is a diskette drive. For more information see Disk Write Test. ═══ 13.12. Diskette Read ═══ The Diskette Read Test is the same as the Disk Read Test, except that less data is written, and the target drive is a diskette drive. For more information see Disk Read Test. ═══ 13.13. Diskette Random Access ═══ The Diskette Random Access Test is the same as the Disk Random Access Test, except that less data is written, and the target drive is a diskette drive. For more information see Disk Random Access Test. ═══ 13.14. Metafile ═══ The Metafile Test measures the amount of time it takes to draw a predefined set of metafiles to your display. Definition of "metafile." ═══ 13.15. Bitmap Bit-Block Transfer ═══ The Bitmap Bit-Block Transfer Test measures how long it takes to redraw a bitmap on your display. Each time the bitmap is re-drawn, the test uses a different mixing pattern and raster operation. Definition of "bitmap." The process of drawing the bitmap is called a Bit-Block transfer or BitBlt for short. This test uses the BitBlt function provided by the OS/2 Graphics Programming Interface (GPI). This test does not stretch the bitmap. Definition of "BitBlt." ═══ 13.16. Bitmap Scaling ═══ The Bitmap Scaling Test measures how long it takes to redraw a bitmap on your display. Each time the bitmap is re-drawn, the test uses a different mixing pattern and raster operation. Definition of "bitmap". The process of drawing the bitmap is called a Bit-Block transfer or BitBlt for short. This test uses the BitBlt function provided by the OS/2 Graphics Programming Interface (GPI). This test is very similar to the Bitmap Bit-Block Transfer Test, except that it scales the bitmap to the size of the graphics output window. Since the effect of scaling the bitmap greatly increases the amount of time required to draw the bitmap, the number of iterations of drawing the bitmap is less in this test than in the Bitmap Bit-Block Transfer Test. Definition of "BitBlt". ═══ 13.17. Text Display ═══ The Text Display Test measures the time it takes to write a predefined sequence of text lines to the display. This test uses the OS/2 Graphics Programming Interface (GPI) to write the text to the display. ═══ 13.18. Simple Vectors ═══ The Simple Vectors Test measures how long it takes to draw a set of straight lines. Although the test may look complicated, only straight lines are actually drawn. Note: This tests will utilize a math co-processor if one is installed in your system. ═══ 13.19. Complex Vectors ═══ The Complex Vectors Test uses three OS/2 GPI functions to draw more complex vectors on the display. This test measures the total time to draw the vectors. 1. Poly line (line segments) 2. Poly Fillet (fillet curves generated from multiple reference points) 3. Poly Spline (spline curves generated from multiple reference points) ═══ 13.20. Filled Vectors ═══ The Filled Vectors Test takes advantage of the OS/2 GPIs ability to fill a dynamically drawn area. The test measures the time it takes to draw a series of filled shapes. ═══ 13.21. Towers of Hanoi ═══ The Towers of Hanoi is BenchTech's implementation of the well known "Towers of Hanoi" simulation. The test measures how long it takes to solve the Towers of Hanoi puzzle. This puzzle consists of three "towers" and a number of "disks," where each disk is of a different width. Initially, the disks are placed on one of the towers so that the widest disk is at the bottom of the tower, and the next largest is on top of it and so on. The object of the puzzle is to move the disks from the initial tower to one of the other towers. Only one disk can be moved at a time, and it must be moved to one of the other towers such that no larger disk is ever on top of a smaller one. Although Hanoi is a fairly trivial benchmark, it is interesting because it demonstrates two things about your computer, video re-draw and CPU speed. There are two parameters that you can modify, the number of disks and also the size of each disk. As you might expect, the time to complete the puzzle tends to vary exponentially as the number of disks are increased. The time to complete the puzzle is also influenced by the size of the disks, however, since larger disks take slightly longer to draw and erase than smaller ones. Note: The maximum allowable value for the number of disks is 16. The maximum allowable value for the disk size is 18. If you enter a higher number, the test will use these values. ═══ 13.22. Application Load ═══ The Application Load Test measures how long it takes to load (and dismiss) a special version of the BenchTech program (x$$$bt00.exe). This test will generally run faster the second and subsequent runs. See the "Fairness/Testing Methodology" section of the users manual for a discussion of this aspect of testing. ═══ 13.23. Excel ═══ The Excel Test measures how long it takes to load Microsoft (r) Excel for OS/2 and execute an predefined Excel macro. The Excel macro creates a spreadsheet and fills it with data. It then creates a graph based on the data and forces several re-calculations of the spreadsheet data and re-drawing of the graph. The Excel Test requires Microsoft Excel for OS/2, which is a commercial spreadsheet package that is not included with BenchTech. Version 1.0 of BenchTech was tested with version 3.0 of Microsoft Excel for OS/2. BenchTech will search your disk for a directory called \EXCEL. If found it will fill this directory in the Excel pathname entry field. If the directory is not found, the pathname entry field will display "(not found)". If this directory is incorrect, type over it with the pathname where Excel for OS/2 can be found. Note: There are a few things you may need to know about this test: 1. This test is dependant on your installation of Microsoft Excel and OS/2. Care should be taken when comparing results from this test with those from another computer system. 