diff --git a/README.md b/README.md
index 7fe87ca71..982ba4056 100644
--- a/README.md
+++ b/README.md
@@ -3,7 +3,6 @@
 [SRB2Kart](https://srb2.org/mods/) is a kart racing mod based on the 3D Sonic the Hedgehog fangame [Sonic Robo Blast 2](https://srb2.org/), based on a modified version of [Doom Legacy](http://doomlegacy.sourceforge.net/).
 
 ## Dependencies
-- NASM (x86 builds only)
 - SDL2 (Linux/OS X only)
 - SDL2-Mixer (Linux/OS X only)
 - libupnp (Linux/OS X only)
diff --git a/SRB2.cbp b/SRB2.cbp
index acdc61c7f..b4cf543db 100644
--- a/SRB2.cbp
+++ b/SRB2.cbp
@@ -1996,24 +1996,6 @@ HW3SOUND for 3D hardware sound  support
 			<Option compilerVar="CC" />
 		</Unit>
 		<Unit filename="src/v_video.h" />
-		<Unit filename="src/vid_copy.s">
-			<Option compilerVar="CC" />
-			<Option compiler="avrgcc" use="1" buildCommand="$compiler $options -x assembler-with-cpp -c $file -o $object" />
-			<Option compiler="gnu_gcc_compiler_for_mingw32" use="1" buildCommand="$compiler $options -x assembler-with-cpp -c $file -o $object" />
-			<Option compiler="gnu_gcc_compiler_for_mingw64" use="1" buildCommand="$compiler $options -x assembler-with-cpp -c $file -o $object" />
-			<Option compiler="armelfgcc" use="1" buildCommand="$compiler $options -x assembler-with-cpp -c $file -o $object" />
-			<Option compiler="tricoregcc" use="1" buildCommand="$compiler $options -x assembler-with-cpp -c $file -o $object" />
-			<Option compiler="ppcgcc" use="1" buildCommand="$compiler $options -x assembler-with-cpp -c $file -o $object" />
-			<Option compiler="gcc" use="1" buildCommand="$compiler $options -x assembler-with-cpp -c $file -o $object" />
-			<Option target="Debug Native/SDL" />
-			<Option target="Release Native/SDL" />
-			<Option target="Debug Linux/SDL" />
-			<Option target="Release Linux/SDL" />
-			<Option target="Debug Mingw/SDL" />
-			<Option target="Release Mingw/SDL" />
-			<Option target="Debug Mingw/DirectX" />
-			<Option target="Release Mingw/DirectX" />
-		</Unit>
 		<Unit filename="src/w_wad.c">
 			<Option compilerVar="CC" />
 		</Unit>
diff --git a/SRB2_common.props b/SRB2_common.props
index 0f80ceb17..6a0d53484 100644
--- a/SRB2_common.props
+++ b/SRB2_common.props
@@ -25,9 +25,6 @@
     </Link>
   </ItemDefinitionGroup>
   <ItemDefinitionGroup Condition="'$(PlatformTarget)'=='x86'">
-    <ClCompile>
-      <PreprocessorDefinitions>USEASM;%(PreprocessorDefinitions)</PreprocessorDefinitions>
-    </ClCompile>
     <Link>
       <ImageHasSafeExceptionHandlers>false</ImageHasSafeExceptionHandlers>
     </Link>
diff --git a/Srb2.dev b/Srb2.dev
index 21683e7c3..8bd36cf49 100644
--- a/Srb2.dev
+++ b/Srb2.dev
@@ -5,7 +5,7 @@ Ver=3
 IsCpp=0
 Type=0
 UnitCount=279
-Folders=A_Asm,B_Bot,BLUA,D_Doom,F_Frame,G_Game,H_Hud,Hw_Hardware,Hw_Hardware/r_opengl,I_Interface,I_Interface/Dummy,I_Interface/SDL,I_Interface/Win32,LUA,M_Misc,P_Play,R_Rend,S_Sounds,W_Wad
+Folders=B_Bot,BLUA,D_Doom,F_Frame,G_Game,H_Hud,Hw_Hardware,Hw_Hardware/r_opengl,I_Interface,I_Interface/Dummy,I_Interface/SDL,I_Interface/Win32,LUA,M_Misc,P_Play,R_Rend,S_Sounds,W_Wad
 CommandLine=
 CompilerSettings=00000000000100000111e1
 PchHead=-1
@@ -1473,36 +1473,6 @@ Priority=1000
 OverrideBuildCmd=0
 BuildCmd=
 
-[Unit149]
-FileName=src\tmap.nas
-Folder=A_Asm
-Compile=0
-CompileCpp=0
-Link=0
-Priority=1000
-OverrideBuildCmd=1
-BuildCmd=nasm.exe -g -o $@ -f win32 src/tmap.nas
-
-[Unit150]
-FileName=src\asm_defs.inc
-Folder=A_Asm
-Compile=0
-CompileCpp=0
-Link=0
-Priority=1000
-OverrideBuildCmd=0
-BuildCmd=
-
-[Unit151]
-FileName=src\vid_copy.s
-Folder=A_Asm
-Compile=1
-CompileCpp=0
-Link=1
-Priority=1000
-OverrideBuildCmd=1
-BuildCmd=$(CC) $(CFLAGS) -x assembler-with-cpp -c src/vid_copy.s -o $@
-
 [Unit152]
 FileName=src\y_inter.h
 Folder=H_Hud
@@ -1543,26 +1513,6 @@ Priority=1000
 OverrideBuildCmd=0
 BuildCmd=
 
-[Unit156]
-FileName=src\p5prof.h
-Folder=A_Asm
-Compile=1
-CompileCpp=0
-Link=1
-Priority=1000
-OverrideBuildCmd=0
-BuildCmd=
-
-[Unit157]
-FileName=src\tmap_mmx.nas
-Folder=A_Asm
-Compile=0
-CompileCpp=0
-Link=0
-Priority=1000
-OverrideBuildCmd=1
-BuildCmd=nasm.exe -g -o $@ -f win32 src/tmap_mmx.nas
-
 [Unit159]
 FileName=src\lzf.h
 Folder=W_Wad
diff --git a/appveyor.yml b/appveyor.yml
index 474a2adb3..f4428a090 100644
--- a/appveyor.yml
+++ b/appveyor.yml
@@ -7,8 +7,6 @@ environment:
  # c:\mingw-w64 i686 has gcc 6.3.0, so use c:\msys64 7.3.0 instead
  MINGW_SDK: c:\msys64\mingw32
  CFLAGS: -Wno-implicit-fallthrough
- NASM_ZIP: nasm-2.12.01
- NASM_URL: http://www.nasm.us/pub/nasm/releasebuilds/2.12.01/win64/nasm-2.12.01-win64.zip
  UPX_ZIP: upx391w
  UPX_URL: http://upx.sourceforge.net/download/upx391w.zip
  CCACHE_EXE: ccache.exe
@@ -40,17 +38,12 @@ environment:
  ASSET_CLEAN: 0
 
 cache:
-- nasm-2.12.01.zip
 - upx391w.zip
 - ccache.exe
 - C:\Users\appveyor\.ccache
 - C:\Users\appveyor\srb2_cache
 
 install:
-- if not exist "%NASM_ZIP%.zip" appveyor DownloadFile "%NASM_URL%" -FileName "%NASM_ZIP%.zip"
-- 7z x -y "%NASM_ZIP%.zip" -o%TMP% >null
-- robocopy /S /xx /ns /nc /nfl /ndl /np /njh /njs "%TMP%\%NASM_ZIP%" "%MINGW_SDK%\bin" nasm.exe || exit 0
-
 - if not exist "%UPX_ZIP%.zip" appveyor DownloadFile "%UPX_URL%" -FileName "%UPX_ZIP%.zip"
 - 7z x -y "%UPX_ZIP%.zip" -o%TMP% >null
 - robocopy /S /xx /ns /nc /nfl /ndl /np /njh /njs "%TMP%\%UPX_ZIP%" "%MINGW_SDK%\bin" upx.exe || exit 0
@@ -65,7 +58,6 @@ configuration:
 before_build:
 - set "Path=%MINGW_SDK%\bin;%Path%"
 - mingw32-make --version
-- nasm -v
 - if not [%NOUPX%] == [1] ( upx -V )
 - ccache -V
 - ccache -s
diff --git a/cmake/Modules/CMakeASM_YASMInformation.cmake b/cmake/Modules/CMakeASM_YASMInformation.cmake
deleted file mode 100644
index 176518085..000000000
--- a/cmake/Modules/CMakeASM_YASMInformation.cmake
+++ /dev/null
@@ -1,46 +0,0 @@
-
-#=============================================================================
-# Copyright 2010 Kitware, Inc.
-#
-# Distributed under the OSI-approved BSD License (the "License");
-# see accompanying file Copyright.txt for details.
-#
-# This software is distributed WITHOUT ANY WARRANTY; without even the
-# implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
-# See the License for more information.
-#=============================================================================
-# (To distribute this file outside of CMake, substitute the full
-#  License text for the above reference.)
-
-# support for the yasm assembler
-
-set(CMAKE_ASM_YASM_SOURCE_FILE_EXTENSIONS nasm yasm asm)
-
-if(NOT CMAKE_ASM_YASM_OBJECT_FORMAT)
-  if(WIN32)
-    if(CMAKE_C_SIZEOF_DATA_PTR EQUAL 8)
-      set(CMAKE_ASM_YASM_OBJECT_FORMAT win64)
-    else()
-      set(CMAKE_ASM_YASM_OBJECT_FORMAT win32)
-    endif()
-  elseif(APPLE)
-    if(CMAKE_C_SIZEOF_DATA_PTR EQUAL 8)
-      set(CMAKE_ASM_YASM_OBJECT_FORMAT macho64)
-    else()
-      set(CMAKE_ASM_YASM_OBJECT_FORMAT macho)
-    endif()
-  else()
-    if(CMAKE_C_SIZEOF_DATA_PTR EQUAL 8)
-      set(CMAKE_ASM_YASM_OBJECT_FORMAT elf64)
-    else()
-      set(CMAKE_ASM_YASM_OBJECT_FORMAT elf)
-    endif()
-  endif()
-endif()
-
-set(CMAKE_ASM_YASM_COMPILE_OBJECT "<CMAKE_ASM_YASM_COMPILER> <FLAGS> -f ${CMAKE_ASM_YASM_OBJECT_FORMAT} -o <OBJECT> <SOURCE>")
-
-# Load the generic ASMInformation file:
-set(ASM_DIALECT "_YASM")
-include(CMakeASMInformation)
-set(ASM_DIALECT)
diff --git a/cmake/Modules/CMakeDetermineASM_YASMCompiler.cmake b/cmake/Modules/CMakeDetermineASM_YASMCompiler.cmake
deleted file mode 100644
index a5e7c9e58..000000000
--- a/cmake/Modules/CMakeDetermineASM_YASMCompiler.cmake
+++ /dev/null
@@ -1,27 +0,0 @@
-
-#=============================================================================
-# Copyright 2010 Kitware, Inc.
-#
-# Distributed under the OSI-approved BSD License (the "License");
-# see accompanying file Copyright.txt for details.
-#
-# This software is distributed WITHOUT ANY WARRANTY; without even the
-# implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
-# See the License for more information.
-#=============================================================================
-# (To distribute this file outside of CMake, substitute the full
-#  License text for the above reference.)
-
-# Find the nasm assembler. yasm (http://www.tortall.net/projects/yasm/) is nasm compatible
-
-set(CMAKE_ASM_YASM_COMPILER_LIST nasm yasm)
-
-if(NOT CMAKE_ASM_YASM_COMPILER)
-  find_program(CMAKE_ASM_YASM_COMPILER yasm
-    "$ENV{ProgramFiles}/YASM")
-endif()
-
-# Load the generic DetermineASM compiler file with the DIALECT set properly:
-set(ASM_DIALECT "_YASM")
-include(CMakeDetermineASMCompiler)
-set(ASM_DIALECT)
diff --git a/cmake/Modules/CMakeTestASM_YASMCompiler.cmake b/cmake/Modules/CMakeTestASM_YASMCompiler.cmake
deleted file mode 100644
index 745f7125c..000000000
--- a/cmake/Modules/CMakeTestASM_YASMCompiler.cmake
+++ /dev/null
@@ -1,23 +0,0 @@
-
-#=============================================================================
-# Copyright 2010 Kitware, Inc.
-#
-# Distributed under the OSI-approved BSD License (the "License");
-# see accompanying file Copyright.txt for details.
-#
-# This software is distributed WITHOUT ANY WARRANTY; without even the
-# implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
-# See the License for more information.
-#=============================================================================
-# (To distribute this file outside of CMake, substitute the full
-#  License text for the above reference.)
-
-# This file is used by EnableLanguage in cmGlobalGenerator to
-# determine that the selected ASM_NASM "compiler" works.
-# For assembler this can only check whether the compiler has been found,
-# because otherwise there would have to be a separate assembler source file
-# for each assembler on every architecture.
-
-set(ASM_DIALECT "_YASM")
-include(CMakeTestASMCompiler)
-set(ASM_DIALECT)
diff --git a/debian-template/README.source b/debian-template/README.source
index f63a42cac..acd47d2f1 100644
--- a/debian-template/README.source
+++ b/debian-template/README.source
@@ -9,7 +9,6 @@ GNU/Linux
 Dependencies:
   SDL 1.2.7 or better (from libsdl.org)
   SDL_Mixer 1.2.2(.7 for file-less music playback) (from libsdl.org)
-  Nasm (use NOASM=1 if you don't have it or have an non-i386 system, I think)
   libPNG 1.2.7
   Zlib 1.2.3
   The Xiph.org libogg and libvorbis libraries
diff --git a/src/Android.mk b/src/Android.mk
index a461da224..035d48887 100644
--- a/src/Android.mk
+++ b/src/Android.mk
@@ -76,7 +76,7 @@ LOCAL_SRC_FILES :=      am_map.c \
                         android/i_system.c \
                         android/i_video.c
 
-LOCAL_CFLAGS += -DPLATFORM_ANDROID -DNONX86 -DLINUX -DDEBUGMODE -DNOASM -DNOPIX -DUNIXCOMMON -DNOTERMIOS
+LOCAL_CFLAGS += -DPLATFORM_ANDROID -DNONX86 -DLINUX -DDEBUGMODE -DNOPIX -DUNIXCOMMON -DNOTERMIOS
 
 LOCAL_MODULE := libsrb2
 
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index 74bbf1013..c7af7bf91 100644
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -11,15 +11,7 @@ target_compile_features(SRB2SDL2 PRIVATE c_std_11 cxx_std_17)
 target_sourcefile(c)
 target_sources(SRB2SDL2 PRIVATE comptime.c md5.c config.h.in)
 
-set(SRB2_ASM_SOURCES vid_copy.s)
-
-set(SRB2_NASM_SOURCES tmap_mmx.nas tmap.nas)
-
 ### Configuration
-set(SRB2_CONFIG_USEASM OFF CACHE BOOL
-	"Enable NASM tmap implementation for software mode speedup.")
-set(SRB2_CONFIG_YASM OFF CACHE BOOL
-	"Use YASM in place of NASM.")
 set(SRB2_CONFIG_DEV_BUILD OFF CACHE BOOL
 	"Compile a development build of SRB2Kart.")
 	
@@ -176,33 +168,6 @@ if(${CMAKE_SYSTEM} MATCHES Windows AND ${CMAKE_C_COMPILER_ID} MATCHES "GNU" AND
 	target_include_directories(SRB2SDL2 PRIVATE "${CMAKE_CURRENT_SOURCE_DIR}/../thirdparty/drmingw/include")
 endif()
 
-if(${SRB2_CONFIG_USEASM})
-	#SRB2_ASM_FLAGS can be used to pass flags to either nasm or yasm.
-	if("${CMAKE_SYSTEM_NAME}" MATCHES "Linux")
-		set(SRB2_ASM_FLAGS "-DLINUX ${SRB2_ASM_FLAGS}")
-	endif()
-
-	if(${SRB2_CONFIG_YASM})
-		set(CMAKE_ASM_YASM_SOURCE_FILE_EXTENSIONS ${CMAKE_ASM_YASM_SOURCE_FILE_EXTENSIONS} nas)
-		set(CMAKE_ASM_YASM_FLAGS "${SRB2_ASM_FLAGS}" CACHE STRING "Flags used by the assembler during all build types.")
-		enable_language(ASM_YASM)
-	else()
-		set(CMAKE_ASM_NASM_SOURCE_FILE_EXTENSIONS ${CMAKE_ASM_NASM_SOURCE_FILE_EXTENSIONS} nas)
-		set(CMAKE_ASM_NASM_FLAGS "${SRB2_ASM_FLAGS}" CACHE STRING "Flags used by the assembler during all build types.")
-		enable_language(ASM_NASM)
-	endif()
-
-	set(SRB2_USEASM ON)
-	target_compile_definitions(SRB2SDL2 PRIVATE -DUSEASM)
-	target_compile_options(SRB2SDL2 PRIVATE -msse3 -mfpmath=sse)
-
-	target_sources(SRB2SDL2 PRIVATE ${SRB2_ASM_SOURCES}
-		${SRB2_NASM_SOURCES})
-else()
-	set(SRB2_USEASM OFF)
-	target_compile_definitions(SRB2SDL2 PRIVATE -DNONX86 -DNORUSEASM)
-endif()
-
 # Targets
 
 # If using CCACHE, then force it.
diff --git a/src/Makefile b/src/Makefile
index 51e85504b..dd309ba41 100644
--- a/src/Makefile
+++ b/src/Makefile
@@ -47,8 +47,6 @@
 # HAVE_MINIUPNPC=1 - Enable automated port forwarding.
 #                    Already enabled by default for 32-bit
 #                    Windows.
-# NOASM=1 - Disable hand optimized assembly code for the
-#           Software renderer.
 # NOPNG=1 - Disable PNG graphics support. (TODO: double
 #           check netplay compatible.)
 # NOCURL=1 - Disable libcurl--HTTP capability.
@@ -87,7 +85,6 @@
 #        executable.
 # WINDOWSHELL=1 - Use Windows commands.
 # PREFIX= - Prefix to many commands, for cross compiling.
-# YASM=1 - Use Yasm instead of NASM assembler.
 # STABS=1 - ?
 # ECHO=1 - Print out each command in the build process.
 # NOECHOFILENAMES=1 - Don't print out each that is being
@@ -147,22 +144,6 @@ OBJCOPY?=$(call Prefix,objcopy)
 OBJDUMP?=$(call Prefix,objdump)
 WINDRES?=$(call Prefix,windres)
 
-ifdef YASM
-NASM?=yasm
-else
-NASM?=nasm
-endif
-
-ifdef YASM
-ifdef STABS
-NASMOPTS?=-g stabs
-else
-NASMOPTS?=-g dwarf2
-endif
-else
-NASMOPTS?=-g
-endif
-
 GZIP?=gzip
 GZIP_OPTS?=-9 -f -n
 ifdef WINDOWSHELL
@@ -186,8 +167,6 @@ makedir:=../make
 opts:=-DCOMPVERSION -g
 libs:=
 
-nasm_format:=
-
 # This is a list of variables names, of which if defined,
 # also defines the name as a macro to the compiler.
 passthru_opts:=
@@ -337,7 +316,6 @@ endif
 
 LD:=$(CC)
 cc:=$(cc) $(opts)
-nasm=$(NASM) $(NASMOPTS) -f $(nasm_format)
 ifdef UPX
 upx=$(UPX) $(UPX_OPTS)
 endif
@@ -415,7 +393,6 @@ $(objdir)/%.$(1) : %.$(2) | $$$$(@D)/
 endef
 
 $(eval $(call _recipe,o,c,$(cc) -c -o $$@ $$<))
-$(eval $(call _recipe,o,nas,$(nasm) -o $$@ $$<))
 $(eval $(call _recipe,o,cpp,$(cxx) -c -o $$@ $$<))
 $(eval $(call _recipe,o,s,$(cc) $(asflags) -c -o $$@ $$<))
 $(eval $(call _recipe,res,rc,$(windres) -i $$< -o $$@))
diff --git a/src/Makefile.d/features.mk b/src/Makefile.d/features.mk
index 6492c3f77..6b26ebe53 100644
--- a/src/Makefile.d/features.mk
+++ b/src/Makefile.d/features.mk
@@ -1,82 +1,75 @@
-#
-# Makefile for feature flags.
-#
-
-passthru_opts+=\
-	NO_IPV6 NOHW NOMD5 NOPOSTPROCESSING\
-	MOBJCONSISTANCY PACKETDROP ZDEBUG\
-	HAVE_MINIUPNPC\
-	HAVE_DISCORDRPC DEVELOP
-
-# build with debugging information
-ifdef DEBUGMODE
-MOBJCONSISTANCY=1
-PACKETDROP=1
-opts+=-DPARANOIA -DRANGECHECK
-endif
-
-ifndef NOHW
-opts+=-DHWRENDER
-sources+=$(call List,hardware/Sourcefile)
-endif
-
-ifndef NOASM
-ifndef NONX86
-sources+=tmap.nas tmap_mmx.nas
-opts+=-DUSEASM
-endif
-endif
-
-ifndef NOMD5
-sources+=md5.c
-endif
-
-ifndef NOZLIB
-ifndef NOPNG
-ifdef PNG_PKGCONFIG
-$(eval $(call Use_pkg_config,PNG_PKGCONFIG))
-else
-PNG_CONFIG?=$(call Prefix,libpng-config)
-$(eval $(call Configure,PNG,$(PNG_CONFIG) \
-	$(if $(PNG_STATIC),--static),,--ldflags))
-endif
-ifdef LINUX
-opts+=-D_LARGEFILE64_SOURCE
-endif
-opts+=-DHAVE_PNG
-sources+=apng.c
-endif
-endif
-
-ifndef NOCURL
-CURLCONFIG?=curl-config
-$(eval $(call Configure,CURL,$(CURLCONFIG)))
-opts+=-DHAVE_CURL
-endif
-
-ifdef HAVE_MINIUPNPC
-libs+=-lminiupnpc
-endif
-
-ifdef HAVE_DISCORDRPC
-$(eval $(call Propogate_flags,DISCORDRPC))
-libs+=-ldiscord-rpc
-opts+=-DUSE_STUN
-sources+=discord.c stun.c
-endif
-
-# (Valgrind is a memory debugger.)
-ifdef VALGRIND
-VALGRIND_PKGCONFIG?=valgrind
-$(eval $(call Use_pkg_config,VALGRIND))
-ZDEBUG=1
-opts+=-DHAVE_VALGRIND
-endif
-
-default_packages:=\
-	GME/libgme/LIBGME\
-	OPENMPT/libopenmpt/LIBOPENMPT\
-	ZLIB/zlib\
-
-$(foreach p,$(default_packages),\
-	$(eval $(call Check_pkg_config,$(p))))
+#
+# Makefile for feature flags.
+#
+
+passthru_opts+=\
+	NO_IPV6 NOHW NOMD5 NOPOSTPROCESSING\
+	MOBJCONSISTANCY PACKETDROP ZDEBUG\
+	HAVE_MINIUPNPC\
+	HAVE_DISCORDRPC DEVELOP
+
+# build with debugging information
+ifdef DEBUGMODE
+MOBJCONSISTANCY=1
+PACKETDROP=1
+opts+=-DPARANOIA -DRANGECHECK
+endif
+
+ifndef NOHW
+opts+=-DHWRENDER
+sources+=$(call List,hardware/Sourcefile)
+endif
+
+ifndef NOMD5
+sources+=md5.c
+endif
+
+ifndef NOZLIB
+ifndef NOPNG
+ifdef PNG_PKGCONFIG
+$(eval $(call Use_pkg_config,PNG_PKGCONFIG))
+else
+PNG_CONFIG?=$(call Prefix,libpng-config)
+$(eval $(call Configure,PNG,$(PNG_CONFIG) \
+	$(if $(PNG_STATIC),--static),,--ldflags))
+endif
+ifdef LINUX
+opts+=-D_LARGEFILE64_SOURCE
+endif
+opts+=-DHAVE_PNG
+sources+=apng.c
+endif
+endif
+
+ifndef NOCURL
+CURLCONFIG?=curl-config
+$(eval $(call Configure,CURL,$(CURLCONFIG)))
+opts+=-DHAVE_CURL
+endif
+
+ifdef HAVE_MINIUPNPC
+libs+=-lminiupnpc
+endif
+
+ifdef HAVE_DISCORDRPC
+$(eval $(call Propogate_flags,DISCORDRPC))
+libs+=-ldiscord-rpc
+opts+=-DUSE_STUN
+sources+=discord.c stun.c
+endif
+
+# (Valgrind is a memory debugger.)
+ifdef VALGRIND
+VALGRIND_PKGCONFIG?=valgrind
+$(eval $(call Use_pkg_config,VALGRIND))
+ZDEBUG=1
+opts+=-DHAVE_VALGRIND
+endif
+
+default_packages:=\
+	GME/libgme/LIBGME\
+	OPENMPT/libopenmpt/LIBOPENMPT\
+	ZLIB/zlib\
+
+$(foreach p,$(default_packages),\
+	$(eval $(call Check_pkg_config,$(p))))
diff --git a/src/Makefile.d/nix.mk b/src/Makefile.d/nix.mk
index 97bcf018e..5f7d1b6a3 100644
--- a/src/Makefile.d/nix.mk
+++ b/src/Makefile.d/nix.mk
@@ -5,10 +5,6 @@
 opts+=-DUNIXCOMMON -DLUA_USE_POSIX
 libs+=-lm
 
-ifndef nasm_format
-nasm_format:=elf -DLINUX
-endif
-
 ifndef NOHW
 opts+=-I/usr/X11R6/include
 libs+=-L/usr/X11R6/lib
@@ -31,7 +27,6 @@ endif
 # FIXME: UNTESTED
 #ifdef SOLARIS
 #NOIPX=1
-#NOASM=1
 #opts+=-I/usr/local/include -I/opt/sfw/include \
 #		-DSOLARIS -DINADDR_NONE=INADDR_ANY -DBSD_COMP
 #libs+=-L/opt/sfw/lib -lsocket -lnsl
diff --git a/src/Makefile.d/platform.mk b/src/Makefile.d/platform.mk
index fad4be191..b4e2b8aa1 100644
--- a/src/Makefile.d/platform.mk
+++ b/src/Makefile.d/platform.mk
@@ -39,7 +39,6 @@ else ifdef SOLARIS # FIXME: UNTESTED
 UNIX=1
 platform=solaris
 else ifdef CYGWIN32 # FIXME: UNTESTED
-nasm_format=win32
 platform=cygwin
 else ifdef MINGW
 ifdef MINGW64
diff --git a/src/Makefile.d/sdl.mk b/src/Makefile.d/sdl.mk
index 99ca624e6..a1bfa3303 100644
--- a/src/Makefile.d/sdl.mk
+++ b/src/Makefile.d/sdl.mk
@@ -56,13 +56,6 @@ SDL_LDFLAGS?=$(shell $(SDL_CONFIG) \
 $(eval $(call Propogate_flags,SDL))
 endif
 
-# use the x86 asm code
-ifndef CYGWIN32
-ifndef NOASM
-USEASM=1
-endif
-endif
-
 ifdef MINGW
 ifndef NOSDLMAIN
 SDLMAIN=1
diff --git a/src/Makefile.d/win32.mk b/src/Makefile.d/win32.mk
index a6c66c84e..e1609fe20 100644
--- a/src/Makefile.d/win32.mk
+++ b/src/Makefile.d/win32.mk
@@ -13,9 +13,7 @@ sources+=win32/Srb2win.rc
 opts+=-DSTDC_HEADERS
 libs+=-ladvapi32 -lkernel32 -lmsvcrt -luser32
 
-nasm_format:=win32
-
-SDL=1
+SDL?=1
 
 ifndef NOHW
 opts+=-DUSE_WGL_SWAP
diff --git a/src/Sourcefile b/src/Sourcefile
index c3d4bca7e..36b9a0e79 100644
--- a/src/Sourcefile
+++ b/src/Sourcefile
@@ -92,7 +92,6 @@ mserv.c
 http-mserv.c
 i_tcp.c
 lzf.c
-vid_copy.s
 lua_script.c
 lua_baselib.c
 lua_mathlib.c
diff --git a/src/d_netcmd.c b/src/d_netcmd.c
index a1f2dcc87..e7e54842c 100644
--- a/src/d_netcmd.c
+++ b/src/d_netcmd.c
@@ -4904,11 +4904,6 @@ static void Command_Version_f(void)
 	else // 16-bit? 128-bit?
 		CONS_Printf("Bits Unknown ");
 
-	// No ASM?
-#ifdef NOASM
-	CONS_Printf("\x85" "NOASM " "\x80");
-#endif
-
 	// Debug build
 #ifdef _DEBUG
 	CONS_Printf("\x85" "DEBUG " "\x80");
diff --git a/src/m_fixed.h b/src/m_fixed.h
index 57dde19bd..7dc2f6778 100644
--- a/src/m_fixed.h
+++ b/src/m_fixed.h
@@ -168,7 +168,6 @@ FUNCMATH FUNCINLINE static ATTRINLINE fixed_t FloatToFixed(float f)
 	#define __USE_C_FIXEDDIV__ // no 64b asm div in MIPS land
 #elif defined (__GNUC__) && defined (__sh__) && 0 // DC: SH4 CPU
 #elif defined (__GNUC__) && defined (__m68k__) && 0 // DEAD: Motorola 6800 CPU
-#elif defined (_MSC_VER) && defined(USEASM) && FRACBITS == 16
 	// Microsoft Visual C++ (no asm inline)
 	fixed_t __cdecl FixedMul(fixed_t a, fixed_t b);
 	fixed_t __cdecl FixedDiv2(fixed_t a, fixed_t b);
diff --git a/src/p5prof.h b/src/p5prof.h
deleted file mode 100644
index a9ed3965e..000000000
--- a/src/p5prof.h
+++ /dev/null
@@ -1,278 +0,0 @@
-/*********************************************************
- *
- * File:  p5prof.h
- * By:    Kevin Baca
- *
- * MODIFIED BY Fab SO THAT RDMSR(...) WRITES EDX : EAX TO A LONG LONG
- * (WHICH MEANS WRITE THE LOW DWORD FIRST)
- *
- * Now in yer code do:
- *   INT64 count,total;
- *
- *   ...
- *   RDMSR(0x10,&count);        //inner loop count
- *   total += count;
- *   ...
- *
- *   printf("0x%x %x", (INT32)total, *((INT32 *)&total+1));
- *   //                  HIGH        LOW
- *
- *********************************************************/
-/**\file
-	\brief  This file provides macros to profile your code.
-
- Here's how they work...
-
- As you may or may not know, the Pentium class of
- processors provides extremely fine grained profiling
- capabilities through the use of what are called
- Machine Specific Registers (MSRs). These registers
- can provide information about almost any aspect of
- CPU performance down to a single cycle.
-
- The MSRs of interest for profiling are specified by
- indices 0x10, 0x11, 0x12, and 0x13.  Here is a brief
- description of each of these registers:
-
- MSR 0x10
-    This register is simple a cycle counter.
-
- MSR 0x11
-    This register controls what type of profiling data
- will be gathered.
-
- MSRs 0x12 and 0x13
-    These registers gather the profiling data specified in
- MSR 0x11.
-
- Each MSR is 64 bits wide.  For the Pentium processor,
- only the lower 32 bits of MSR 0x11 are valid.  Bits 0-15
- specify what data will be gathered in MSR 0x12.  Bits 16-31
- specify what data will be gathered in MSR 0x13.  Both sets
- of bits have the same format:
-
- Bits 0-5 specify which hardware event will be tracked.
- Bit 6, if set, indicates events will be tracked in
- rings 0-2.
- Bit 7, if set, indicates events will be tracked in
- ring 3.
- Bit 8, if set, indicates cycles should be counted for
- the specified event.  If clear, it indicates the
- number of events should be counted.
-
- Two instructions are provided for manupulating the MSRs.
- RDMSR (Read Machine Specific Register) and WRMSR
- (Write Machine Specific Register).  These opcodes were
- originally undocumented and therefore most assemblers don't
- recognize them.  Their byte codes are provided in the
- macros below.
-
- RDMSR takes the MSR index in ecx and the profiling criteria
- in edx : eax.
-
- WRMSR takes the MSR index in ecx and returns the profile data
- in edx : eax.
-
- Two profiling registers limits profiling capability to
- gathering only two types of information.  The register
- usage can, however, be combined in interesting ways.
- For example, you can set one register to gather the
- number of a specific type of event while the other gathers
- the number of cycles for the same event.  Or you can
- gather the number of two separate events while using
- MSR 0x10 to gather the number of cycles.
-
- The enumerated list provides somewhat readable labels for
- the types of events that can be tracked.
-
- For more information, get ahold of appendix H from the
- Intel Pentium programmer's manual (I don't remember the
- order number) or go to
- http://green.kaist.ac.kr/jwhahn/art3.htm.
- That's an article by Terje Mathisen where I got most of
- my information.
-
- You may use this code however you wish.  I hope it's
- useful and I hope I got everything right.
-
- -Kevin
-
- kbaca@skygames.com
-
-*/
-
-#ifdef __GNUC__
-
-#define RDTSC(_dst) \
-__asm__("
-     .byte 0x0F,0x31
-     movl %%edx,(%%edi)
-     movl %%eax,4(%%edi)"\
-: : "D" (_dst) : "eax", "edx", "edi")
-
-// the old code... swapped it
-//     movl %%edx,(%%edi)
-//     movl %%eax,4(%%edi)"
-#define RDMSR(_msri, _msrd) \
-__asm__("
-     .byte 0x0F,0x32
-     movl %%eax,(%%edi)
-     movl %%edx,4(%%edi)"\
-: : "c" (_msri), "D" (_msrd) : "eax", "ecx", "edx", "edi")
-
-#define WRMSR(_msri, _msrd) \
-__asm__("
-     xorl %%edx,%%edx
-     .byte 0x0F,0x30"\
-: : "c" (_msri), "a" (_msrd) : "eax", "ecx", "edx")
-
-#define RDMSR_0x12_0x13(_msr12, _msr13) \
-__asm__("
-     movl $0x12,%%ecx
-     .byte 0x0F,0x32
-     movl %%edx,(%%edi)
-     movl %%eax,4(%%edi)
-     movl $0x13,%%ecx
-     .byte 0x0F,0x32
-     movl %%edx,(%%esi)
-     movl %%eax,4(%%esi)"\
-: : "D" (_msr12), "S" (_msr13) : "eax", "ecx", "edx", "edi")
-
-#define ZERO_MSR_0x12_0x13() \
-__asm__("
-     xorl %%edx,%%edx
-     xorl %%eax,%%eax
-     movl $0x12,%%ecx
-     .byte 0x0F,0x30
-     movl $0x13,%%ecx
-     .byte 0x0F,0x30"\
-: : : "eax", "ecx", "edx")
-
-#elif defined (__WATCOMC__)
-
-extern void RDTSC(UINT32 *dst);
-#pragma aux RDTSC =\
-   "db 0x0F,0x31"\
-   "mov [edi],edx"\
-   "mov [4+edi],eax"\
-   parm [edi]\
-   modify [eax edx edi];
-
-extern void RDMSR(UINT32 msri, UINT32 *msrd);
-#pragma aux RDMSR =\
-   "db 0x0F,0x32"\
-   "mov [edi],edx"\
-   "mov [4+edi],eax"\
-   parm [ecx] [edi]\
-   modify [eax ecx edx edi];
-
-extern void WRMSR(UINT32 msri, UINT32 msrd);
-#pragma aux WRMSR =\
-   "xor edx,edx"\
-   "db 0x0F,0x30"\
-   parm [ecx] [eax]\
-   modify [eax ecx edx];
-
-extern void RDMSR_0x12_0x13(UINT32 *msr12, UINT32 *msr13);
-#pragma aux RDMSR_0x12_0x13 =\
-   "mov ecx,0x12"\
-   "db 0x0F,0x32"\
-   "mov [edi],edx"\
-   "mov [4+edi],eax"\
-   "mov ecx,0x13"\
-   "db 0x0F,0x32"\
-   "mov [esi],edx"\
-   "mov [4+esi],eax"\
-   parm [edi] [esi]\
-   modify [eax ecx edx edi esi];
-
-extern void ZERO_MSR_0x12_0x13(void);
-#pragma aux ZERO_MSR_0x12_0x13 =\
-   "xor edx,edx"\
-   "xor eax,eax"\
-   "mov ecx,0x12"\
-   "db 0x0F,0x30"\
-   "mov ecx,0x13"\
-   "db 0x0F,0x30"\
-   modify [eax ecx edx];
-
-#endif
-
-typedef enum
-{
-   DataRead,
-     DataWrite,
-     DataTLBMiss,
-     DataReadMiss,
-     DataWriteMiss,
-     WriteHitEM,
-     DataCacheLinesWritten,
-     DataCacheSnoops,
-     DataCacheSnoopHit,
-     MemAccessBothPipes,
-     BankConflict,
-     MisalignedDataRef,
-     CodeRead,
-     CodeTLBMiss,
-     CodeCacheMiss,
-     SegRegLoad,
-     RESERVED0,
-     RESERVED1,
-     Branch,
-     BTBHit,
-     TakenBranchOrBTBHit,
-     PipelineFlush,
-     InstructionsExeced,
-     InstructionsExecedVPipe,
-     BusUtilizationClocks,
-     PipelineStalledWriteBackup,
-     PipelineStalledDateMemRead,
-     PipeLineStalledWriteEM,
-     LockedBusCycle,
-     IOReadOrWriteCycle,
-     NonCacheableMemRef,
-     AGI,
-     RESERVED2,
-     RESERVED3,
-     FPOperation,
-     Breakpoint0Match,
-     Breakpoint1Match,
-     Breakpoint2Match,
-     Breakpoint3Match,
-     HWInterrupt,
-     DataReadOrWrite,
-     DataReadOrWriteMiss
-};
-
-#define PROF_CYCLES (0x100)
-#define PROF_EVENTS (0x000)
-#define RING_012    (0x40)
-#define RING_3      (0x80)
-#define RING_0123   (RING_012 | RING_3)
-
-/*void ProfSetProfiles(UINT32 msr12, UINT32 msr13);*/
-#define ProfSetProfiles(_msr12, _msr13)\
-{\
-   UINT32 prof;\
-\
-   prof = (_msr12) | ((_msr13) << 16);\
-   WRMSR(0x11, prof);\
-}
-
-/*void ProfBeginProfiles(void);*/
-#define ProfBeginProfiles()\
-   ZERO_MSR_0x12_0x13();
-
-/*void ProfGetProfiles(UINT32 msr12[2], UINT32 msr13[2]);*/
-#define ProfGetProfiles(_msr12, _msr13)\
-   RDMSR_0x12_0x13(_msr12, _msr13);
-
-/*void ProfZeroTimer(void);*/
-#define ProfZeroTimer()\
-   WRMSR(0x10, 0);
-
-/*void ProfReadTimer(UINT32 timer[2]);*/
-#define ProfReadTimer(timer)\
-   RDMSR(0x10, timer);
-
-/*EOF*/
diff --git a/src/screen.c b/src/screen.c
index ddc9f5af0..2a4111426 100644
--- a/src/screen.c
+++ b/src/screen.c
@@ -45,10 +45,6 @@
 // SRB2Kart
 #include "r_fps.h" // R_GetFramerateCap
 
-#if defined (USEASM) && !defined (NORUSEASM)//&& (!defined (_MSC_VER) || (_MSC_VER <= 1200))
-#define RUSEASM //MSC.NET can't patch itself
-#endif
-
 // ------------------
 // global video state
 // ------------------
diff --git a/src/sdl/CMakeLists.txt b/src/sdl/CMakeLists.txt
index 03b18d219..e84dfafcc 100644
--- a/src/sdl/CMakeLists.txt
+++ b/src/sdl/CMakeLists.txt
@@ -13,11 +13,6 @@ target_sources(SRB2SDL2 PRIVATE
 	hwsym_sdl.c
 )
 
-if(${SRB2_USEASM})
-	set_source_files_properties(${SRB2_ASM_SOURCES} PROPERTIES LANGUAGE C)
-	set_source_files_properties(${SRB2_ASM_SOURCES} PROPERTIES COMPILE_FLAGS "-x assembler-with-cpp")
-endif()
-
 if("${CMAKE_SYSTEM_NAME}" MATCHES Windows)
 	target_sources(SRB2SDL2 PRIVATE
 		../win32/win_dbg.c
@@ -66,18 +61,6 @@ if("${CMAKE_SYSTEM_NAME}" MATCHES Linux)
 	target_link_libraries(SRB2SDL2 PRIVATE m rt)
 endif()
 
-if(${SRB2_USEASM})
-	if(${SRB2_CONFIG_YASM})
-		set(ASM_ASSEMBLER_TEMP ${CMAKE_ASM_YASM_COMPILER})
-		set(ASM_ASSEMBLER_OBJFORMAT ${CMAKE_ASM_YASM_OBJECT_FORMAT})
-		set_source_files_properties(${SRB2_NASM_SOURCES} LANGUAGE ASM_YASM)
-	else()
-		set(ASM_ASSEMBLER_TEMP ${CMAKE_ASM_NASM_COMPILER})
-		set(ASM_ASSEMBLER_OBJFORMAT ${CMAKE_ASM_NASM_OBJECT_FORMAT})
-		set_source_files_properties(${SRB2_NASM_SOURCES} LANGUAGE ASM_NASM)
-	endif()
-endif()
-
 if("${CMAKE_SYSTEM_NAME}" MATCHES Windows)
 	target_link_libraries(SRB2SDL2 PRIVATE
 		ws2_32
diff --git a/src/sdl/MakeCYG.cfg b/src/sdl/MakeCYG.cfg
index 5907579c1..b78316b00 100644
--- a/src/sdl/MakeCYG.cfg
+++ b/src/sdl/MakeCYG.cfg
@@ -7,7 +7,6 @@
 
 	NOHW=1
 	NOHS=1
-	NOASM=1
 
 	OPTS+=-DLINUX
 
diff --git a/src/sdl/Srb2SDL-vc9.vcproj b/src/sdl/Srb2SDL-vc9.vcproj
index f1574feba..de60e4785 100644
--- a/src/sdl/Srb2SDL-vc9.vcproj
+++ b/src/sdl/Srb2SDL-vc9.vcproj
@@ -50,7 +50,7 @@
 				Name="VCCLCompilerTool"
 				Optimization="0"
 				AdditionalIncludeDirectories="&quot;$(ProjectDir)..\..\libs\libpng-src&quot;;&quot;$(ProjectDir)..\..\libs\zlib&quot;"
-				PreprocessorDefinitions="_DEBUG;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;USEASM;HAVE_PNG;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE"
+				PreprocessorDefinitions="_DEBUG;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;HAVE_PNG;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE"
 				MinimalRebuild="true"
 				BasicRuntimeChecks="3"
 				RuntimeLibrary="1"
@@ -248,7 +248,7 @@
 				FavorSizeOrSpeed="1"
 				OmitFramePointers="true"
 				AdditionalIncludeDirectories="&quot;$(ProjectDir)..\..\libs\libpng-src&quot;;&quot;$(ProjectDir)..\..\libs\zlib&quot;"
-				PreprocessorDefinitions="NDEBUG;SDLMAIN;NO_STDIO_REDIRECT;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;USEASM;HAVE_PNG;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE"
+				PreprocessorDefinitions="NDEBUG;SDLMAIN;NO_STDIO_REDIRECT;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;HAVE_PNG;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE"
 				StringPooling="true"
 				RuntimeLibrary="0"
 				PrecompiledHeaderFile=".\..\..\objs\VC9\$(Platform)\$(Configuration)\SDL\Srb2SDL-vc9.pch"
diff --git a/src/sdl/Srb2SDL.dsp b/src/sdl/Srb2SDL.dsp
index 9f6dd7b33..2297c2cfd 100644
--- a/src/sdl/Srb2SDL.dsp
+++ b/src/sdl/Srb2SDL.dsp
@@ -45,7 +45,7 @@ MTL=midl.exe
 # ADD BASE MTL /nologo /D "NDEBUG" /mktyplib203 /win32
 # ADD MTL /nologo /D "NDEBUG" /mktyplib203 /win32
 # ADD BASE CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c
-# ADD CPP /nologo /G5 /W3 /GX /Zi /Ot /Og /Oi /Op /Oy /Ob1 /I "..\..\libs\libpng-src" /I "..\..\libs\zlib" /D "NDEBUG" /D "SDLMAIN" /D "NO_STDIO_REDIRECT" /D "USE_WGL_SWAP" /D "DIRECTFULLSCREEN" /D "HAVE_SDL" /D "HWRENDER" /D "HW3SOUND" /D "HAVE_FILTER" /D "HAVE_MIXER" /D "USEASM" /D "HAVE_PNG" /FR /FD /GF /c
+# ADD CPP /nologo /G5 /W3 /GX /Zi /Ot /Og /Oi /Op /Oy /Ob1 /I "..\..\libs\libpng-src" /I "..\..\libs\zlib" /D "NDEBUG" /D "SDLMAIN" /D "NO_STDIO_REDIRECT" /D "USE_WGL_SWAP" /D "DIRECTFULLSCREEN" /D "HAVE_SDL" /D "HWRENDER" /D "HW3SOUND" /D "HAVE_FILTER" /D "HAVE_MIXER" /D "HAVE_PNG" /FR /FD /GF /c
 # ADD BASE RSC /l 0x409 /d "NDEBUG"
 # ADD RSC /l 0x409 /d "NDEBUG"
 # SUBTRACT RSC /x
@@ -72,7 +72,7 @@ LINK32=link.exe
 # PROP Target_Dir ""
 MTL=midl.exe
 # ADD BASE CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /GZ /c
-# ADD CPP /nologo /G6 /W4 /WX /Gm /GX /ZI /Od /Op /I "..\..\libs\libpng-src" /I "..\..\libs\zlib" /D "_DEBUG" /D "USE_WGL_SWAP" /D "DIRECTFULLSCREEN" /D "HAVE_SDL" /D "HWRENDER" /D "HW3SOUND" /D "HAVE_FILTER" /D "HAVE_MIXER" /D "USEASM" /D "HAVE_PNG" /FR /FD /GZ /c
+# ADD CPP /nologo /G6 /W4 /WX /Gm /GX /ZI /Od /Op /I "..\..\libs\libpng-src" /I "..\..\libs\zlib" /D "_DEBUG" /D "USE_WGL_SWAP" /D "DIRECTFULLSCREEN" /D "HAVE_SDL" /D "HWRENDER" /D "HW3SOUND" /D "HAVE_FILTER" /D "HAVE_MIXER" /D "HAVE_PNG" /FR /FD /GZ /c
 # SUBTRACT CPP /YX
 # ADD BASE RSC /l 0x409 /d "_DEBUG"
 # ADD RSC /l 0x409 /d "_DEBUG"
diff --git a/src/sdl/i_main.c b/src/sdl/i_main.c
index 157a1f0be..77387aa8e 100644
--- a/src/sdl/i_main.c
+++ b/src/sdl/i_main.c
@@ -77,39 +77,6 @@ char logfilename[1024];
 typedef BOOL (WINAPI *p_IsDebuggerPresent)(VOID);
 #endif
 
-#if defined (_WIN32)
-static inline VOID MakeCodeWritable(VOID)
-{
-#ifdef USEASM // Disable write-protection of code segment
-	DWORD OldRights;
-	const DWORD NewRights = PAGE_EXECUTE_READWRITE;
-	PBYTE pBaseOfImage = (PBYTE)GetModuleHandle(NULL);
-	PIMAGE_DOS_HEADER dosH =(PIMAGE_DOS_HEADER)pBaseOfImage;
-	PIMAGE_NT_HEADERS ntH = (PIMAGE_NT_HEADERS)(pBaseOfImage + dosH->e_lfanew);
-	PIMAGE_OPTIONAL_HEADER oH = (PIMAGE_OPTIONAL_HEADER)
-		((PBYTE)ntH + sizeof (IMAGE_NT_SIGNATURE) + sizeof (IMAGE_FILE_HEADER));
-	LPVOID pA = pBaseOfImage+oH->BaseOfCode;
-	SIZE_T pS = oH->SizeOfCode;
-#if 1 // try to find the text section
-	PIMAGE_SECTION_HEADER ntS = IMAGE_FIRST_SECTION (ntH);
-	WORD s;
-	for (s = 0; s < ntH->FileHeader.NumberOfSections; s++)
-	{
-		if (memcmp (ntS[s].Name, ".text\0\0", 8) == 0)
-		{
-			pA = pBaseOfImage+ntS[s].VirtualAddress;
-			pS = ntS[s].Misc.VirtualSize;
-			break;
-		}
-	}
-#endif
-
-	if (!VirtualProtect(pA,pS,NewRights,&OldRights))
-		I_Error("Could not make code writable\n");
-#endif
-}
-#endif
-
 #ifdef LOGMESSAGES
 static void InitLogging(void)
 {
@@ -270,7 +237,6 @@ int main(int argc, char **argv)
 #ifndef __MINGW32__
 	prevExceptionFilter = SetUnhandledExceptionFilter(RecordExceptionInfo);
 #endif
-	MakeCodeWritable();
 #endif
 
 	// startup SRB2
diff --git a/src/sdl12/Srb2SDL-vc10.vcxproj b/src/sdl12/Srb2SDL-vc10.vcxproj
index 1edb3d2db..5d4b48708 100644
--- a/src/sdl12/Srb2SDL-vc10.vcxproj
+++ b/src/sdl12/Srb2SDL-vc10.vcxproj
@@ -96,7 +96,7 @@
     <ClCompile>
       <Optimization>Disabled</Optimization>
       <AdditionalIncludeDirectories>$(ProjectDir)..\..\libs\libpng-src;$(ProjectDir)..\..\libs\zlib;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
-      <PreprocessorDefinitions>_DEBUG;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;USEASM;HAVE_PNG;COMPVERSION;HAVE_BLUA;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
+      <PreprocessorDefinitions>_DEBUG;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;HAVE_PNG;COMPVERSION;HAVE_BLUA;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
       <MinimalRebuild>true</MinimalRebuild>
       <BasicRuntimeChecks>EnableFastChecks</BasicRuntimeChecks>
       <RuntimeLibrary>MultiThreadedDebug</RuntimeLibrary>
@@ -202,7 +202,7 @@
       <FavorSizeOrSpeed>Speed</FavorSizeOrSpeed>
       <OmitFramePointers>true</OmitFramePointers>
       <AdditionalIncludeDirectories>$(ProjectDir)..\..\libs\libpng-src;$(ProjectDir)..\..\libs\zlib;%(AdditionalIncludeDirectories)</AdditionalIncludeDirectories>
-      <PreprocessorDefinitions>NDEBUG;SDLMAIN;NO_STDIO_REDIRECT;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;USEASM;HAVE_PNG;COMPVERSION;HAVE_BLUA;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
+      <PreprocessorDefinitions>NDEBUG;SDLMAIN;NO_STDIO_REDIRECT;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;HAVE_PNG;COMPVERSION;HAVE_BLUA;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE;%(PreprocessorDefinitions)</PreprocessorDefinitions>
       <StringPooling>true</StringPooling>
       <RuntimeLibrary>MultiThreaded</RuntimeLibrary>
       <PrecompiledHeaderOutputFile>.\..\..\objs\VC10\$(Platform)\$(Configuration)\SDL\Srb2SDL-vc9.pch</PrecompiledHeaderOutputFile>
diff --git a/src/sdl12/Srb2SDL-vc9.vcproj b/src/sdl12/Srb2SDL-vc9.vcproj
index 16c00454d..85ee8940c 100644
--- a/src/sdl12/Srb2SDL-vc9.vcproj
+++ b/src/sdl12/Srb2SDL-vc9.vcproj
@@ -50,7 +50,7 @@
 				Name="VCCLCompilerTool"
 				Optimization="0"
 				AdditionalIncludeDirectories="&quot;$(ProjectDir)..\..\libs\libpng-src&quot;;&quot;$(ProjectDir)..\..\libs\zlib&quot;"
-				PreprocessorDefinitions="_DEBUG;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;USEASM;HAVE_PNG;HAVE_BLUA;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE"
+				PreprocessorDefinitions="_DEBUG;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;HAVE_PNG;HAVE_BLUA;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE"
 				MinimalRebuild="true"
 				BasicRuntimeChecks="3"
 				RuntimeLibrary="1"
@@ -248,7 +248,7 @@
 				FavorSizeOrSpeed="1"
 				OmitFramePointers="true"
 				AdditionalIncludeDirectories="&quot;$(ProjectDir)..\..\libs\libpng-src&quot;;&quot;$(ProjectDir)..\..\libs\zlib&quot;"
-				PreprocessorDefinitions="NDEBUG;SDLMAIN;NO_STDIO_REDIRECT;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;USEASM;HAVE_PNG;HAVE_BLUA;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE"
+				PreprocessorDefinitions="NDEBUG;SDLMAIN;NO_STDIO_REDIRECT;USE_WGL_SWAP;DIRECTFULLSCREEN;HAVE_SDL;HWRENDER;HW3SOUND;HAVE_FILTER;HAVE_MIXER;HAVE_PNG;HAVE_BLUA;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE"
 				StringPooling="true"
 				RuntimeLibrary="0"
 				PrecompiledHeaderFile=".\..\..\objs\VC9\$(Platform)\$(Configuration)\SDL\Srb2SDL-vc9.pch"
diff --git a/src/sdl12/Srb2SDL.dsp b/src/sdl12/Srb2SDL.dsp
index ce35e2e23..15c214f51 100644
--- a/src/sdl12/Srb2SDL.dsp
+++ b/src/sdl12/Srb2SDL.dsp
@@ -45,7 +45,7 @@ MTL=midl.exe
 # ADD BASE MTL /nologo /D "NDEBUG" /mktyplib203 /win32
 # ADD MTL /nologo /D "NDEBUG" /mktyplib203 /win32
 # ADD BASE CPP /nologo /W3 /GX /O2 /D "WIN32" /D "NDEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /c
-# ADD CPP /nologo /G5 /W3 /GX /Zi /Ot /Og /Oi /Op /Oy /Ob1 /I "..\..\libs\libpng-src" /I "..\..\libs\zlib" /D "NDEBUG" /D "SDLMAIN" /D "NO_STDIO_REDIRECT" /D "USE_WGL_SWAP" /D "DIRECTFULLSCREEN" /D "HAVE_SDL" /D "HWRENDER" /D "HW3SOUND" /D "HAVE_FILTER" /D "HAVE_MIXER" /D "USEASM" /D "HAVE_PNG" /FR /FD /GF /c
+# ADD CPP /nologo /G5 /W3 /GX /Zi /Ot /Og /Oi /Op /Oy /Ob1 /I "..\..\libs\libpng-src" /I "..\..\libs\zlib" /D "NDEBUG" /D "SDLMAIN" /D "NO_STDIO_REDIRECT" /D "USE_WGL_SWAP" /D "DIRECTFULLSCREEN" /D "HAVE_SDL" /D "HWRENDER" /D "HW3SOUND" /D "HAVE_FILTER" /D "HAVE_MIXER" /D "HAVE_PNG" /FR /FD /GF /c
 # ADD BASE RSC /l 0x409 /d "NDEBUG"
 # ADD RSC /l 0x409 /d "NDEBUG"
 # SUBTRACT RSC /x
@@ -72,7 +72,7 @@ LINK32=link.exe
 # PROP Target_Dir ""
 MTL=midl.exe
 # ADD BASE CPP /nologo /W3 /Gm /GX /ZI /Od /D "WIN32" /D "_DEBUG" /D "_CONSOLE" /D "_MBCS" /YX /FD /GZ /c
-# ADD CPP /nologo /G6 /W4 /WX /Gm /GX /ZI /Od /Op /I "..\..\libs\libpng-src" /I "..\..\libs\zlib" /D "_DEBUG" /D "USE_WGL_SWAP" /D "DIRECTFULLSCREEN" /D "HAVE_SDL" /D "HWRENDER" /D "HW3SOUND" /D "HAVE_FILTER" /D "HAVE_MIXER" /D "USEASM" /D "HAVE_PNG" /FR /FD /GZ /c
+# ADD CPP /nologo /G6 /W4 /WX /Gm /GX /ZI /Od /Op /I "..\..\libs\libpng-src" /I "..\..\libs\zlib" /D "_DEBUG" /D "USE_WGL_SWAP" /D "DIRECTFULLSCREEN" /D "HAVE_SDL" /D "HWRENDER" /D "HW3SOUND" /D "HAVE_FILTER" /D "HAVE_MIXER" /D "HAVE_PNG" /FR /FD /GZ /c
 # SUBTRACT CPP /YX
 # ADD BASE RSC /l 0x409 /d "_DEBUG"
 # ADD RSC /l 0x409 /d "_DEBUG"
diff --git a/src/sdl12/i_main.c b/src/sdl12/i_main.c
index 7817196bd..372d7d00c 100644
--- a/src/sdl12/i_main.c
+++ b/src/sdl12/i_main.c
@@ -94,33 +94,7 @@ KOS_INIT_FLAGS(INIT_DEFAULT
 #if defined (_WIN32) && !defined (_XBOX) && !defined (_WIN32_WCE)
 static inline VOID MakeCodeWritable(VOID)
 {
-#ifdef USEASM // Disable write-protection of code segment
-	DWORD OldRights;
-	const DWORD NewRights = PAGE_EXECUTE_READWRITE;
-	PBYTE pBaseOfImage = (PBYTE)GetModuleHandle(NULL);
-	PIMAGE_DOS_HEADER dosH =(PIMAGE_DOS_HEADER)pBaseOfImage;
-	PIMAGE_NT_HEADERS ntH = (PIMAGE_NT_HEADERS)(pBaseOfImage + dosH->e_lfanew);
-	PIMAGE_OPTIONAL_HEADER oH = (PIMAGE_OPTIONAL_HEADER)
-		((PBYTE)ntH + sizeof (IMAGE_NT_SIGNATURE) + sizeof (IMAGE_FILE_HEADER));
-	LPVOID pA = pBaseOfImage+oH->BaseOfCode;
-	SIZE_T pS = oH->SizeOfCode;
-#if 1 // try to find the text section
-	PIMAGE_SECTION_HEADER ntS = IMAGE_FIRST_SECTION (ntH);
-	WORD s;
-	for (s = 0; s < ntH->FileHeader.NumberOfSections; s++)
-	{
-		if (memcmp (ntS[s].Name, ".text\0\0", 8) == 0)
-		{
-			pA = pBaseOfImage+ntS[s].VirtualAddress;
-			pS = ntS[s].Misc.VirtualSize;
-			break;
-		}
-	}
-#endif
-
-	if (!VirtualProtect(pA,pS,NewRights,&OldRights))
-		I_Error("Could not make code writable\n");
-#endif
+	;
 }
 #endif
 
