Re: My thoughts on C# and gaming.
Miguel de Icaza <[email protected]>
| Newsgroups | gmane.games.devel.sweng |
|---|---|
| Organization | Novell, Inc. |
| Message-ID | <[email protected]> |
> The Great Language Game http://shootout.alioth.debian.org/ says that > compared to GNU g++ > Java 6 server is up to 3x slower > Java 6 steady state is up to 3x slower > Mono C# is up to 10x slower > LuaJIT is up to 24x slower > Mono F# is up to 48x slower > Java 6 -Xint interpreted mode is up to 69x slower > Lua is up to 148x slower Right, that uses the standard Mono installation, not the LLVM one which doubles Mono's speed. And also runs Mono in safe mode, if you make it unsafe, it doubles the speed again. On SciMark the results on my machine are as follows: Tests on Jan 29, 2009 Mono 2.6 Composite Score: 462.10 MFlops FFT : 287.74 - (1024) SOR : 824.82 - (100x100) Monte Carlo : 111.53 Sparse MatMult : 439.06 - (N=1000, nz=5000) LU : 647.37 - (100x100) Mono Trunk using LLVM Composite88 Score: 640.94 MFlops FFT : 496.85 - (1024) SOR : 849.50 - (100x100) Monte Carlo : 163.69 Sparse MatMult : 738.69 - (N=1000, nz=5000) LU : 955.99 - (100x100) Mono trunk with LLVM and no bounds checking: Composite Score: 805.14 MFlops FFT : 599.11 - (1024) SOR : 831.50 - (100x100) Monte Carlo : 165.99 Sparse MatMult : 1107.51 - (N=1000, nz=5000) LU : 1321.57 - (100x100) Java 1.6: Composite Score: 882.9892682796393 FFT (1024): 532.7552466699907 SOR (100x100): 1113.8468131532127 Monte Carlo : 353.08841271319847 Sparse matmult (N=1000, nz=5000): 777.1835376921547 LU (100x100): 1638.0723311696404 Miguel _______________________________________________ Sweng-Gamedev mailing list [email protected] http://lists.midnightryder.com/listinfo.cgi/sweng-gamedev-midnightryder.com