M-HT
Very Active Member
Hi,
I updated the dynrec in Dosbox - you can get the patch in the dosbox forums.
I compared the speed of my version (latest svn + my patch) against the old dosbox pnd with following results...
Even without my patch applied, my version was 20% faster than dosbox pnd.
With my patch applied, my version is 80% faster than dosbox pnd - using doom2 timedemo (100% faster using benchmark without graphics output).
(I also tested version by urjaman called winbox, but the speed was roughly the same as dosbox pnd.)
My version was compiled with crosscompiler by Ivanovic.
First I configured dosbox using this script:
Next I edited config.h and changed the following definitions:
Then I compiled dosbox.
I updated the dynrec in Dosbox - you can get the patch in the dosbox forums.
I compared the speed of my version (latest svn + my patch) against the old dosbox pnd with following results...
Even without my patch applied, my version was 20% faster than dosbox pnd.
With my patch applied, my version is 80% faster than dosbox pnd - using doom2 timedemo (100% faster using benchmark without graphics output).
(I also tested version by urjaman called winbox, but the speed was roughly the same as dosbox pnd.)
My version was compiled with crosscompiler by Ivanovic.
First I configured dosbox using this script:
Code:
#!/bin/sh
HOST=arm-none-linux-gnueabi
TARGET=arm-none-linux-gnueabi
export HOST
export TARGET
PATH=$PATH:$PNDSDK/bin
export PATH
./configure --prefix=$PNDSDK --with-sdl-exec-prefix=$PNDSDK/usr --host=$HOST --target=$TARGET --build=`uname -m` CPPFLAGS="-I$PNDSDK/include" CXXFLAGS="-O3 -fmerge-all-constants -ffast-math -pipe -march=armv7-a -mcpu=cortex-a8 -mtune=cortex-a8 -mfpu=vfpv3 -mfloat-abi=softfp -mthumb-interwork" LDFLAGS="-s -L$PNDSDK/lib" --enable-core-inline
Next I edited config.h and changed the following definitions:
Code:
#define C_DYNREC 1
#define C_TARGETCPU ARMV7LE
#define C_UNALIGNED_MEMORY 1
Then I compiled dosbox.
Last edited by a moderator: