HOW_FAST.PRG

- Purpose of the program
  ----------------------

 With How-Fast it becomes possible for the first time (at least we 
 do not know of any similar program for the ATARI computer) to 
 subject mass storage devices such as hard disks, interchangeable 
 media drives, floppy disk drives, RAM-disks, networks etc. to a
 test that represents very well their working speed under normal 
 operating conditions.

 To do this, files are written and read via the operating system 
 functions, the time taken measured and the file size incremented.
 Finally a graph is drawn of the results.

 To obtain comparable results with this test, it is imperative that 
 the hard disk drives etc. are partitioned and filled identically 
 (a full 4Mb partition always gives poorer results than an empty 
 16 Mb partition). As a side-effect, the program shows up quite 
 well the loss of performance due to fragmentation of the data 
 after repeated deletion and re-use of disk space. Disk caches too 
 can be recognised without a doubt from the shape of the graphs 
 (this includes write-caches!). In addition the program reveals 
 timing problems of various SCSI host adapters without mercy.

 The results may also be saved permanently as a numerical file 
 for further graphical processing with, say, X-Act.

- Operation
  ---------

 How-Fast is really self-explanatory. The box at far right allows 
 you to  select the drive letter (A to P) of the drive or partition 
 you want to test. The 'Status' box below shows the amount of free 
 space on the selected drive.
 
 The initial and final sizes of the test files as well as the 
 number of test-steps to be run can be set conveniently with the 
 sliders; as you change any of them, the amount by which the size 
 of the test file will be incremented for each step is shown in the 
 field next to 'Block growth:'.
 
 The scaling of the Y-axis on the results graph can be set in two 
 ways: The default is 'Automatic', which before the start of the 
 test proper measures the speed of the medium with a 100 KB file 
 and then uses twice the resulting value as a guide for the 
 vertical scaling.

 The automation can be switched off by clicking on the 'Y-axis...'
 button and after deselecting 'Automatic' in the new dialog you can 
 enter a maximum value for the Y-axis (up to 9999 KB/s).

 In the main dialog you can also input a description of the medium 
 (e.g. Quantum LPS120 16MB partition). This text will be displayed 
 as the title of the GEM-window containing the graph (useful for 
 hard-copies/screen-shots) and also stored in the data file.
 
 To start the test, click on 'Run' or use the [Return] key. A 
 growing graphical plot will show the speed for various block sizes 
 for both reading (solid line/blue) and writing (broken line/red) 
 in monochrome/16-colour-and up resolutions respectively. If you 
 wish to interrupt a test before it is finished (say because the 
 plots are running off the top of the graph), hold down the [Esc] 
 key. You will be asked whether you wish to continue with the test; 
 if so, click on 'Go on', otherwise 'OK' (or [Return]) stops 
 further testing and returns you to the main dialog where you can 
 make fresh choices if desired.
 
 Clicking on the finished graph brings up a dialog asking whether 
 you wish to save the result data. Clicking on 'Yes' brings up the 
 file selector, allowing you to choose a suitable path and 
 filename.
 
 As may be expected, the 'Read only' button performs tests without 
 writing anything to the disk - useful for speed measurement of 
 CD-ROMs, or disks that have little free space on them. Clicking on 
 the button brings up the file selector, where you can select any 
 desired file with which the read performance is then measured. For 
 this you should use only files larger than 100 KB ....

 A previously saved data file may be read in with the 'Load...' 
 button, which will create an identical graph to the original.
 
 The 'i' button gives information about the program, its author and 
 distributor.
 
 The 'Exit' button quits the program and returns you to the 
 desktop.

- Problems/Use at your own risk?
  -----------------------------

 As How-Fast uses only GEMDOS operating system calls for reading 
 and writing to the mass storage devices, undesired loss of data is 
 very unlikely. The test file is always written to the root 
 directory of the selected dfrive under the name 'HOW_FAST.TST', 
 and deleted immediately afterwards.

 Despite taking all possible precautions, we have to state here 
 explicitly that we can accept no liability for any damage or 
 loss of data reulting from the use of How-Fast.

Questions to:
Editor, ST-Computer
Industriestrae 26
W-6236 Eschborn
z.Hd. Christian Mller


History
----------------------------------------------------------------

V0.94    28.05.1992

- With the 'Y-axis' button one can switch off the automatic 
  Y-axis scaling and input a maximum value manually.

- Only one grid will be created in the graph.

- In addition the step size of the Y-axis now adapts automatically
  to the measured maximum value.

- The measurements can be interrupted by holding down the [Esc] 
  key (it make take a short while before the program reacts).

----------------------------------------------------------------

V0.96     ??.07.1992

- From now on colour screens are also supported. For this the 
  program enquires via the VDI the number of bitplanes. If it is 4 
  or more than coloured lines will be drawn. This assumes the 
  standard palette is in use.

----------------------------------------------------------------

V1.00    15.07.1992

- This is the first version that may be used officially.

- Design and handling slightly revised.

- The ST-Computer logo has been incorporated.

- Small changes to the graphics routines.

- Bug in graphic output during loading removed.

V1.03     29.10.1992

- The manual Y-axis scaling now also works during loading. This 
  means that you can alter the display scale subsequently.

V1.11    09.08.1993

- CAREFUL! STILL A BETA!
  A new button 'Read only' has been added. It serves mainly for the 
  testing of CD-ROMs. With it one can select any desired file with 
  which then read performance is measured. For this you should use 
  only files larger than 100 KB ...

V1.12	  10.08.1993
- CAREFUL! STILL A BETA!
  Adapted for MultiTOS and MagiC!. Should now run OK ...
  
----------------------------------------------------------------

Bugs:

- The maximum size of a test file depends on the amount of space 
  free on the disk or partition, but also on the available RAM 
  memory. However this is not tested for, so take care if you are 
  short of memory!

- Crash when accessing interchangeable media, including floppies, 
  that are not ready (no medium loaded or not yet running).

Have fun!.....Chris

------------------------------------------------------------------

Translation by Peter West of DDP. Email: xlator@cix.co.uk

