******************************************* * * * Extend-O-Save Overhead Tester * * by John Eidsvoog & Charles F. Johnson * * Copyright 1994 CodeHead Technologies * * * ******************************************* Release date: Saturday, March 19, 1994 The purpose of the Extend-O-Save Overhead Tester is to show how much processor time is consumed by your Extend-O-Save modules. This may or may not be important to you. If you've ever tried to do a download or printing operation and had a busy Extend-O-Save module kick in, you know that it can grind the operation to a halt, resulting in an aborted download or a printout that never finishes. By carefully choosing which modules you use, you can eliminate or minimize this problem. Operation of the Extend-O-Save Overhead Tester is quite simple. First make sure you have Warp 9 installed with Extend-O-Save enabled. Then just run EOS_TEST.PRG. When the window appears on the screen, you'll see a message telling you that a reference value is being obtained. Do not to move the mouse or your tests will not show true results. After the results are shown, you can hit any key and test the currently installed Extend-O-Save module. The tester will automatically kick in the screensaver for the duration of the test (3 seconds). If you want an accurate test, don't hit a key or move the mouse because it will kick out the screensaver. (However, you may wish to use this feature to test the amount of CPU time taken up in moving the mouse around). When the test is done, you'll see the test value and two percentages. The first percentage is the rate at which the CPU can perform its tasks while the EOS module is kicked in. The second percentage is the amount of the "hit" that the module puts on the CPU. By definition, these two values add up to 100%. When a test starts, you'll hear an ascending sound and when it ends you'll hear a descending sound. The test runs for 3 seconds and is simply a program loop which counts (by continually adding one to a counter) as fast as it can. Since EOS modules operate in the vertical blank interrupt, they will affect the speed at which the tester can count. The amount by which they affect the count depends on how much activity occurs in the module. The Extend-O-Save Overhead Tester is a GEM program with a menu bar. This gives you access to the Warp 9 Control Panel so that you may load different modules or change the speed settings. By changing the speed of modules which allow it, you can find a speed which does not affect your computer as much (if this is important to you). For instance, on a Mega ST4 running with a T-16 accelerator, the Mandala module uses a full 100% of the CPU time with any setting faster than the middle one. When a module consumes this much time, it even affects the ability of the test to get started. You may sometimes hear very long tones in the ascending sound. This is an indication that the module is using 100% of the CPU time (because the timing of the sound is dependent on the vertical blank interrupt). On the other hand, Mandala takes as little as 2% of the CPU time in its slowest setting. Other things will affect the results you get in this test program, including the type of computer, its clock speed, and the resident software you have installed. We've even found mysterious circumstances which will drastically affect the CPU overhead. The Clock module normally takes about 3-7% of the CPU time with its "Slow" setting. But after doing a number of activities during a computer session, it may report a much higher figure -- as much as 84%. The exact cause of this is unknown at the time, but it is not within Warp 9. We hope that the Extend-O-Save Overhead Tester will be of some use in choosing Extend-O-Save modules and speed settings which will achieve acceptable results on your system. Copyright stuff --------------- EOS_TEST.PRG, and this document are Copyright 1994 CodeHead Technologies. Extend-O-Save Overhead Tester is freeware and may be freely distributed as long as neither this document file nor the program file are altered in any way. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~