The new version is looking good; I haven't found any errors but it does seem to run noticeably slower (approx. 5% slower) than the previous version I was using (v0.8.2).
The speed is slower no matter which compiler options are specified for each version as long as the options are the same.
I personally find -Os to be faster than -O2 or -O3.
From what I have read, -Os is basically the same as -O2 except that -Os also reduces the code size a little more.
The program I've tested is mainly math and string comparison/manipulation in SCREEN 0 -- no loading of graphics files...
Has anyone else tried a benchmark program to test for speed differences between v0.8.2 and v3.1.0 ??
More Info : I'm using the 32 bit version of QB64pe on Windows Vista SP2.
Also,
The size of the .exe generated by v0.8.2 is 2165 KB.
The size of the .exe generated by v3.1.0 is 2243 KB.
The speed is slower no matter which compiler options are specified for each version as long as the options are the same.
I personally find -Os to be faster than -O2 or -O3.
From what I have read, -Os is basically the same as -O2 except that -Os also reduces the code size a little more.
The program I've tested is mainly math and string comparison/manipulation in SCREEN 0 -- no loading of graphics files...
Has anyone else tried a benchmark program to test for speed differences between v0.8.2 and v3.1.0 ??
More Info : I'm using the 32 bit version of QB64pe on Windows Vista SP2.
Also,
The size of the .exe generated by v0.8.2 is 2165 KB.
The size of the .exe generated by v3.1.0 is 2243 KB.