home *** CD-ROM | disk | FTP | other *** search
-
- Documentation? Ok, here's some... I wish that someone would write a
- program which will grade the quality of the ram boards that are being used.
- In my beginning attempts at writing in the Lattice-C language and the Amiga
- PC, I've programmed this test. The results may mean something but who
- knows and I only can hope. I need some explanation on what I'm doing. I
- need to find the access speeds of Fast and Chip ram. Really. How can a
- pocket watch clock the speed of light? How? Use a computer? WORDS OF
- CAUTION! The program sets the priority of the current task to 127,
- Forbid()s, and Disable()s the machine during testing. DON'T REMOVE OR
- INSERT DISKS DURING THE TEST! Afterwards, things SHOULD return to normal
- again. The timings returned are how long it took for a 1K long ram copy
- program to copy itself to another 1K area of memory. Inorder to get a more
- accurate reading of the time, the copying is repeated 15,000 or so times
- and reported in milliseconds from the V_BLANK timer. This timer is 116.7
- milliseconds accurate. TimeRam has two self-copying routines; One routine
- resides in ChipRam and the other resides in FastRam. Also, two 1K areas of
- ram are allocated as targets; One target lives in ChipRam and the other in
- FastRam. The program proceeds as follows:
-
- 1) ChipRam copier copies itself to ChipRam and is timed.
- 2) FastRam copier copies itself to FastRam and is timed.
- 3) ChipRam copier copies itself to FastRam and is timed.
- 4) FastRam copier copies itself to ChipRam and is timed.
-
- None of this information tells much about the QUALITY of the ramboards out
- there for sale; Unless, of course, we spread the word to others of our
- ramboards and these figures. It may be of some use to programmers to
- understand how to speed up their programs though. Who knows? I had fun
- writing the program, anyway. Tell me about it. Bruce's repair and testing
- (415) 525-6973 Berkeley, California, USA, Third Planet Ol' Sol, 94710-1706.
-
- Special thanks to Bryce Nesbitt for getting me to do what I wouldn't have
- done 'cause I didn't know how to do it and didn't know it existed anyway.
- Ya Know? I didn't.
-
- Here're the test results of an internal ramboard from a company with morals
- not to sell, but honest enough to let me test and they improve. More power
- to 'em.
-
- ----------------------------------
- chipram to chipram 11.883 seconds
- fastram to fastram 12.833 seconds
- chipram to fastram 13.333 seconds
- fastram to chipram 12.166 seconds
-
- With HAM displayed on the monitor:
-
- chipram to chipram 14.550 seconds
- fastram to fastram 12.833 seconds
- chipram to fastram 13.516 seconds
- fastram to chipram 12.950 seconds
- ----------------------------------
-
- TimeRam is absolutely free as long as you own an Amiga PC. If you don't
- have an Amiga, you can treasure it anyway.
-
- Bruce T.
-
-
- You may also send test results to:
-
- bryce@cogsci.Berkeley.EDU
- ucbvax!cogsci!bryce
-
- -or, if that does not work, try-
- bryce@hoser.Berkeley.EDU
- ucbvax!hoser!bryce
-
- When test results are in for all the memory boards, the findings will
- be summaried to comp.sys.amiga.
-
-