2. This test will generally run faster the second and subsequent runs. See the "Fairness/Testing Methodology" section of the users manual for a discussion of this aspect of testing. 3. This test uses the clipboard to add data to the Excel spreadsheet. Do not run the test if there is clipboard data that you need to save. 4. The test copies a file called x$$$bt00.xlm to your XLSTART directory, which is a subdirectory of the directory that contains Excel for OS/2. At the end of the test, this file will be deleted. a. If this directory does not exist, the test will fail. To create it, open an OS/2 window and change directories to the directory containing Excel (i.e. cd \excel), and create a directory called XLSTART (i.e. mkdir xlstart). b. If XLSTART does exist, but has Excel files already in it, the test results will potentially not be valid for comparison because these Excel files will also be executed. 5. This test will not work with Microsoft Excel for Windows. 6. If you want to use your own Excel macro, just replace x$$$bt00.xlm with your macro. The macro should not wait for user input, and it should close the Excel program when it is finished. ═══ 13.24. DeScribe(r) ═══ The DeScribe Test measures how long it takes to load DeScribe and execute an predefined DeScribe macro. The DeScribe Test requires DeScribe for OS/2, which is a commercial word processor package that is not included with BenchTech. Version 1.0 of BenchTech was tested with version 3.0 of DeScribe for OS/2. BenchTech will search your disk for a directory called \DESCRIBE. If found it will fill this directory in the DeScribe pathname entry field. If the directory is not found, the pathname entry field will display "(not found)." If this directory is incorrect, type over it with the pathname where DeScribe for OS/2 can be found. Note: There are several things you may need to know about this test: 1. This test is dependant on your installation of DeScribe and OS/2. Care should be taken when comparing results from this test with those from another computer system. 2. This test will generally run faster the second and subsequent runs. See the "Fairness/Testing Methodology" section of the users manual for a discussion of this aspect of testing. 3. This test uses the clipboard to add data to the DeScribe document. Do not run the test if there is clipboard data that you need to save. 4. This test will not work with DeScribe for Windows. 5. If you want to use your own DeScribe macro, just replace x$$$bt00.dml with your macro. The macro should not wait for user input, and it should close the DeScribe program when it is finished. ═══ 13.25. User Test ═══ The user test allows you to specify your own BenchTech test. There are three parts to this test, the setup program(s), the timed program(s), and the cleanup program(s). Use the setup program to setup any special needs that your program has. Use the cleanup program to cleanup after the program. The setup and cleanup programs are optional. The timed program is timed from when it is started until OS/2 notified BenchTech that it has completed. In a nutshell, BenchTech does the following with this test: 1. If you have specified a setup program, BenchTech launches it with any parameters you have specified. BenchTech waits for the setup program to finish. 2. A timer is started, and the timed program is launched with the parameters you specified, if any. 3. When the timed program finishes, the timer is stopped and the results are reported. 4. If you have specified a cleanup program, BenchTech launches it with any parameters you have specified. BenchTech waits for the cleanup program to finish, and then goes on to the next test, if any. Note: Since BenchTech waits for the programs to complete, it is important that they terminate. If it does not terminate, you must terminate it manually, so that BenchTech will continue to run. If you are unable to terminate it, you will have to close BenchTech. Note: To run commands that are built into OS/2, such as "copy" and "del," enter "CMD.EXE" in the command line, and "/C copy" or "/C del" in the parameter line. To run DOS programs, enter the name of the program in the command line. The program will be run in a full screen session. To run Windows programs, enter "WINOS2.COM" in the command line and the Windows program name in the parameter line. Example: This example replicates the Excel test, and demonstrates the use of the setup and cleanup fields. (The following assumes that the Excel is located in d:\excel and that the xlstart directory exists.) Setup command line cmd.exe Setup parameter line /C copy x$$$bt00.xlm d:\excel\xlstart Timed command line d:\excel\excel.exe Timed parameter line Cleanup command line cmd.exe Cleanup parameter line /C del d:\excel\xlstart\x$$$bt00.xlm ═══ 14. Units and Test Duration ═══ This section lists, for each test, the approximate amount of time each test is expected to take. In addition, this section lists the units that each test reports its results in, or whether higher or lower indicates better performance. Note: The ranges given for the duration are approximate and will vary with your configuration. The numbers given here are some observed times noted when tested on computer systems ranging from a 386SX/16 to a 486DX/33. ┌───────────────┬──────────┬────────────┬───────┐ │Test Name │Duration │Units │High or│ │ │in seconds│ │low │ │ │ │ │better │ ├───────────────┼──────────┼────────────┼───────┤ │BenchTech CPU │7 - 57 │units per │  │ │ │ │second │ │ ├───────────────┼──────────┼────────────┼───────┤ │BenchTech │25 - 93 │units per │  │ │Composite │ │second │ │ ├───────────────┼──────────┼────────────┼───────┤ │Floating Point │1 - 307** │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Integer Math │7 - 32 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Data Movement │9 - 45 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Memory Access │60 - 280 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Sieve of │3 - 50** │seconds │  │ │Eratosthenes │ │ │ │ ├───────────────┼──────────┼────────────┼───────┤ │CPU Cycles │7 - 56 │million │  │ │ │ │cycles per │ │ │ │ │second │ │ ├───────────────┼──────────┼────────────┼───────┤ │Dhrystone │14 - 100**│dhrystones │  │ │ │ │per second │ │ ├───────────────┼──────────┼────────────┼───────┤ │Disk Sequential│6 - 15 │megabytes │  │ │Write │ │per second │ │ ├───────────────┼──────────┼────────────┼───────┤ │Disk Sequential│12 - 25 │megabytes │  │ │Read │ │per second │ │ ├───────────────┼──────────┼────────────┼───────┤ │Disk Random │40 - 60 │megabytes │  │ │Access │ │per second │ │ │Read/Write │ │ │ │ ├───────────────┼──────────┼────────────┼───────┤ │Diskette │30 - 40 │megabytes │  │ │Sequential │ │per second │ │ │Write │ │ │ │ ├───────────────┼──────────┼────────────┼───────┤ │Diskette │40 - 60 │megabytes │  │ │Sequential Read│ │per second │ │ ├───────────────┼──────────┼────────────┼───────┤ │Diskette Random│100 - 200 │megabytes │  │ │Access │ │per second │ │ │Read/Write │ │ │ │ ├───────────────┼──────────┼────────────┼───────┤ │Metafile │17 - 66 │seconds │  │ │Display │ │ │ │ ├───────────────┼──────────┼────────────┼───────┤ │Bitmap │30 - 68 │seconds │  │ │Bit-Block │ │ │ │ │Transfer │ │ │ │ │(BitBlt) │ │ │ │ ├───────────────┼──────────┼────────────┼───────┤ │Bitmap Scaling │36 - 121 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Text Display │7 - 17 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Simple Vectors │8 - 35 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Complex Vectors│13 - 57 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Filled Vectors │9 - 38 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Towers of Hanoi│39 - 159 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │Application │3 - 8 │seconds │  │ │Load │ │ │ │ ├───────────────┼──────────┼────────────┼───────┤ │Excel │35 - 70 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │DeScribe │32 - 64 │seconds │  │ ├───────────────┼──────────┼────────────┼───────┤ │User Tests │N/A │seconds │  │ └───────────────┴──────────┴────────────┴───────┘ ═══ 15. BenchTech glossary ═══ This section contains definitions of some of the terms used in BenchTech and in the BenchTech help panels. ═══ 15.1. Bitmap ═══ bitmap A bitmap is an array of bits organized into rows and columns in which each bit (monochrome bitmap) or series of bits (color bitmap) represent pixels on the screen. Bitmaps are device dependent. ═══ 15.2. BitBlt ═══ BitBlt BitBlt is short for Bit-Block Transfer, a process of moving the contents of a bitmap to a device, such as a display. It can be a simple process where the bits contained in the bitmap are represented directly, or it can involve more complex operations including stretching the bitmap, inverting it, mixing it with another bitmap, or mixing it with a pattern. ═══ 15.3. Clipboard ═══ clipboard The clipboard is a temporary place in memory where text (as well as other types of data) is stored. The clipboard stores text that you have copied or cut so that you can paste it into a file. The clipboard holds one block of text at a time. When you cut or copy a block of text, you replace the text that was there. ═══ 15.4. CPU ═══ CPU CPU stands for Central Processing Unit. The CPU in computers that run OS/2 (at least this version of OS/2) are usually the Intel 80386 or 80486, or equivalent. ═══ 15.5. Logging ═══ logging Logging, in the context of the BenchTech program, means the saving of result data (the data that appears in your results window) to a disk or diskette file. Log data are generally written to the log file when BenchTech is ended. When you close the BenchTech program, if this option is set to Write or Append, the log data will be written to the log file. In addition, if the results window contents reach approximately 30,000 characters, the contents of the window are logged according to this option and the results window is cleared. ═══ 15.6. Macro ═══ macro In the context of this program, a macro is a subprogram that allows an operator to group multiple commands into one file. In BenchTech, macros use the Rexx programming language plus some extra BenchTech specific functions. A BenchTech macro is an ASCII file containing Rexx and BenchTech commands. ═══ 15.7. Megabyte ═══ megabyte A megabyte is defined as 1,048,576 bytes, which is 2 to the 20th power. ═══ 15.8. Metafile ═══ metafile A metafile is a collection of drawing instructions saved in a file. Most metafiles are device independent. An example of a device dependent metafile is one that contains a bitmap. ═══ ═══ The time to complete the floating point test depends largely on whether or not a floating point co-processor is installed. On machine with an installed co-processor, the time to complete this test is typically 1 to 2 seconds. Without a co-processor, it can take 200 - 300 seconds. ═══ ═══ Using the default number of iterations