The place for coders and contributers to discuss about the code base of the DragonScript language, VM and compiler.
- Forum Administrator
- Posts: 609
- Joined: Fri Jul 30, 2004 4:30 pm
- Location: Switzerland
thanOS wrote:This comparative test looks very interesting!
It is most probable that the java compiler and VM takes advantage of the features of the underlying architecture to gain some speed.
responsible for that is JIT, not some VM trick. what it does is just-in-time compile code to native machine code (like math operations) hence getting up there with C. in terms of math JIT can do a lot but really interesting would be a test of OOP features (namely many function calls, instantiations or late-binding calls for languages making there a difference).
Why did'nt you use a "simple" math operator like that of multiplication and went on doing floatvalue & 2048 / 8.0f (???). What is the equivalent of these logical operators on a float? (In terms of actual operations)
on my old model-editor (which is now R.I.P.) i wrote a simple import plugin for unreal-models. unreal uses some packing to store 3-float vector in 1-int value (hence 96bit -> 32bit). for fun i then took this un-packing code and put it into DS making a speed test. marschall then took it further and put it into other languages hence we ended up where we did
Oh to be a dragon, of silkworm size or immense...