Hi James :)
I'd love to get to the bottom of this as I am doing the test across several machines and compilers.
So far, Blitzmax is indeed comparable - but it's very touchy about what you do! small changes result in drastic peformance differences.
The above code approaches the MSVC++ version in speed until you use globals. With most of us coming from Blitz3D, you can understand why we are concerned.
Here are the speed test results:
Blitz3D 2.58 seconds
Blitzmax 0.95 seconds
MSVC++ 0.71 seconds
0.7 vs 0.9 isn't bad!... but making it GLOBAL results in *worse than Blitz3D performance* - 2.6 !
I'd love to get to the bottom of this as I am doing the test across several machines and compilers.
So far, Blitzmax is indeed comparable - but it's very touchy about what you do! small changes result in drastic peformance differences.
The above code approaches the MSVC++ version in speed until you use globals. With most of us coming from Blitz3D, you can understand why we are concerned.
Here are the speed test results:
Blitz3D 2.58 seconds
Blitzmax 0.95 seconds
MSVC++ 0.71 seconds
0.7 vs 0.9 isn't bad!... but making it GLOBAL results in *worse than Blitz3D performance* - 2.6 !