home *** CD-ROM | disk | FTP | other *** search
- Hello everybody!
-
- Like many others on this list I have recently switched to
- AMOS Pro thanks to the coverdisk of the Amiga Format
- magazine. I have also directly ordered the AMOS Pro
- compiler which I have received a few days ago. As the
- AP compiler manual stated that the compiler had been
- optimized compared to the AMOS Creator compiler, I have made
- a few simple benchmarks to test that. They are by no way
- extensive or very meaningful, but they might interest you anyway.
-
- The benchmarks have been tested in the following way:
-
- T=Timer
- For H=1 to 200000
- A=B*16
- Next H
- PRINT Timer-T
-
- The result is therefore the number of 1/50th seconds on
- my PAL Amiga 4000/40 25 MHz.
-
- The most interesting thing is probably that the AP interpreter
- is about 10% slower than the AMOS interpreter, probably due to
- the higher number of commands it has to check all the time.
- You'll also see that the AP compiler is indeed faster than
- the AMOS Creator compiler, but the difference is only really
- significant in a few cases that have been specially optimized
- (A=B*16 is translated with the assembler shift commands as
- shown in the AP Compiler manual and the A#=B#^4 statement
- has probably been simplified to A#=B#*B#*B#*B#).
-
- UNCOMPILED COMPILED CODE
- AMOS APro A Comp AP Comp
-
- A=B*16 114 124 30 13
- A=B*C 123 133 31 30
- A=B/C 134 141 45 39
- A#=B#*C# 126 148 36 34
- A#=B#^4 318 296 222 147
- A#=Sin(B#) 135 139 54 50
-
-
- By the way: I would be interested to know the correspnding times
- on other Amiga machines (A1200, A500, 68030 systems etc.)
- Maybe some of you can check that...
-
- Cheers,
-
- Chris
-
-