diff --git a/src/tmap.nas b/src/tmap.nas
deleted file mode 100644
index 69282d0b4..000000000
--- a/src/tmap.nas
+++ /dev/null
@@ -1,957 +0,0 @@
-;; SONIC ROBO BLAST 2
-;;-----------------------------------------------------------------------------
-;; Copyright (C) 1998-2000 by DooM Legacy Team.
-;; Copyright (C) 1999-2020 by Sonic Team Junior.
-;;
-;; This program is free software distributed under the
-;; terms of the GNU General Public License, version 2.
-;; See the 'LICENSE' file for more details.
-;;-----------------------------------------------------------------------------
-;; FILE:
-;;      tmap.nas
-;; DESCRIPTION:
-;;      Assembler optimised rendering code for software mode.
-;;      Draw wall columns.
-
-
-[BITS 32]
-
-%define FRACBITS 16
-%define TRANSPARENTPIXEL 255
-
-%ifdef LINUX
-%macro cextern 1
-[extern %1]
-%endmacro
-
-%macro cglobal 1
-[global %1]
-%endmacro
-
-%else
-%macro cextern 1
-%define %1 _%1
-[extern %1]
-%endmacro
-
-%macro cglobal 1
-%define %1 _%1
-[global %1]
-%endmacro
-
-%endif
-
-
-; The viddef_s structure. We only need the width field.
-struc viddef_s
-        resb 12
-.width: resb 4
-        resb 44
-endstruc
-
-;; externs
-;; columns
-cextern dc_x
-cextern dc_yl
-cextern dc_yh
-cextern ylookup
-cextern columnofs
-cextern dc_source
-cextern dc_texturemid
-cextern dc_texheight
-cextern dc_iscale
-cextern dc_hires
-cextern centery
-cextern centeryfrac
-cextern dc_colormap
-cextern dc_transmap
-cextern colormaps
-cextern vid
-cextern topleft
-
-; DELME
-cextern R_DrawColumn_8
-
-; polygon edge rasterizer
-cextern prastertab
-
-[SECTION .data]
-
-;;.align        4
-loopcount       dd      0
-pixelcount      dd      0
-tystep          dd      0
-
-[SECTION .text]
-
-;;----------------------------------------------------------------------
-;;
-;; R_DrawColumn : 8bpp column drawer
-;;
-;; New  optimised version 10-01-1998 by D.Fabrice and P.Boris
-;; Revised by G. Dick July 2010 to support the intervening twelve years'
-;; worth of changes to the renderer. Since I only vaguely know what I'm
-;; doing, this is probably rather suboptimal. Help appreciated!
-;;
-;;----------------------------------------------------------------------
-;; fracstep, vid.width in memory
-;; eax = accumulator
-;; ebx = colormap
-;; ecx = count
-;; edx = heightmask
-;; esi = source
-;; edi = dest
-;; ebp = frac
-;;----------------------------------------------------------------------
-
-cglobal R_DrawColumn_8_ASM
-;       align   16
-R_DrawColumn_8_ASM:
-        push    ebp                     ;; preserve caller's stack frame pointer
-        push    esi                     ;; preserve register variables
-        push    edi
-        push    ebx
-;;
-;; dest = ylookup[dc_yl] + columnofs[dc_x];
-;;
-        mov     ebp,[dc_yl]
-        mov     edi,[ylookup+ebp*4]
-        mov     ebx,[dc_x]
-        add     edi,[columnofs+ebx*4]  ;; edi = dest
-;;
-;; pixelcount = yh - yl + 1
-;;
-        mov     ecx,[dc_yh]
-        add     ecx,1
-        sub     ecx,ebp                 ;; pixel count
-        jle     near .done              ;; nothing to scale
-;;
-;; fracstep = dc_iscale;	// But we just use [dc_iscale]
-;; frac = (dc_texturemid + FixedMul((dc_yl << FRACBITS) - centeryfrac, fracstep));
-;;
-        mov     eax,ebp                 ;; dc_yl
-        shl     eax,FRACBITS
-        sub     eax,[centeryfrac]
-        imul    dword [dc_iscale]
-        shrd    eax,edx,FRACBITS
-        add     eax,[dc_texturemid]
-        mov     ebp,eax                 ;; ebp = frac
-
-        mov     ebx,[dc_colormap]
-
-        mov     esi,[dc_source]
-;;
-;; if (dc_hires) frac = 0;
-;;
-        test    byte [dc_hires],0x01
-        jz      .texheightcheck
-        xor     ebp,ebp
-
-;;
-;; Check for power of two
-;;
-.texheightcheck:
-        mov     edx,[dc_texheight]
-        sub     edx,1                   ;; edx = heightmask
-        test    edx,[dc_texheight]
-        jnz     .notpowertwo
-
-        test    ecx,0x01                ;; Test for odd no. pixels
-        jnz     .odd
-
-;;
-;; Texture height is a power of two, so we get modular arithmetic by
-;; masking
-;;
-.powertwo:
-        mov     eax,ebp                 ;; eax = frac
-        sar     eax,FRACBITS            ;; Integer part
-        and     eax,edx                 ;; eax &= heightmask
-        movzx   eax,byte [esi + eax]    ;; eax = texel
-        add     ebp,[dc_iscale]         ;; frac += fracstep
-        movzx   eax,byte [ebx+eax]      ;; Map through colormap
-        mov     [edi],al                ;; Write pixel
-                                        ;; dest += vid.width
-        add     edi,[vid + viddef_s.width]
-
-.odd:
-        mov     eax,ebp                 ;; eax = frac
-        sar     eax,FRACBITS            ;; Integer part
-        and     eax,edx                 ;; eax &= heightmask
-        movzx   eax,byte [esi + eax]    ;; eax = texel
-        add     ebp,[dc_iscale]         ;; frac += fracstep
-        movzx   eax,byte [ebx+eax]      ;; Map through colormap
-        mov     [edi],al                ;; Write pixel
-                                        ;; dest += vid.width
-        add     edi,[vid + viddef_s.width]
-
-
-        sub     ecx,2                   ;; count -= 2
-        jg      .powertwo
-
-        jmp     .done
-
-.notpowertwo:
-        add     edx,1
-        shl     edx,FRACBITS
-        test    ebp,ebp
-        jns     .notpowtwoloop
-
-.makefracpos:
-        add     ebp,edx                 ;; frac is negative; make it positive
-        js      .makefracpos
-
-.notpowtwoloop:
-        cmp     ebp,edx                 ;; Reduce mod height
-        jl      .writenonpowtwo
-        sub     ebp,edx
-        jmp     .notpowtwoloop
-
-.writenonpowtwo:
-        mov     eax,ebp                 ;; eax = frac
-        sar     eax,FRACBITS            ;; Integer part.
-        mov     bl,[esi + eax]          ;; ebx = colormap + texel
-        add     ebp,[dc_iscale]         ;; frac += fracstep
-        movzx   eax,byte [ebx]          ;; Map through colormap
-        mov     [edi],al                ;; Write pixel
-                                        ;; dest += vid.width
-        add     edi,[vid + viddef_s.width]
-
-        sub     ecx,1
-        jnz     .notpowtwoloop
-
-;;
-
-.done:
-        pop     ebx                     ;; restore register variables
-        pop     edi
-        pop     esi
-        pop     ebp                     ;; restore caller's stack frame pointer
-        ret
-
-
-;;----------------------------------------------------------------------
-;;
-;; R_Draw2sMultiPatchColumn : Like R_DrawColumn, but omits transparent
-;;                            pixels.
-;;
-;; New  optimised version 10-01-1998 by D.Fabrice and P.Boris
-;; Revised by G. Dick July 2010 to support the intervening twelve years'
-;; worth of changes to the renderer. Since I only vaguely know what I'm
-;; doing, this is probably rather suboptimal. Help appreciated!
-;;
-;;----------------------------------------------------------------------
-;; fracstep, vid.width in memory
-;; eax = accumulator
-;; ebx = colormap
-;; ecx = count
-;; edx = heightmask
-;; esi = source
-;; edi = dest
-;; ebp = frac
-;;----------------------------------------------------------------------
-
-cglobal R_Draw2sMultiPatchColumn_8_ASM
-;       align   16
-R_Draw2sMultiPatchColumn_8_ASM:
-        push    ebp                     ;; preserve caller's stack frame pointer
-        push    esi                     ;; preserve register variables
-        push    edi
-        push    ebx
-;;
-;; dest = ylookup[dc_yl] + columnofs[dc_x];
-;;
-        mov     ebp,[dc_yl]
-        mov     edi,[ylookup+ebp*4]
-        mov     ebx,[dc_x]
-        add     edi,[columnofs+ebx*4]  ;; edi = dest
-;;
-;; pixelcount = yh - yl + 1
-;;
-        mov     ecx,[dc_yh]
-        add     ecx,1
-        sub     ecx,ebp                 ;; pixel count
-        jle     near .done              ;; nothing to scale
-;;
-;; fracstep = dc_iscale;	// But we just use [dc_iscale]
-;; frac = (dc_texturemid + FixedMul((dc_yl << FRACBITS) - centeryfrac, fracstep));
-;;
-        mov     eax,ebp                 ;; dc_yl
-        shl     eax,FRACBITS
-        sub     eax,[centeryfrac]
-        imul    dword [dc_iscale]
-        shrd    eax,edx,FRACBITS
-        add     eax,[dc_texturemid]
-        mov     ebp,eax                 ;; ebp = frac
-
-        mov     ebx,[dc_colormap]
-
-        mov     esi,[dc_source]
-;;
-;; if (dc_hires) frac = 0;
-;;
-        test    byte [dc_hires],0x01
-        jz      .texheightcheck
-        xor     ebp,ebp
-
-;;
-;; Check for power of two
-;;
-.texheightcheck:
-        mov     edx,[dc_texheight]
-        sub     edx,1                   ;; edx = heightmask
-        test    edx,[dc_texheight]
-        jnz     .notpowertwo
-
-        test    ecx,0x01                ;; Test for odd no. pixels
-        jnz     .odd
-
-;;
-;; Texture height is a power of two, so we get modular arithmetic by
-;; masking
-;;
-.powertwo:
-        mov     eax,ebp                 ;; eax = frac
-        sar     eax,FRACBITS            ;; Integer part
-        and     eax,edx                 ;; eax &= heightmask
-        movzx   eax,byte [esi + eax]    ;; eax = texel
-        add     ebp,[dc_iscale]         ;; frac += fracstep
-        cmp     al,TRANSPARENTPIXEL     ;; Is pixel transparent?
-        je      .nextpowtwoeven         ;; If so, advance.
-        movzx   eax,byte [ebx+eax]      ;; Map through colormap
-        mov	    [edi],al                ;; Write pixel
-.nextpowtwoeven:
-                                        ;; dest += vid.width
-        add     edi,[vid + viddef_s.width]
-
-.odd:
-        mov     eax,ebp                 ;; eax = frac
-        sar     eax,FRACBITS            ;; Integer part
-        and     eax,edx                 ;; eax &= heightmask
-        movzx   eax,byte [esi + eax]    ;; eax = texel
-        add     ebp,[dc_iscale]         ;; frac += fracstep
-        cmp     al,TRANSPARENTPIXEL     ;; Is pixel transparent?
-        je      .nextpowtwoodd          ;; If so, advance.
-        movzx   eax,byte [ebx+eax]      ;; Map through colormap
-        mov     [edi],al                ;; Write pixel
-.nextpowtwoodd:
-                                        ;; dest += vid.width
-        add     edi,[vid + viddef_s.width]
-
-
-        sub     ecx,2                   ;; count -= 2
-        jg      .powertwo
-
-        jmp     .done
-
-.notpowertwo:
-        add     edx,1
-        shl     edx,FRACBITS
-        test    ebp,ebp
-        jns     .notpowtwoloop
-
-.makefracpos:
-        add     ebp,edx                 ;; frac is negative; make it positive
-        js      .makefracpos
-
-.notpowtwoloop:
-        cmp     ebp,edx                 ;; Reduce mod height
-        jl      .writenonpowtwo
-        sub     ebp,edx
-        jmp     .notpowtwoloop
-
-.writenonpowtwo:
-        mov     eax,ebp                 ;; eax = frac
-        sar     eax,FRACBITS            ;; Integer part.
-        mov     bl,[esi + eax]          ;; ebx = colormap + texel
-        add     ebp,[dc_iscale]         ;; frac += fracstep
-        cmp     bl,TRANSPARENTPIXEL     ;; Is pixel transparent?
-        je      .nextnonpowtwo          ;; If so, advance.
-        movzx   eax,byte [ebx]          ;; Map through colormap
-        mov     [edi],al                ;; Write pixel
-.nextnonpowtwo:
-                                        ;; dest += vid.width
-        add     edi,[vid + viddef_s.width]
-
-        sub     ecx,1
-        jnz     .notpowtwoloop
-
-;;
-
-.done:
-        pop     ebx                     ;; restore register variables
-        pop     edi
-        pop     esi
-        pop     ebp                     ;; restore caller's stack frame pointer
-        ret
-
-;;----------------------------------------------------------------------
-;; R_DrawTranslucentColumnA_8
-;;
-;; Vertical column texture drawer, with transparency. Replaces Doom2's
-;; 'fuzz' effect, which was not so beautiful.
-;; Transparency is always impressive in some way, don't know why...
-;;----------------------------------------------------------------------
-
-cglobal R_DrawTranslucentColumn_8_ASM
-R_DrawTranslucentColumn_8_ASM:
-        push    ebp                     ;; preserve caller's stack frame pointer
-        push    esi                     ;; preserve register variables
-        push    edi
-        push    ebx
-;;
-;; dest = ylookup[dc_yl] + columnofs[dc_x];
-;;
-        mov     ebp,[dc_yl]
-        mov     ebx,ebp
-        mov     edi,[ylookup+ebx*4]
-        mov     ebx,[dc_x]
-        add     edi,[columnofs+ebx*4]   ;; edi = dest
-;;
-;; pixelcount = yh - yl + 1
-;;
-        mov     eax,[dc_yh]
-        inc     eax
-        sub     eax,ebp                 ;; pixel count
-        mov     [pixelcount],eax        ;; save for final pixel
-        jle     near    vtdone         ;; nothing to scale
-;;
-;; frac = dc_texturemid - (centery-dc_yl)*fracstep;
-;;
-        mov     ecx,[dc_iscale]        ;; fracstep
-        mov     eax,[centery]
-        sub     eax,ebp
-        imul    eax,ecx
-        mov     edx,[dc_texturemid]
-        sub     edx,eax
-        mov     ebx,edx
-
-        shr     ebx,16                  ;; frac int.
-        and     ebx,0x7f
-        shl     edx,16                  ;; y frac up
-
-        mov     ebp,ecx
-        shl     ebp,16                  ;; fracstep f. up
-        shr     ecx,16                  ;; fracstep i. ->cl
-        and     cl,0x7f
-        push    cx
-        mov     ecx,edx
-        pop     cx
-        mov     edx,[dc_colormap]
-        mov     esi,[dc_source]
-;;
-;; lets rock :) !
-;;
-        mov     eax,[pixelcount]
-        shr     eax,0x2
-        test    byte [pixelcount],0x3
-        mov     ch,al                   ;; quad count
-        mov     eax,[dc_transmap]
-        je      vt4quadloop
-;;
-;;  do un-even pixel
-;;
-        test    byte [pixelcount],0x1
-        je      trf2
-
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        add     ecx,ebp
-        adc     bl,cl
-        mov     al,[edi]                ;; fetch dest  : index into colormap
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     dl,[edx]
-        mov     [edi],dl
-pf:     add     edi,0x12345678
-;;
-;;  do two non-quad-aligned pixels
-;;
-trf2:    test    byte [pixelcount],0x2
-        je      trf3
-
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        add     ecx,ebp
-        adc     bl,cl
-        mov     al,[edi]                ;; fetch dest  : index into colormap
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     dl,[edx]
-        mov     [edi],dl
-pg:     add     edi,0x12345678
-
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        add     ecx,ebp
-        adc     bl,cl
-        mov     al,[edi]                ;; fetch dest  : index into colormap
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     dl,[edx]
-        mov     [edi],dl
-ph:     add     edi,0x12345678
-;;
-;;  test if there was at least 4 pixels
-;;
-trf3:   test    ch,0xff                 ;; test quad count
-        je near vtdone
-
-;;
-;; ebp : ystep frac. upper 24 bits
-;; edx : y     frac. upper 24 bits
-;; ebx : y     i.    lower 7 bits,  masked for index
-;; ecx : ch = counter, cl = y step i.
-;; eax : colormap aligned 256
-;; esi : source texture column
-;; edi : dest screen
-;;
-vt4quadloop:
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     [tystep],ebp
-pi:     add     edi,0x12345678
-        mov     al,[edi]                ;; fetch dest  : index into colormap
-pj:     sub     edi,0x12345678
-        mov     ebp,edi
-pk:     sub     edi,0x12345678
-        jmp short inloop
-align 4
-vtquadloop:
-        add     ecx,[tystep]
-        adc     bl,cl
-q1:     add     ebp,0x23456789
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     dl,[edx]
-        mov     [edi],dl
-        mov     al,[ebp]                ;; fetch dest   : index into colormap
-inloop:
-        add     ecx,[tystep]
-        adc     bl,cl
-q2:     add     edi,0x23456789
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     dl,[edx]
-        mov     [ebp+0x0],dl
-        mov     al,[edi]                ;; fetch dest   : index into colormap
-
-        add     ecx,[tystep]
-        adc     bl,cl
-q3:     add     ebp,0x23456789
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     dl,[edx]
-        mov     [edi],dl
-        mov     al,[ebp]                ;; fetch dest   : index into colormap
-
-        add     ecx,[tystep]
-        adc     bl,cl
-q4:     add     edi,0x23456789
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     dl,[edx]
-        mov     [ebp],dl
-        mov     al,[edi]                ;; fetch dest   : index into colormap
-
-        dec     ch
-        jne     vtquadloop
-vtdone:
-        pop     ebx
-        pop     edi
-        pop     esi
-        pop     ebp
-        ret
-
-;;----------------------------------------------------------------------
-;; R_DrawShadeColumn
-;;
-;;   for smoke..etc.. test.
-;;----------------------------------------------------------------------
-cglobal R_DrawShadeColumn_8_ASM
-R_DrawShadeColumn_8_ASM:
-        push    ebp                     ;; preserve caller's stack frame pointer
-        push    esi                     ;; preserve register variables
-        push    edi
-        push    ebx
-
-;;
-;; dest = ylookup[dc_yl] + columnofs[dc_x];
-;;
-        mov     ebp,[dc_yl]
-        mov     ebx,ebp
-        mov     edi,[ylookup+ebx*4]
-        mov     ebx,[dc_x]
-        add     edi,[columnofs+ebx*4]  ;; edi = dest
-;;
-;; pixelcount = yh - yl + 1
-;;
-        mov     eax,[dc_yh]
-        inc     eax
-        sub     eax,ebp                 ;; pixel count
-        mov     [pixelcount],eax       ;; save for final pixel
-        jle near shdone                ;; nothing to scale
-;;
-;; frac = dc_texturemid - (centery-dc_yl)*fracstep;
-;;
-        mov     ecx,[dc_iscale]        ;; fracstep
-        mov     eax,[centery]
-        sub     eax,ebp
-        imul    eax,ecx
-        mov     edx,[dc_texturemid]
-        sub     edx,eax
-        mov     ebx,edx
-        shr     ebx,16                  ;; frac int.
-        and     ebx,byte +0x7f
-        shl     edx,16                  ;; y frac up
-
-        mov     ebp,ecx
-        shl     ebp,16                  ;; fracstep f. up
-        shr     ecx,16                  ;; fracstep i. ->cl
-        and     cl,0x7f
-
-        mov     esi,[dc_source]
-;;
-;; lets rock :) !
-;;
-        mov     eax,[pixelcount]
-        mov     dh,al
-        shr     eax,2
-        mov     ch,al                   ;; quad count
-        mov     eax,[colormaps]
-        test    dh,3
-        je      sh4quadloop
-;;
-;;  do un-even pixel
-;;
-        test    dh,0x1
-        je      shf2
-
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        add     edx,ebp
-        adc     bl,cl
-        mov     al,[edi]                ;; fetch dest  : index into colormap
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     [edi],dl
-pl:     add     edi,0x12345678
-;;
-;;  do two non-quad-aligned pixels
-;;
-shf2:
-        test    dh,0x2
-        je      shf3
-
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        add     edx,ebp
-        adc     bl,cl
-        mov     al,[edi]                ;; fetch dest  : index into colormap
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     [edi],dl
-pm:     add     edi,0x12345678
-
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        add     edx,ebp
-        adc     bl,cl
-        mov     al,[edi]                ;; fetch dest  : index into colormap
-        and     bl,0x7f
-        mov     dl,[eax]
-        mov     [edi],dl
-pn:     add     edi,0x12345678
-;;
-;;  test if there was at least 4 pixels
-;;
-shf3:
-        test    ch,0xff                 ;; test quad count
-        je near shdone
-
-;;
-;; ebp : ystep frac. upper 24 bits
-;; edx : y     frac. upper 24 bits
-;; ebx : y     i.    lower 7 bits,  masked for index
-;; ecx : ch = counter, cl = y step i.
-;; eax : colormap aligned 256
-;; esi : source texture column
-;; edi : dest screen
-;;
-sh4quadloop:
-        mov     dh,0x7f                 ;; prep mask
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     [tystep],ebp
-po:     add     edi,0x12345678
-        mov     al,[edi]                ;; fetch dest  : index into colormap
-pp:     sub     edi,0x12345678
-        mov     ebp,edi
-pq:     sub     edi,0x12345678
-        jmp short shinloop
-
-align  4
-shquadloop:
-        add     edx,[tystep]
-        adc     bl,cl
-        and     bl,dh
-q5:     add     ebp,0x12345678
-        mov     dl,[eax]
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     [edi],dl
-        mov     al,[ebp]                ;; fetch dest : index into colormap
-shinloop:
-        add     edx,[tystep]
-        adc     bl,cl
-        and     bl,dh
-q6:     add     edi,0x12345678
-        mov     dl,[eax]
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     [ebp],dl
-        mov     al,[edi]                ;; fetch dest : index into colormap
-
-        add     edx,[tystep]
-        adc     bl,cl
-        and     bl,dh
-q7:     add     ebp,0x12345678
-        mov     dl,[eax]
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     [edi],dl
-        mov     al,[ebp]                ;; fetch dest : index into colormap
-
-        add     edx,[tystep]
-        adc     bl,cl
-        and     bl,dh
-q8:     add     edi,0x12345678
-        mov     dl,[eax]
-        mov     ah,[esi+ebx]            ;; fetch texel : colormap number
-        mov     [ebp],dl
-        mov     al,[edi]                ;; fetch dest : index into colormap
-
-        dec     ch
-        jne     shquadloop
-
-shdone:
-        pop     ebx                     ;; restore register variables
-        pop     edi
-        pop     esi
-        pop     ebp                     ;; restore caller's stack frame pointer
-        ret
-
-
-;; ========================================================================
-;;  Rasterization of the segments of a LINEAR polygne textur of manire.
-;;  It is thus a question of interpolating coordinate them at the edges of texture in
-;;  the time that the X-coordinates minx/maxx for each line.
-;;  the argument ' dir' indicates which edges of texture are Interpol?:
-;;    0:  segments associs at edge TOP? and BOTTOM? (constant TY)
-;;    1:  segments associs at the LEFT and RIGHT edge (constant TX)
-;; ========================================================================
-;;
-;;  void   rasterize_segment_tex( LONG x1, LONG y1, LONG x2, LONG y2, LONG tv1, LONG tv2, LONG tc, LONG dir );
-;;                                   ARG1     ARG2     ARG3     ARG4      ARG5      ARG6     ARG7       ARG8
-;;
-;;  Pour dir = 0, (tv1,tv2) = (tX1,tX2), tc = tY, en effet TY est constant.
-;;
-;;  Pour dir = 1, (tv1,tv2) = (tY1,tY2), tc = tX, en effet TX est constant.
-;;
-;;
-;;  Uses:  extern struct rastery *_rastertab;
-;;
-
-MINX            EQU    0
-MAXX            EQU    4
-TX1             EQU    8
-TY1             EQU    12
-TX2             EQU    16
-TY2             EQU    20
-RASTERY_SIZEOF  EQU    24
-
-cglobal rasterize_segment_tex_asm
-rasterize_segment_tex_asm:
-        push    ebp
-        mov     ebp,esp
-
-        sub     esp,byte +0x8           ;; allocate the local variables
-
-        push    ebx
-        push    esi
-        push    edi
-        o16 mov ax,es
-        push    eax
-
-;;        #define DX       [ebp-4]
-;;        #define TD       [ebp-8]
-
-        mov     eax,[ebp+0xc]           ;; y1
-        mov     ebx,[ebp+0x14]          ;; y2
-        cmp     ebx,eax
-        je near .L_finished             ;; special (y1==y2) segment horizontal, exit!
-
-        jg near .L_rasterize_right
-
-;;rasterize_left:       ;; one rasterize a segment LEFT of the polygne
-
-        mov     ecx,eax
-        sub     ecx,ebx
-        inc     ecx                     ;; y1-y2+1
-
-        mov     eax,RASTERY_SIZEOF
-        mul     ebx                     ;; * y2
-        mov     esi,[prastertab]
-        add     esi,eax                 ;; point into rastertab[y2]
-
-        mov     eax,[ebp+0x8]           ;; ARG1
-        sub     eax,[ebp+0x10]          ;; ARG3
-        shl     eax,0x10                ;;     ((x1-x2)<<PRE) ...
-        cdq
-        idiv    ecx                     ;; dx =     ...        / (y1-y2+1)
-        mov     [ebp-0x4],eax           ;; DX
-
-        mov     eax,[ebp+0x18]          ;; ARG5
-        sub     eax,[ebp+0x1c]          ;; ARG6
-        shl     eax,0x10
-        cdq
-        idiv    ecx                     ;;      tdx =((tx1-tx2)<<PRE) / (y1-y2+1)
-        mov     [ebp-0x8],eax           ;; idem tdy =((ty1-ty2)<<PRE) / (y1-y2+1)
-
-        mov     eax,[ebp+0x10]          ;; ARG3
-        shl     eax,0x10                ;; x = x2<<PRE
-
-        mov     ebx,[ebp+0x1c]          ;; ARG6
-        shl     ebx,0x10                ;; tx = tx2<<PRE    d0
-                                        ;; ty = ty2<<PRE    d1
-        mov     edx,[ebp+0x20]          ;; ARG7
-        shl     edx,0x10                ;; ty = ty<<PRE     d0
-                                        ;; tx = tx<<PRE     d1
-        push    ebp
-        mov     edi,[ebp-0x4]           ;; DX
-        cmp     dword [ebp+0x24],byte +0x0      ;; ARG8   direction ?
-
-        mov     ebp,[ebp-0x8]           ;; TD
-        je      .L_rleft_h_loop
-;;
-;; TY varies, TX is constant
-;;
-.L_rleft_v_loop:
-        mov     [esi+MINX],eax           ;; rastertab[y].minx = x
-          add     ebx,ebp
-        mov     [esi+TX1],edx           ;;             .tx1  = tx
-          add     eax,edi
-        mov     [esi+TY1],ebx           ;;             .ty1  = ty
-
-        ;;addl    DX, %eax        // x     += dx
-        ;;addl    TD, %ebx        // ty    += tdy
-
-        add     esi,RASTERY_SIZEOF      ;; next raster line into rastertab[]
-        dec     ecx
-        jne     .L_rleft_v_loop
-        pop     ebp
-        jmp     .L_finished
-;;
-;; TX varies, TY is constant
-;;
-.L_rleft_h_loop:
-        mov     [esi+MINX],eax           ;; rastertab[y].minx = x
-          add     eax,edi
-        mov     [esi+TX1],ebx           ;;             .tx1  = tx
-          add     ebx,ebp
-        mov     [esi+TY1],edx           ;;             .ty1  = ty
-
-        ;;addl    DX, %eax        // x     += dx
-        ;;addl    TD, %ebx        // tx    += tdx
-
-        add     esi,RASTERY_SIZEOF      ;; next raster line into rastertab[]
-        dec     ecx
-        jne     .L_rleft_h_loop
-        pop     ebp
-        jmp     .L_finished
-;;
-;; one rasterize a segment LINE of the polygne
-;;
-.L_rasterize_right:
-        mov     ecx,ebx
-        sub     ecx,eax
-        inc     ecx                     ;; y2-y1+1
-
-        mov     ebx,RASTERY_SIZEOF
-        mul     ebx                     ;;   * y1
-        mov     esi,[prastertab]
-        add     esi,eax                 ;;  point into rastertab[y1]
-
-        mov     eax,[ebp+0x10]          ;; ARG3
-        sub     eax,[ebp+0x8]           ;; ARG1
-        shl     eax,0x10                ;; ((x2-x1)<<PRE) ...
-        cdq
-        idiv    ecx                     ;;  dx =     ...        / (y2-y1+1)
-        mov     [ebp-0x4],eax           ;; DX
-
-        mov     eax,[ebp+0x1c]          ;; ARG6
-        sub     eax,[ebp+0x18]          ;; ARG5
-        shl     eax,0x10
-        cdq
-        idiv    ecx                     ;;       tdx =((tx2-tx1)<<PRE) / (y2-y1+1)
-        mov     [ebp-0x8],eax           ;;  idem tdy =((ty2-ty1)<<PRE) / (y2-y1+1)
-
-        mov     eax,[ebp+0x8]           ;; ARG1
-        shl     eax,0x10                ;; x  = x1<<PRE
-
-        mov     ebx,[ebp+0x18]          ;; ARG5
-        shl     ebx,0x10                ;; tx = tx1<<PRE    d0
-                                        ;; ty = ty1<<PRE    d1
-        mov     edx,[ebp+0x20]          ;; ARG7
-        shl     edx,0x10                ;; ty = ty<<PRE     d0
-                                        ;; tx = tx<<PRE     d1
-        push    ebp
-        mov     edi,[ebp-0x4]           ;; DX
-
-        cmp     dword [ebp+0x24], 0     ;; direction ?
-
-         mov     ebp,[ebp-0x8]          ;; TD
-        je      .L_rright_h_loop
-;;
-;; TY varies, TX is constant
-;;
-.L_rright_v_loop:
-
-        mov     [esi+MAXX],eax           ;; rastertab[y].maxx = x
-          add     ebx,ebp
-        mov     [esi+TX2],edx          ;;             .tx2  = tx
-          add     eax,edi
-        mov     [esi+TY2],ebx          ;;             .ty2  = ty
-
-        ;;addl    DX, %eax        // x     += dx
-        ;;addl    TD, %ebx        // ty    += tdy
-
-        add     esi,RASTERY_SIZEOF
-        dec     ecx
-        jne     .L_rright_v_loop
-
-        pop     ebp
-
-        jmp     short .L_finished
-;;
-;; TX varies, TY is constant
-;;
-.L_rright_h_loop:
-        mov     [esi+MAXX],eax           ;; rastertab[y].maxx = x
-          add     eax,edi
-        mov     [esi+TX2],ebx          ;;             .tx2  = tx
-          add     ebx,ebp
-        mov     [esi+TY2],edx          ;;             .ty2  = ty
-
-        ;;addl    DX, %eax        // x     += dx
-        ;;addl    TD, %ebx        // tx    += tdx
-
-        add     esi,RASTERY_SIZEOF
-        dec     ecx
-        jne     .L_rright_h_loop
-
-        pop     ebp
-
-.L_finished:
-        pop     eax
-        o16 mov es,ax
-        pop     edi
-        pop     esi
-        pop     ebx
-
-        mov     esp,ebp
-        pop     ebp
-        ret
diff --git a/src/tmap.s b/src/tmap.s
deleted file mode 100644
index 3a4cf2e1a..000000000
--- a/src/tmap.s
+++ /dev/null
@@ -1,1587 +0,0 @@
-// SONIC ROBO BLAST 2
-//-----------------------------------------------------------------------------
-// Copyright (C) 1998-2000 by DooM Legacy Team.
-// Copyright (C) 1999-2020 by Sonic Team Junior.
-//
-// This program is free software distributed under the
-// terms of the GNU General Public License, version 2.
-// See the 'LICENSE' file for more details.
-//-----------------------------------------------------------------------------
-/// \file  tmap.s
-/// \brief optimised drawing routines for span/column rendering
-
-// structures, must match the C structures!
-#include "asm_defs.inc"
-
-// Rappel: seuls EAX, ECX, EDX peuvent �tre �cras�s librement.
-//         il faut sauver esi,edi, cd...gs
-
-/* Attention aux comparaisons!                                              */
-/*                                                                          */
-/*      Intel_compare:                                                      */
-/*                                                                          */
-/*              cmp     A,B                     // A-B , set flags          */
-/*              jg      A_greater_than_B                                    */
-/*                                                                          */
-/*      AT&T_compare:                                                       */
-/*                                                                          */
-/*              cmp     A,B                     // B-A , set flags          */
-/*              jg      B_greater_than_A                                    */
-/*                                                                          */
-/*        (soustrait l'op�rande source DE l'op�rande destination,           */
-/*         comme sur Motorola! )                                            */
-
-// RAPPEL: Intel
-//         SECTION:[BASE+INDEX*SCALE+DISP]
-// devient SECTION:DISP(BASE,INDEX,SCALE)
-
-//----------------------------------------------------------------------
-//
-// R_DrawColumn
-//
-//   New optimised version 10-01-1998 by D.Fabrice and P.Boris
-//   TO DO: optimise it much farther... should take at most 3 cycles/pix
-//          once it's fixed, add code to patch the offsets so that it
-//          works in every screen width.
-//
-//----------------------------------------------------------------------
-
-    .data
-#ifdef LINUX
-    .align 2
-#else
-    .align 4
-#endif
-C(loopcount):   .long   0
-C(pixelcount):  .long   0
-C(tystep):      .long   0
-
-C(vidwidth):    .long   0       //use this one out of the inner loops
-                                //so you don't need to patch everywhere...
-
-#ifdef USEASM
-#if !defined( LINUX)
-    .text
-#endif
-.globl C(ASM_PatchRowBytes)
-C(ASM_PatchRowBytes):
-    pushl   %ebp
-    movl    %esp, %ebp      // assure l'"adressabilit� du stack"
-
-    movl    ARG1, %edx         // read first arg
-    movl    %edx, C(vidwidth)
-
-    // 1 * vidwidth
-    movl    %edx,p1+2
-    movl    %edx,w1+2   //water
-    movl    %edx,p1b+2  //sky
-
-    movl    %edx,p5+2
-      movl    %edx,sh5+2        //smokie test
-
-    // 2 * vidwidth
-    addl    ARG1,%edx
-
-    movl    %edx,p2+2
-    movl    %edx,w2+2   //water
-    movl    %edx,p2b+2  //sky
-
-    movl    %edx,p6+2
-    movl    %edx,p7+2
-    movl    %edx,p8+2
-    movl    %edx,p9+2
-      movl    %edx,sh6+2         //smokie test
-      movl    %edx,sh7+2
-      movl    %edx,sh8+2
-      movl    %edx,sh9+2
-
-    // 3 * vidwidth
-    addl    ARG1,%edx
-
-    movl    %edx,p3+2
-    movl    %edx,w3+2   //water
-    movl    %edx,p3b+2  //sky
-
-    // 4 * vidwidth
-    addl    ARG1,%edx
-
-    movl    %edx,p4+2
-    movl    %edx,w4+2   //water
-    movl    %edx,p4b+2  //sky
-
-    popl    %ebp
-    ret
-
-
-#ifdef LINUX
-    .align 2
-#else
-    .align 5
-#endif
-.globl C(R_DrawColumn_8)
-C(R_DrawColumn_8):
-    pushl   %ebp                // preserve caller's stack frame pointer
-    pushl   %esi                // preserve register variables
-    pushl   %edi
-    pushl   %ebx
-
-//
-// dest = ylookup[dc_yl] + columnofs[dc_x];
-//
-    movl     C(dc_yl),%ebp
-    movl     %ebp,%ebx
-    movl     C(ylookup)(,%ebx,4),%edi
-    movl     C(dc_x),%ebx
-    addl     C(columnofs)(,%ebx,4),%edi  // edi = dest
-
-//
-// pixelcount = yh - yl + 1
-//
-    movl     C(dc_yh),%eax
-    incl     %eax
-    subl     %ebp,%eax                   // pixel count
-    movl     %eax,C(pixelcount)          // save for final pixel
-    jle      vdone                       // nothing to scale
-
-//
-// frac = dc_texturemid - (centery-dc_yl)*fracstep;
-//
-    movl     C(dc_iscale),%ecx           // fracstep
-    movl     C(centery),%eax
-    subl     %ebp,%eax
-    imul     %ecx,%eax
-    movl     C(dc_texturemid),%edx
-    subl     %eax,%edx
-     movl     %edx,%ebx
-     shrl     $16,%ebx          // frac int.
-     andl     $0x0000007f,%ebx
-     shll     $16,%edx          // y frac up
-
-     movl     %ecx,%ebp
-     shll     $16,%ebp          // fracstep f. up
-     shrl     $16,%ecx          // fracstep i. ->cl
-     andb     $0x7f,%cl
-
-    movl     C(dc_source),%esi
-
-//
-// lets rock :) !
-//
-    movl    C(pixelcount),%eax
-    movb    %al,%dh
-    shrl    $2,%eax
-    movb    %al,%ch             // quad count
-    movl    C(dc_colormap),%eax
-    testb   $3,%dh
-    jz      v4quadloop
-
-//
-//  do un-even pixel
-//
-    testb   $1,%dh
-    jz      2f
-
-    movb    (%esi,%ebx),%al     // prep un-even loops
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    andb    $0x7f,%bl            // mask 0-127 texture index
-     movb    %dl,(%edi)           // output pixel
-    addl    C(vidwidth),%edi
-
-//
-//  do two non-quad-aligned pixels
-//
-2:
-    testb   $2,%dh
-    jz      3f
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    andb    $0x7f,%bl            // mask 0-127 texture index
-     movb    %dl,(%edi)           // output pixel
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    andb    $0x7f,%bl            // mask 0-127 texture index
-    addl    C(vidwidth),%edi
-     movb    %dl,(%edi)           // output pixel
-
-    addl    C(vidwidth),%edi
-
-//
-//  test if there was at least 4 pixels
-//
-3:
-    testb   $0xFF,%ch           // test quad count
-    jz      vdone
-
-//
-// ebp : ystep frac. upper 24 bits
-// edx : y     frac. upper 24 bits
-// ebx : y     i.    lower 7 bits,  masked for index
-// ecx : ch = counter, cl = y step i.
-// eax : colormap aligned 256
-// esi : source texture column
-// edi : dest screen
-//
-v4quadloop:
-    movb    $0x7f,%dh           // prep mask
-//    .align  4
-vquadloop:
-    movb    (%esi,%ebx),%al     // prep loop
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    movb    %dl,(%edi)           // output pixel
-     andb    $0x7f,%bl            // mask 0-127 texture index
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-p1:    movb    %dl,0x12345678(%edi)
-     andb    $0x7f,%bl
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-p2:    movb    %dl,2*0x12345678(%edi)
-     andb    $0x7f,%bl
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-p3:    movb    %dl,3*0x12345678(%edi)
-     andb    $0x7f,%bl
-
-p4:    addl    $4*0x12345678,%edi
-
-    decb   %ch
-     jnz    vquadloop
-
-vdone:
-    popl    %ebx                // restore register variables
-    popl    %edi
-    popl    %esi
-    popl    %ebp                // restore caller's stack frame pointer
-    ret
-
-#ifdef HORIZONTALDRAW
-// --------------------------------------------------------------------------
-// Horizontal Column Drawer Optimisation
-// --------------------------------------------------------------------------
-
-#ifdef LINUX
-    .align 2
-#else
-    .align 5
-#endif
-.globl C(R_DrawHColumn_8)
-C(R_DrawHColumn_8):
-    pushl   %ebp
-    pushl   %esi
-    pushl   %edi
-    pushl   %ebx
-
-//
-// dest = yhlookup[dc_x] + hcolumnofs[dc_yl];
-//
-    movl    C(dc_x),%ebx
-    movl    C(yhlookup)(,%ebx,4),%edi
-    movl    C(dc_yl),%ebp
-    movl    %ebp,%ebx
-    addl    C(hcolumnofs)(,%ebx,4),%edi  // edi = dest
-
-//
-// pixelcount = yh - yl + 1
-//
-    movl     C(dc_yh),%eax
-    incl     %eax
-    subl     %ebp,%eax                   // pixel count
-    movl     %eax,C(pixelcount)          // save for final pixel
-    jle      vhdone                      // nothing to scale
-
-//
-// frac = dc_texturemid - (centery-dc_yl)*fracstep;
-//
-    movl     C(dc_iscale),%ecx           // fracstep
-    movl     C(centery),%eax
-    subl     %ebp,%eax
-    imul     %ecx,%eax
-    movl     C(dc_texturemid),%edx
-    subl     %eax,%edx
-     movl     %edx,%ebx
-     shrl     $16,%ebx          // frac int.
-     andl     $0x0000007f,%ebx
-     shll     $16,%edx          // y frac up
-
-     movl     %ecx,%ebp
-     shll     $16,%ebp          // fracstep f. up
-     shrl     $16,%ecx          // fracstep i. ->cl
-     andb     $0x7f,%cl
-
-    movl     C(dc_source),%esi
-
-//
-// lets rock :) !
-//
-
-    movl    C(pixelcount),%eax
-    movb    %al,%dh
-    shrl    $2,%eax
-    movb    %al,%ch     // quad count
-
-    testb   %ch, %ch
-    jz      vhnearlydone
-
-    movl    C(dc_colormap),%eax
-    decl    %edi                  //-----
-
-vhloop:
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     andb    $0x7f,%bl
-    incl    %edi                 //-----
-     movb    (%eax),%dh
-    movb    %dh,(%edi)           //-----
-
-     movb    (%esi,%ebx),%al      // fetch source texel
-    addl    %ebp,%edx
-     incl    %edi                //-----
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-    andb    $0x7f,%bl
-     movb    %dl,(%edi)          //-----
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-//    shll    $16,%edx
-     andb    $0x7f,%bl
-    incl    %edi                //-----
-     movb    (%eax),%dh
-    movb    %dh,(%edi)          //-----
-
-     movb    (%esi,%ebx),%al      // fetch source texel
-    addl    %ebp,%edx
-     incl    %edi               //-----
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-    andb    $0x7f,%bl
-     movb    %dl,(%edi)
-//     movl    %edx,(%edi)
-//    addl    $4,%edi
-
-    decb   %ch
-     jnz    vhloop
-
-vhnearlydone:
-//    movl    C(pixelcount)
-
-vhdone:
-    popl    %ebx
-    popl    %edi
-    popl    %esi
-    popl    %ebp
-    ret
-
-
-// --------------------------------------------------------------------------
-// Rotate a buffer 90 degree in clockwise order after horiz.col. draws
-// --------------------------------------------------------------------------
-
-#ifdef LINUX
-    .align 2
-#else
-    .align 5
-#endif
-.globl C(R_RotateBuffer)
-C(R_RotateBuffer):
-    pushl   %ebp
-    pushl   %esi
-    pushl   %edi
-    pushl   %ebx
-
-
-    movl    C(dc_source),%esi
-    movl    C(dc_colormap),%edi
-
-
-    movb    (%esi),%ah
-     addl    $200,%esi
-    movb    (%ebx),%al
-     addl    $200,%ebx
-    bswap    %eax
-    movb    (%esi),%ah
-     addl    $200,%esi
-    movb    (%ebx),%al
-     addl    $200,%ebx
-    movl    %eax,(%edi)
-     addl    $4,%edi
-
-
-    popl    %ebx
-    popl    %edi
-    popl    %esi
-    popl    %ebp
-    ret
-#endif
-
-//----------------------------------------------------------------------
-//13-02-98:
-//   R_DrawSkyColumn : same as R_DrawColumn but:
-//
-//            - wrap around 256 instead of 127.
-//   this is needed because we have a higher texture for mouselook,
-//   we need at least 200 lines for the sky.
-//
-//   NOTE: the sky should never wrap, so it could use a faster method.
-//         for the moment, we'll still use a wrapping method...
-//
-//  IT S JUST A QUICK CUT N PASTE, WAS NOT OPTIMISED AS IT SHOULD BE !!!
-//
-//----------------------------------------------------------------------
-
-#ifdef LINUX
-    .align 2
-#else
-    .align 5
-#endif
-.globl C(R_DrawSkyColumn_8)
-C(R_DrawSkyColumn_8):
-    pushl   %ebp
-    pushl   %esi
-    pushl   %edi
-    pushl   %ebx
-
-//
-// dest = ylookup[dc_yl] + columnofs[dc_x];
-//
-    movl     C(dc_yl),%ebp
-    movl     %ebp,%ebx
-    movl     C(ylookup)(,%ebx,4),%edi
-    movl     C(dc_x),%ebx
-    addl     C(columnofs)(,%ebx,4),%edi  // edi = dest
-
-//
-// pixelcount = yh - yl + 1
-//
-    movl     C(dc_yh),%eax
-    incl     %eax
-    subl     %ebp,%eax                   // pixel count
-    movl     %eax,C(pixelcount)          // save for final pixel
-    jle      vskydone                       // nothing to scale
-
-//
-// frac = dc_texturemid - (centery-dc_yl)*fracstep;
-//
-    movl     C(dc_iscale),%ecx           // fracstep
-    movl     C(centery),%eax
-    subl     %ebp,%eax
-    imul     %ecx,%eax
-    movl     C(dc_texturemid),%edx
-    subl     %eax,%edx
-     movl     %edx,%ebx
-     shrl     $16,%ebx          // frac int.
-     andl     $0x000000ff,%ebx
-     shll     $16,%edx          // y frac up
-
-     movl     %ecx,%ebp
-     shll     $16,%ebp          // fracstep f. up
-     shrl     $16,%ecx          // fracstep i. ->cl
-
-    movl     C(dc_source),%esi
-
-//
-// lets rock :) !
-//
-    movl    C(pixelcount),%eax
-    movb    %al,%dh
-    shrl    $2,%eax
-    movb    %al,%ch             // quad count
-    movl    C(dc_colormap),%eax
-    testb   $3,%dh
-    jz      v4skyquadloop
-
-//
-//  do un-even pixel
-//
-    testb   $1,%dh
-    jz      2f
-
-    movb    (%esi,%ebx),%al     // prep un-even loops
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-     movb    %dl,(%edi)           // output pixel
-    addl    C(vidwidth),%edi
-
-//
-//  do two non-quad-aligned pixels
-//
-2:
-    testb   $2,%dh
-    jz      3f
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-     movb    %dl,(%edi)           // output pixel
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    addl    C(vidwidth),%edi
-     movb    %dl,(%edi)           // output pixel
-
-    addl    C(vidwidth),%edi
-
-//
-//  test if there was at least 4 pixels
-//
-3:
-    testb   $0xFF,%ch           // test quad count
-    jz      vskydone
-
-//
-// ebp : ystep frac. upper 24 bits
-// edx : y     frac. upper 24 bits
-// ebx : y     i.    lower 7 bits,  masked for index
-// ecx : ch = counter, cl = y step i.
-// eax : colormap aligned 256
-// esi : source texture column
-// edi : dest screen
-//
-v4skyquadloop:
-//    .align  4
-vskyquadloop:
-    movb    (%esi,%ebx),%al     // prep loop
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    movb    %dl,(%edi)           // output pixel
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-p1b:    movb    %dl,0x12345678(%edi)
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-p2b:    movb    %dl,2*0x12345678(%edi)
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-p3b:    movb    %dl,3*0x12345678(%edi)
-
-p4b:    addl    $4*0x12345678,%edi
-
-    decb   %ch
-     jnz    vskyquadloop
-
-vskydone:
-    popl    %ebx                // restore register variables
-    popl    %edi
-    popl    %esi
-    popl    %ebp                // restore caller's stack frame pointer
-    ret
-
-
-
-//----------------------------------------------------------------------
-//
-// R_DrawSpan
-//
-// Horizontal texture mapping
-//
-//----------------------------------------------------------------------
-
-    .data
-
-ystep:          .long   0
-xstep:          .long   0
-C(texwidth):    .long   64      // texture width
-#if !defined( LINUX)
-    .text
-#endif
-#ifdef LINUX
-    .align 2
-#else
-    .align 4
-#endif
-.globl C(R_DrawSpan_8)
-C(R_DrawSpan_8):
-    pushl   %ebp                // preserve caller's stack frame pointer
-    pushl   %esi                // preserve register variables
-    pushl   %edi
-    pushl   %ebx
-
-
-//
-// find loop count
-//
-    movl    C(ds_x2),%eax
-    incl    %eax
-    subl    C(ds_x1),%eax               // pixel count
-    movl    %eax,C(pixelcount)          // save for final pixel
-    js      hdone                       // nothing to scale
-    shrl    $1,%eax                     // double pixel count
-    movl    %eax,C(loopcount)
-
-//
-// build composite position
-//
-    movl    C(ds_xfrac),%ebp
-    shll    $10,%ebp
-    andl    $0x0ffff0000,%ebp
-    movl    C(ds_yfrac),%eax
-    shrl    $6,%eax
-    andl    $0x0ffff,%eax
-    movl    C(ds_y),%edi
-    orl     %eax,%ebp
-
-    movl    C(ds_source),%esi
-
-//
-// calculate screen dest
-//
-
-    movl    C(ylookup)(,%edi,4),%edi
-    movl    C(ds_x1),%eax
-    addl    C(columnofs)(,%eax,4),%edi
-
-//
-// build composite step
-//
-    movl    C(ds_xstep),%ebx
-    shll    $10,%ebx
-    andl    $0x0ffff0000,%ebx
-    movl    C(ds_ystep),%eax
-    shrl    $6,%eax
-    andl    $0x0ffff,%eax
-    orl     %eax,%ebx
-
-    //movl        %eax,OFFSET hpatch1+2        // convice tasm to modify code...
-    movl    %ebx,hpatch1+2
-    //movl        %eax,OFFSET hpatch2+2        // convice tasm to modify code...
-    movl    %ebx,hpatch2+2
-    movl    %esi,hpatch3+2
-    movl    %esi,hpatch4+2
-// %eax      aligned colormap
-// %ebx      aligned colormap
-// %ecx,%edx  scratch
-// %esi      virtual source
-// %edi      moving destination pointer
-// %ebp      frac
-    movl    C(ds_colormap),%eax
-//    shld    $22,%ebp,%ecx           // begin calculating third pixel (y units)
-//    shld    $6,%ebp,%ecx            // begin calculating third pixel (x units)
-     movl    %ebp,%ecx
-    addl    %ebx,%ebp               // advance frac pointer
-     shrw    $10,%cx
-     roll    $6,%ecx
-    andl    $4095,%ecx              // finish calculation for third pixel
-//    shld    $22,%ebp,%edx           // begin calculating fourth pixel (y units)
-//    shld    $6,%ebp,%edx            // begin calculating fourth pixel (x units)
-     movl    %ebp,%edx
-     shrw    $10,%dx
-     roll    $6,%edx
-    addl    %ebx,%ebp               // advance frac pointer
-    andl    $4095,%edx              // finish calculation for fourth pixel
-    movl    %eax,%ebx
-    movb    (%esi,%ecx),%al         // get first pixel
-    movb    (%esi,%edx),%bl         // get second pixel
-    testl   $0x0fffffffe,C(pixelcount)
-    movb    (%eax),%dl             // color translate first pixel
-
-//    jnz hdoubleloop             // at least two pixels to map
-//    jmp hchecklast
-
-//    movw $0xf0f0,%dx //see visplanes start
-
-    jz      hchecklast
-    movb    (%ebx),%dh              // color translate second pixel
-    movl    C(loopcount),%esi
-//    .align  4
-hdoubleloop:
-//    shld    $22,%ebp,%ecx        // begin calculating third pixel (y units)
-//    shld    $6,%ebp,%ecx         // begin calculating third pixel (x units)
-    movl    %ebp,%ecx
-    shrw    $10,%cx
-    roll    $6,%ecx
-hpatch1:
-    addl    $0x012345678,%ebp    // advance frac pointer
-    movw    %dx,(%edi)           // write first pixel
-    andl    $4095,%ecx           // finish calculation for third pixel
-//    shld    $22,%ebp,%edx        // begin calculating fourth pixel (y units)
-//    shld    $6,%ebp,%edx         // begin calculating fourth pixel (x units)
-    movl    %ebp,%edx
-    shrw    $10,%dx
-    roll    $6,%edx
-hpatch3:
-    movb    0x012345678(%ecx),%al      // get third pixel
-//    movb    %bl,1(%edi)          // write second pixel
-    andl    $4095,%edx           // finish calculation for fourth pixel
-hpatch2:
-    addl    $0x012345678,%ebp    // advance frac pointer
-hpatch4:
-    movb    0x012345678(%edx),%bl      // get fourth pixel
-    movb    (%eax),%dl           // color translate third pixel
-    addl    $2,%edi              // advance to third pixel destination
-    decl    %esi                 // done with loop?
-    movb    (%ebx),%dh           // color translate fourth pixel
-    jnz hdoubleloop
-
-// check for final pixel
-hchecklast:
-    testl   $1,C(pixelcount)
-    jz      hdone
-    movb    %dl,(%edi)           // write final pixel
-
-hdone:
-    popl    %ebx                 // restore register variables
-    popl    %edi
-    popl    %esi
-    popl    %ebp                 // restore caller's stack frame pointer
-    ret
-
-
-//.endif
-
-
-//----------------------------------------------------------------------
-// R_DrawTransColumn
-//
-// Vertical column texture drawer, with transparency. Replaces Doom2's
-// 'fuzz' effect, which was not so beautiful.
-// Transparency is always impressive in some way, don't know why...
-//----------------------------------------------------------------------
-
-#ifdef LINUX
-    .align 2
-#else
-    .align 5
-#endif
-
-.globl C(R_DrawTranslucentColumn_8)
-C(R_DrawTranslucentColumn_8):
-    pushl   %ebp                // preserve caller's stack frame pointer
-    pushl   %esi                // preserve register variables
-    pushl   %edi
-    pushl   %ebx
-
-//
-// dest = ylookup[dc_yl] + columnofs[dc_x];
-//
-    movl     C(dc_yl),%ebp
-    movl     %ebp,%ebx
-    movl     C(ylookup)(,%ebx,4),%edi
-    movl     C(dc_x),%ebx
-    addl     C(columnofs)(,%ebx,4),%edi  // edi = dest
-
-//
-// pixelcount = yh - yl + 1
-//
-    movl     C(dc_yh),%eax
-    incl     %eax
-    subl     %ebp,%eax                   // pixel count
-    movl     %eax,C(pixelcount)          // save for final pixel
-    jle      vtdone                       // nothing to scale
-
-//
-// frac = dc_texturemid - (centery-dc_yl)*fracstep;
-//
-    movl     C(dc_iscale),%ecx           // fracstep
-    movl     C(centery),%eax
-    subl     %ebp,%eax
-    imul     %ecx,%eax
-    movl     C(dc_texturemid),%edx
-    subl     %eax,%edx
-    movl     %edx,%ebx
-
-    shrl     $16,%ebx          // frac int.
-    andl     $0x0000007f,%ebx
-    shll     $16,%edx          // y frac up
-
-    movl     %ecx,%ebp
-    shll     $16,%ebp          // fracstep f. up
-    shrl     $16,%ecx          // fracstep i. ->cl
-    andb     $0x7f,%cl
-    pushw    %cx
-    movl     %edx,%ecx
-    popw     %cx
-    movl     C(dc_colormap),%edx
-    movl     C(dc_source),%esi
-
-//
-// lets rock :) !
-//
-    movl    C(pixelcount),%eax
-    shrl    $2,%eax
-    testb   $0x03,C(pixelcount)
-    movb    %al,%ch             // quad count
-    movl    C(dc_transmap),%eax
-    jz      vt4quadloop
-//
-//  do un-even pixel
-//
-    testb   $1,C(pixelcount)
-    jz      2f
-
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-     addl    %ebp,%ecx
-    adcb    %cl,%bl
-     movb    (%edi),%al           // fetch dest  : index into colormap
-    andb    $0x7f,%bl
-     movb    (%eax),%dl
-    movb    (%edx), %dl          // use colormap now !
-    movb    %dl,(%edi)
-     addl    C(vidwidth),%edi
-//
-//  do two non-quad-aligned pixels
-//
-2:
-    testb   $2,C(pixelcount)
-    jz      3f
-
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-     addl    %ebp,%ecx
-    adcb    %cl,%bl
-     movb    (%edi),%al           // fetch dest  : index into colormap
-    andb    $0x7f,%bl
-     movb    (%eax),%dl
-    movb    (%edx), %dl          // use colormap now !
-    movb    %dl,(%edi)
-     addl    C(vidwidth),%edi
-
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-     addl    %ebp,%ecx
-    adcb    %cl,%bl
-     movb    (%edi),%al           // fetch dest  : index into colormap
-    andb    $0x7f,%bl
-     movb    (%eax),%dl
-    movb    (%edx), %dl          // use colormap now !
-    movb    %dl,(%edi)
-     addl    C(vidwidth),%edi
-
-//
-//  test if there was at least 4 pixels
-//
-3:
-    testb   $0xFF,%ch           // test quad count
-    jz      vtdone
-
-//
-// tystep : ystep frac. upper 24 bits
-// edx : upper 24 bit : colomap
-//  dl : tmp pixel to write
-// ebx : y     i.    lower 7 bits,  masked for index
-// ecx : y     frac. upper 16 bits
-// ecx : ch = counter, cl = y step i.
-// eax : transmap aligned 65535 (upper 16 bit)
-//  ah : background pixel (from the screen buffer)
-//  al : foreground pixel (from the texture)
-// esi : source texture column
-// ebp,edi : dest screen
-//
-vt4quadloop:
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-p5: movb    0x12345678(%edi),%al           // fetch dest  : index into colormap
-
-    movl    %ebp,C(tystep)
-    movl    %edi,%ebp
-    subl    C(vidwidth),%edi
-    jmp inloop
-//    .align  4
-vtquadloop:
-    addl    C(tystep),%ecx
-    adcb    %cl,%bl
-p6: addl    $2*0x12345678,%ebp
-    andb    $0x7f,%bl
-    movb    (%eax),%dl
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-    movb    (%edx), %dl          // use colormap now !
-    movb    %dl,(%edi)
-    movb    (%ebp),%al           // fetch dest  : index into colormap
-inloop:
-    addl    C(tystep),%ecx
-    adcb    %cl,%bl
-p7: addl    $2*0x12345678,%edi
-    andb    $0x7f,%bl
-    movb    (%eax),%dl
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-    movb    (%edx), %dl          // use colormap now !
-    movb    %dl,(%ebp)
-    movb    (%edi),%al           // fetch dest  : index into colormap
-
-    addl    C(tystep),%ecx
-    adcb    %cl,%bl
-p8: addl    $2*0x12345678,%ebp
-    andb    $0x7f,%bl
-    movb    (%eax),%dl
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-    movb    (%edx), %dl          // use colormap now !
-    movb    %dl,(%edi)
-    movb    (%ebp),%al           // fetch dest  : index into colormap
-
-    addl    C(tystep),%ecx
-    adcb    %cl,%bl
-p9: addl    $2*0x12345678,%edi
-    andb    $0x7f,%bl
-    movb    (%eax),%dl
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-    movb    (%edx), %dl          // use colormap now !
-    movb    %dl,(%ebp)
-    movb    (%edi),%al           // fetch dest  : index into colormap
-
-    decb   %ch
-     jnz    vtquadloop
-
-vtdone:
-    popl    %ebx                // restore register variables
-    popl    %edi
-    popl    %esi
-    popl    %ebp                // restore caller's stack frame pointer
-    ret
-
-#endif // ifdef USEASM
-
-
-
-//----------------------------------------------------------------------
-// R_DrawShadeColumn
-//
-//   for smoke..etc.. test.
-//----------------------------------------------------------------------
-
-#ifdef LINUX
-    .align 2
-#else
-    .align 5
-#endif
-.globl C(R_DrawShadeColumn_8)
-C(R_DrawShadeColumn_8):
-    pushl   %ebp                // preserve caller's stack frame pointer
-    pushl   %esi                // preserve register variables
-    pushl   %edi
-    pushl   %ebx
-
-//
-// dest = ylookup[dc_yl] + columnofs[dc_x];
-//
-    movl     C(dc_yl),%ebp
-    movl     %ebp,%ebx
-    movl     C(ylookup)(,%ebx,4),%edi
-    movl     C(dc_x),%ebx
-    addl     C(columnofs)(,%ebx,4),%edi  // edi = dest
-
-//
-// pixelcount = yh - yl + 1
-//
-    movl     C(dc_yh),%eax
-    incl     %eax
-    subl     %ebp,%eax                   // pixel count
-    movl     %eax,C(pixelcount)          // save for final pixel
-    jle      shdone                       // nothing to scale
-
-//
-// frac = dc_texturemid - (centery-dc_yl)*fracstep;
-//
-    movl     C(dc_iscale),%ecx           // fracstep
-    movl     C(centery),%eax
-    subl     %ebp,%eax
-    imul     %ecx,%eax
-    movl     C(dc_texturemid),%edx
-    subl     %eax,%edx
-     movl     %edx,%ebx
-     shrl     $16,%ebx          // frac int.
-     andl     $0x0000007f,%ebx
-     shll     $16,%edx          // y frac up
-
-     movl     %ecx,%ebp
-     shll     $16,%ebp          // fracstep f. up
-     shrl     $16,%ecx          // fracstep i. ->cl
-     andb     $0x7f,%cl
-
-    movl     C(dc_source),%esi
-
-//
-// lets rock :) !
-//
-    movl    C(pixelcount),%eax
-    movb    %al,%dh
-    shrl    $2,%eax
-    movb    %al,%ch             // quad count
-    movl    C(colormaps),%eax
-    testb   $0x03,%dh
-    jz      sh4quadloop
-
-//
-//  do un-even pixel
-//
-    testb   $1,%dh
-    jz      2f
-
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%edi),%al           // fetch dest  : index into colormap
-    andb    $0x7f,%bl
-     movb    (%eax),%dl
-    movb    %dl,(%edi)
-     addl    C(vidwidth),%edi
-
-//
-//  do two non-quad-aligned pixels
-//
-2:
-    testb   $2,%dh
-    jz      3f
-
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%edi),%al           // fetch dest  : index into colormap
-    andb    $0x7f,%bl
-     movb    (%eax),%dl
-    movb    %dl,(%edi)
-     addl    C(vidwidth),%edi
-
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%edi),%al           // fetch dest  : index into colormap
-    andb    $0x7f,%bl
-     movb    (%eax),%dl
-    movb    %dl,(%edi)
-     addl    C(vidwidth),%edi
-
-//
-//  test if there was at least 4 pixels
-//
-3:
-    testb   $0xFF,%ch           // test quad count
-    jz      shdone
-
-//
-// ebp : ystep frac. upper 24 bits
-// edx : y     frac. upper 24 bits
-// ebx : y     i.    lower 7 bits,  masked for index
-// ecx : ch = counter, cl = y step i.
-// eax : colormap aligned 256
-// esi : source texture column
-// edi : dest screen
-//
-sh4quadloop:
-    movb    $0x7f,%dh           // prep mask
-
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-sh5:    movb    0x12345678(%edi),%al           // fetch dest  : index into colormap
-
-    movl    %ebp,C(tystep)
-    movl    %edi,%ebp
-    subl    C(vidwidth),%edi
-    jmp shinloop
-//    .align  4
-shquadloop:
-    addl    C(tystep),%edx
-    adcb    %cl,%bl
-    andb    %dh,%bl
-sh6:    addl    $2*0x12345678,%ebp
-    movb    (%eax),%dl
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-    movb    %dl,(%edi)
-    movb    (%ebp),%al           // fetch dest  : index into colormap
-shinloop:
-    addl    C(tystep),%edx
-    adcb    %cl,%bl
-    andb    %dh,%bl
-sh7:    addl    $2*0x12345678,%edi
-    movb    (%eax),%dl
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-    movb    %dl,(%ebp)
-    movb    (%edi),%al           // fetch dest  : index into colormap
-
-    addl    C(tystep),%edx
-    adcb    %cl,%bl
-    andb    %dh,%bl
-sh8:    addl    $2*0x12345678,%ebp
-    movb    (%eax),%dl
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-    movb    %dl,(%edi)
-    movb    (%ebp),%al           // fetch dest  : index into colormap
-
-    addl    C(tystep),%edx
-    adcb    %cl,%bl
-    andb    %dh,%bl
-sh9:    addl    $2*0x12345678,%edi
-    movb    (%eax),%dl
-    movb    (%esi,%ebx),%ah      // fetch texel : colormap number
-    movb    %dl,(%ebp)
-    movb    (%edi),%al           // fetch dest  : index into colormap
-
-    decb   %ch
-     jnz    shquadloop
-
-shdone:
-    popl    %ebx                // restore register variables
-    popl    %edi
-    popl    %esi
-    popl    %ebp                // restore caller's stack frame pointer
-    ret
-
-
-
-//----------------------------------------------------------------------
-//
-//  R_DrawWaterColumn : basically it's just a copy of R_DrawColumn,
-//                      but it uses dc_colormap from dc_yl to dc_yw-1
-//                      then it uses dc_wcolormap from dc_yw to dc_yh
-//
-//  Thus, the 'underwater' part of the walls is remapped to 'water-like'
-//  colors.
-//
-//----------------------------------------------------------------------
-
-#ifdef LINUX
-    .align 2
-#else
-    .align 5
-#endif
-.globl C(R_DrawWaterColumn)
-C(R_DrawWaterColumn):
-    pushl   %ebp                // preserve caller's stack frame pointer
-    pushl   %esi                // preserve register variables
-    pushl   %edi
-    pushl   %ebx
-
-//
-// dest = ylookup[dc_yl] + columnofs[dc_x];
-//
-    movl     C(dc_yl),%ebp
-    movl     %ebp,%ebx
-    movl     C(ylookup)(,%ebx,4),%edi
-    movl     C(dc_x),%ebx
-    addl     C(columnofs)(,%ebx,4),%edi  // edi = dest
-
-//
-// pixelcount = yh - yl + 1
-//
-    movl     C(dc_yh),%eax
-    incl     %eax
-    subl     %ebp,%eax                   // pixel count
-    movl     %eax,C(pixelcount)          // save for final pixel
-    jle      wdone                       // nothing to scale
-
-//
-// frac = dc_texturemid - (centery-dc_yl)*fracstep;
-//
-    movl     C(dc_iscale),%ecx           // fracstep
-    movl     C(centery),%eax
-    subl     %ebp,%eax
-    imul     %ecx,%eax
-    movl     C(dc_texturemid),%edx
-    subl     %eax,%edx
-     movl     %edx,%ebx
-     shrl     $16,%ebx          // frac int.
-     andl     $0x0000007f,%ebx
-     shll     $16,%edx          // y frac up
-
-     movl     %ecx,%ebp
-     shll     $16,%ebp          // fracstep f. up
-     shrl     $16,%ecx          // fracstep i. ->cl
-     andb     $0x7f,%cl
-
-    movl     C(dc_source),%esi
-
-//
-// lets rock :) !
-//
-    movl    C(pixelcount),%eax
-    movb    %al,%dh
-    shrl    $2,%eax
-    movb    %al,%ch             // quad count
-    movl    C(dc_wcolormap),%eax
-    testb   $3,%dh
-    jz      w4quadloop
-
-//
-//  do un-even pixel
-//
-    testb   $1,%dh
-    jz      2f
-
-    movb    (%esi,%ebx),%al     // prep un-even loops
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    andb    $0x7f,%bl            // mask 0-127 texture index
-     movb    %dl,(%edi)           // output pixel
-    addl    C(vidwidth),%edi
-
-//
-//  do two non-quad-aligned pixels
-//
-2:
-    testb   $2,%dh
-    jz      3f
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    andb    $0x7f,%bl            // mask 0-127 texture index
-     movb    %dl,(%edi)           // output pixel
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    andb    $0x7f,%bl            // mask 0-127 texture index
-    addl    C(vidwidth),%edi
-     movb    %dl,(%edi)           // output pixel
-
-    addl    C(vidwidth),%edi
-
-//
-//  test if there was at least 4 pixels
-//
-3:
-    testb   $0xFF,%ch           // test quad count
-    jz      wdone
-
-//
-// ebp : ystep frac. upper 24 bits
-// edx : y     frac. upper 24 bits
-// ebx : y     i.    lower 7 bits,  masked for index
-// ecx : ch = counter, cl = y step i.
-// eax : colormap aligned 256
-// esi : source texture column
-// edi : dest screen
-//
-w4quadloop:
-    movb    $0x7f,%dh           // prep mask
-//    .align  4
-wquadloop:
-    movb    (%esi,%ebx),%al     // prep loop
-     addl    %ebp,%edx            // ypos f += ystep f
-    adcb    %cl,%bl              // ypos i += ystep i
-     movb    (%eax),%dl           // colormap texel
-    movb    %dl,(%edi)           // output pixel
-     andb    $0x7f,%bl            // mask 0-127 texture index
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-w1:    movb    %dl,0x12345678(%edi)
-     andb    $0x7f,%bl
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-w2:    movb    %dl,2*0x12345678(%edi)
-     andb    $0x7f,%bl
-
-    movb    (%esi,%ebx),%al      // fetch source texel
-     addl    %ebp,%edx
-    adcb    %cl,%bl
-     movb    (%eax),%dl
-w3:    movb    %dl,3*0x12345678(%edi)
-     andb    $0x7f,%bl
-
-w4:    addl    $4*0x12345678,%edi
-
-    decb   %ch
-     jnz    wquadloop
-
-wdone:
-    popl    %ebx                // restore register variables
-    popl    %edi
-    popl    %esi
-    popl    %ebp                // restore caller's stack frame pointer
-    ret
-
-
-
-
-
-
-
-//----------------------------------------------------------------------
-//
-//  R_DrawSpanNoWrap
-//
-//      Horizontal texture mapping, does not remap colors,
-//      neither needs to wrap around the source texture.
-//
-//      Thus, a special optimisation can be used...
-//
-//----------------------------------------------------------------------
-
-    .data
-
-advancetable:   .long   0, 0
-#if !defined( LINUX)
-    .text
-#endif
-#ifdef LINUX
-    .align 2
-#else
-    .align 4
-#endif
-.globl C(R_DrawSpanNoWrap)
-C(R_DrawSpanNoWrap):
-    pushl   %ebp                // preserve caller's stack frame pointer
-    pushl   %esi                // preserve register variables
-    pushl   %edi
-    pushl   %ebx
-
-//
-// find loop count
-//
-
-    movl    C(ds_x2),%eax
-    incl    %eax
-    subl    C(ds_x1),%eax               // pixel count
-    movl    %eax,C(pixelcount)          // save for final pixel
-    jle     htvdone                       // nothing to scale
-//    shrl    $1,%eax                     // double pixel count
-//    movl    %eax,C(loopcount)
-
-//
-// calculate screen dest
-//
-
-    movl    C(ds_y),%edi        //full destination start address
-
-//
-// set up advancetable
-//
-
-    movl    C(ds_xstep),%ebp
-    movl    C(ds_ystep),%ecx
-    movl    %ecx,%eax
-    movl    %ebp,%edx
-    sarl    $16,%edx            // xstep >>= 16;
-    movl    C(vidwidth),%ebx
-    sarl    $16,%eax            // ystep >>= 16;
-    jz      0f
-    imull   %ebx,%eax           // (ystep >> 16) * texwidth;
-0:
-    addl    %edx,%eax           // add in xstep
-                                // (ystep >> 16) * texwidth + (xstep >> 16);
-
-    movl    %eax,advancetable+4 // advance base in y
-    addl    %ebx,%eax           // ((ystep >> 16) + 1) * texwidth +
-                                //  (xstep >> 16);
-    movl    %eax,advancetable   // advance extra in y
-
-    shll    $16,%ebp            // left-justify xstep fractional part
-    movl    %ebp,xstep
-    shll    $16,%ecx            // left-justify ystep fractional part
-    movl    %ecx,ystep
-
-//
-// calculate the texture starting address
-//
-    movl    C(ds_source),%esi       // texture source
-
-     movl    C(ds_yfrac),%eax
-     movl    %eax,%edx
-     sarl    $16,%eax
-    movl    C(ds_xfrac),%ecx
-     imull   %ebx,%eax               // (yfrac >> 16) * texwidth
-    movl    %ecx,%ebx
-    sarl    $16,%ecx
-    movl    %ecx,%ebp
-     addl    %eax,%ebp               // source = (xfrac >> 16) +
-                                    //           ((yfrac >> 16) * texwidth);
-
-//
-//  esi : texture source
-//  edi : screen dest
-//  eax : colormap aligned on 256 boundary, hehehe...
-//  ebx : xfrac << 16
-//  ecx : used in loop, contains either 0 or -1, *4, offset into advancetable
-//  edx : yfrac << 16
-//  ebp : offset into texture
-//
-
-    shll    $16,%edx             // yfrac upper word, lower byte will be used
-    movl    C(ds_colormap),%eax
-    shll    $16,%ebx             // xfrac upper word, lower unused
-
-    movl    C(pixelcount),%ecx
-    shrl    $2,%ecx
-    movb    %cl,%dh             // quad pixels count
-
-    movl    C(pixelcount),%ecx
-    andl    $3,%ecx
-    jz      htvquadloop         // pixelcount is multiple of 4
-    decl    %ecx
-    jz      1f
-    decl    %ecx
-    jz      2f
-
-//
-//  do one to three pixels first
-//
-    addl    ystep,%edx          // yfrac += ystep
-   sbbl    %ecx,%ecx           // turn carry into 0 or -1 if set
-    movb    (%esi,%ebp),%al          // get texture pixel
-   addl    xstep,%ebx           // xfrac += xstep
-//    movb    (%eax),%dl           // pixel goes through colormap
-   adcl    advancetable+4(,%ecx,4),%ebp       // advance source
-    movb    %al,(%edi)           // write pixel dest
-
-   incl    %edi
-
-2:
-    addl    ystep,%edx          // yfrac += ystep
-   sbbl    %ecx,%ecx           // turn carry into 0 or -1 if set
-    movb    (%esi,%ebp),%al          // get texture pixel
-   addl    xstep,%ebx           // xfrac += xstep
-//    movb    (%eax),%dl           // pixel goes through colormap
-   adcl    advancetable+4(,%ecx,4),%ebp       // advance source
-    movb    %al,(%edi)           // write pixel dest
-
-   incl    %edi
-
-1:
-    addl    ystep,%edx          // yfrac += ystep
-   sbbl    %ecx,%ecx           // turn carry into 0 or -1 if set
-    movb    (%esi,%ebp),%al          // get texture pixel
-   addl    xstep,%ebx           // xfrac += xstep
-//    movb    (%eax),%dl           // pixel goes through colormap
-   adcl    advancetable+4(,%ecx,4),%ebp       // advance source
-    movb    %al,(%edi)           // write pixel dest
-
-   incl    %edi
-
-//
-//  test if there was at least 4 pixels
-//
-    testb   $0xFF,%dh
-    jz      htvdone
-
-//
-//  two pixels per loop
-// U
-//  V
-htvquadloop:
-    addl    ystep,%edx             // yfrac += ystep
-   sbbl    %ecx,%ecx               // turn carry into 0 or -1 if set
-    movb    (%esi,%ebp),%al        // get texture pixel
-   addl    xstep,%ebx              // xfrac += xstep
-//    movb    (%eax),%dl             // pixel goes through colormap
-   adcl    advancetable+4(,%ecx,4),%ebp       // advance source
-    movb    %al,(%edi)             // write pixel dest
-
-    addl    ystep,%edx
-   sbbl    %ecx,%ecx
-    movb    (%esi,%ebp),%al
-   addl    xstep,%ebx
-//    movb    (%eax),%dl
-   adcl    advancetable+4(,%ecx,4),%ebp
-    movb    %al,1(%edi)
-
-    addl    ystep,%edx
-   sbbl    %ecx,%ecx
-    movb    (%esi,%ebp),%al
-   addl    xstep,%ebx
-//    movb    (%eax),%dl
-   adcl    advancetable+4(,%ecx,4),%ebp
-    movb    %al,2(%edi)
-
-    addl    ystep,%edx
-   sbbl    %ecx,%ecx
-    movb    (%esi,%ebp),%al
-   addl    xstep,%ebx
-//    movb    (%eax),%dl
-   adcl    advancetable+4(,%ecx,4),%ebp
-    movb    %al,3(%edi)
-
-   addl    $4, %edi
-    incl    %ecx    //dummy
-
-   decb   %dh
-    jnz    htvquadloop          // paire dans V-pipe
-
-htvdone:
-    popl    %ebx                // restore register variables
-    popl    %edi
-    popl    %esi
-    popl    %ebp                // restore caller's stack frame pointer
-    ret
-
-
-//.endif
-
-#ifdef HORIZONTALDRAW
-// void R_RotateBuffere (void)
-
-#ifdef LINUX
-    .align 2
-#else
-    .align 4
-#endif
-.globl C(R_RotateBufferasm)
-C(R_RotateBufferasm):
-    pushl   %ebp                // preserve caller's stack frame pointer
-    pushl   %esi                // preserve register variables
-    pushl   %edi
-    pushl   %ebx
-
-    movl    C(dc_source),%esi
-    movl    C(dc_colormap),%edi
-
-    movl    $200,%edx
-ra2:
-    movl    $40,%ecx
-ra:
-    movb    -2*200(%esi),%al
-    movb    -6*200(%esi),%bl
-    movb    -3*200(%esi),%ah
-    movb    -7*200(%esi),%bh
-    shll    $16,%eax
-    shll    $16,%ebx
-    movb    (%esi),%al
-    movb    -4*200(%esi),%bl
-    movb    -1*200(%esi),%ah
-    movb    -5*200(%esi),%bh
-    movl    %eax,(%edi)
-    subl    $8*200,%esi
-    movl    %ebx,4(%edi)
-    addl    $8,%edi
-    decl    %ecx
-    jnz     ra
-
-    addl    $320*200+1,%esi      //32*480 passe a la ligne suivante
-//    addl    320-32,%edi
-
-    decl    %edx
-    jnz     ra2
-
-    pop   %ebp                // preserve caller's stack frame pointer
-    pop   %esi                // preserve register variables
-    pop   %edi
-    pop   %ebx
-    ret
-#endif
diff --git a/src/tmap_asm.s b/src/tmap_asm.s
deleted file mode 100644
index 3cd0f87cc..000000000
--- a/src/tmap_asm.s
+++ /dev/null
@@ -1,322 +0,0 @@
-// SONIC ROBO BLAST 2
-//-----------------------------------------------------------------------------
-// Copyright (C) 1998-2000 by DooM Legacy Team.
-// Copyright (C) 1999-2020 by Sonic Team Junior.
-//
-// This program is free software distributed under the
-// terms of the GNU General Public License, version 2.
-// See the 'LICENSE' file for more details.
-//-----------------------------------------------------------------------------
-/// \file  tmap_asm.s
-/// \brief ???
-
-//.comm _dc_colormap,4
-//.comm _dc_x,4
-//.comm _dc_yl,4
-//.comm _dc_yh,4
-//.comm _dc_iscale,4
-//.comm _dc_texturemid,4
-//.comm _dc_source,4
-//.comm _ylookup,4
-//.comm _columnofs,4
-//.comm _loopcount,4
-//.comm _pixelcount,4
-.data
-_pixelcount:
-.long 0x00000000
-_loopcount:
-.long 0x00000000
-.align 8
-_mmxcomm:
-.long 0x00000000
-.text
-
-        .align 4
-.globl _R_DrawColumn8_NOMMX
-_R_DrawColumn8_NOMMX:
-   pushl %ebp
-   pushl %esi
-   pushl %edi
-   pushl %ebx
-	movl _dc_yl,%edx
-	movl _dc_yh,%eax
-	subl %edx,%eax
-	leal 1(%eax),%ebx
-	testl %ebx,%ebx
-	jle rdc8ndone
-	movl _dc_x,%eax
-        movl _ylookup, %edi
-	movl (%edi,%edx,4),%esi
-	movl _columnofs, %edi
-	addl (%edi,%eax,4),%esi
-	movl _dc_iscale,%edi
-	movl %edx,%eax
-	imull %edi,%eax
-	movl _dc_texturemid,%ecx
-	addl %eax,%ecx
-
-	movl _dc_source,%ebp
-   xorl %edx, %edx
-   subl $0x12345678, %esi
-.globl rdc8nwidth1
-rdc8nwidth1:
-	.align 4,0x90
-rdc8nloop:
-	movl %ecx,%eax
-	shrl $16,%eax
-	addl %edi,%ecx
-	andl $127,%eax
-	addl $0x12345678,%esi
-.globl rdc8nwidth2
-rdc8nwidth2:
-	movb (%eax,%ebp),%dl
-	movl _dc_colormap,%eax
-	movb (%eax,%edx),%al
-	movb %al,(%esi)
-	decl %ebx
-	jne rdc8nloop
-rdc8ndone:
-   popl %ebx
-   popl %edi
-   popl %esi
-   popl %ebp
-   ret
-
-//
-// Optimised specifically for P54C/P55C (aka Pentium with/without MMX)
-// By ES 1998/08/01
-//
-
-.globl _R_DrawColumn_8_Pentium
-_R_DrawColumn_8_Pentium:
-	pushl %ebp
-        pushl %ebx
-	pushl %esi
-        pushl %edi
-	movl _dc_yl,%eax        // Top pixel
-	movl _dc_yh,%ebx        // Bottom pixel
-        movl _ylookup, %edi
-	movl (%edi,%ebx,4),%ecx
-	subl %eax,%ebx          // ebx=number of pixels-1
-	jl rdc8pdone            // no pixel to draw, done
-	jnz rdc8pmany
-	movl _dc_x,%edx         // Special case: only one pixel
-        movl _columnofs, %edi
-	addl (%edi,%edx,4),%ecx // dest pixel at (%ecx)
-	movl _dc_iscale,%esi
-	imull %esi,%eax
-	movl _dc_texturemid,%edi
-	addl %eax,%edi          // texture index in edi
-	movl _dc_colormap,%edx
-   	shrl $16, %edi
-   	movl _dc_source,%ebp
-	andl $127,%edi
-	movb (%edi,%ebp),%dl    // read texture pixel
-	movb (%edx),%al	        // lookup for light
-	movb %al,0(%ecx) 	// write it
-	jmp rdc8pdone		// done!
-.align 4, 0x90
-rdc8pmany:			// draw >1 pixel
-	movl _dc_x,%edx
-        movl _columnofs, %edi
-	movl (%edi,%edx,4),%edx
-	leal 0x12345678(%edx, %ecx), %edi  // edi = two pixels above bottom
-.globl rdc8pwidth5
-rdc8pwidth5:  // DeadBeef = -2*SCREENWIDTH
-        movl _dc_iscale,%edx	// edx = fracstep
-	imull %edx,%eax
-   	shll $9, %edx           // fixme: Should get 7.25 fix as input
-	movl _dc_texturemid,%ecx
-	addl %eax,%ecx          // ecx = frac
-	movl _dc_colormap,%eax  // eax = lighting/special effects LUT
-   	shll $9, %ecx
-   	movl _dc_source,%esi    // esi = source ptr
-
-	imull $0x12345678, %ebx // ebx = negative offset to pixel
-.globl rdc8pwidth6
-rdc8pwidth6:  // DeadBeef = -SCREENWIDTH
-
-// Begin the calculation of the two first pixels
-        leal (%ecx, %edx), %ebp
-	shrl $25, %ecx
-	movb (%esi, %ecx), %al
-	leal (%edx, %ebp), %ecx
-	shrl $25, %ebp
-        movb (%eax), %dl
-
-// The main loop
-rdc8ploop:
-	movb (%esi,%ebp), %al		// load 1
-        leal (%ecx, %edx), %ebp         // calc frac 3
-
-	shrl $25, %ecx                  // shift frac 2
-        movb %dl, 0x12345678(%edi, %ebx)// store 0
-.globl rdc8pwidth1
-rdc8pwidth1:  // DeadBeef = 2*SCREENWIDTH
-
-        movb (%eax), %al                // lookup 1
-
-        movb %al, 0x12345678(%edi, %ebx)// store 1
-.globl rdc8pwidth2
-rdc8pwidth2:  // DeadBeef = 3*SCREENWIDTH
-        movb (%esi, %ecx), %al          // load 2
-
-        leal (%ebp, %edx), %ecx         // calc frac 4
-
-        shrl $25, %ebp                  // shift frac 3
-        movb (%eax), %dl                // lookup 2
-
-        addl $0x12345678, %ebx          // counter
-.globl rdc8pwidth3
-rdc8pwidth3:  // DeadBeef = 2*SCREENWIDTH
-        jl rdc8ploop                    // loop
-
-// End of loop. Write extra pixel or just exit.
-        jnz rdc8pdone
-        movb %dl, 0x12345678(%edi, %ebx)// Write odd pixel
-.globl rdc8pwidth4
-rdc8pwidth4:  // DeadBeef = 2*SCREENWIDTH
-
-rdc8pdone:
-
-        popl %edi
-	popl %esi
-        popl %ebx
-	popl %ebp
-        ret
-
-//
-// MMX asm version, optimised for K6
-// By ES 1998/07/05
-//
-
-.globl _R_DrawColumn_8_K6_MMX
-_R_DrawColumn_8_K6_MMX:
-	pushl %ebp
-        pushl %ebx
-	pushl %esi
-        pushl %edi
-
-        movl %esp, %eax // Push 8 or 12, so that (%esp) gets aligned by 8
-        andl $7,%eax
-        addl $8,%eax
-        movl %eax, _mmxcomm // Temp storage in mmxcomm: (%esp) is used instead
-        subl %eax,%esp
-
-	movl _dc_yl,%edx        // Top pixel
-	movl _dc_yh,%ebx        // Bottom pixel
-        movl _ylookup, %edi
-	movl (%edi,%ebx,4),%ecx
-	subl %edx,%ebx         // ebx=number of pixels-1
-	jl 0x12345678            // no pixel to draw, done
-.globl rdc8moffs1
-rdc8moffs1:
-	jnz rdc8mmany
-	movl _dc_x,%eax         // Special case: only one pixel
-        movl _columnofs, %edi
-	addl (%edi,%eax,4),%ecx  // dest pixel at (%ecx)
-	movl _dc_iscale,%esi
-	imull %esi,%edx
-	movl _dc_texturemid,%edi
-	addl %edx,%edi         // texture index in edi
-	movl _dc_colormap,%edx
-   	shrl $16, %edi
-   	movl _dc_source,%ebp
-	andl $127,%edi
-	movb (%edi,%ebp),%dl  // read texture pixel
-	movb (%edx),%al	 // lookup for light
-	movb %al,0(%ecx) 	 // write it
-	jmp rdc8mdone		 // done!
-.globl rdc8moffs2
-rdc8moffs2:
-.align 4, 0x90
-rdc8mmany:			 // draw >1 pixel
-	movl _dc_x,%eax
-        movl _columnofs, %edi
-	movl (%edi,%eax,4),%eax
-	leal 0x12345678(%eax, %ecx), %esi  // esi = two pixels above bottom
-.globl rdc8mwidth3
-rdc8mwidth3:  // DeadBeef = -2*SCREENWIDTH
-        movl _dc_iscale,%ecx	 // ecx = fracstep
-	imull %ecx,%edx
-   	shll $9, %ecx           // fixme: Should get 7.25 fix as input
-	movl _dc_texturemid,%eax
-	addl %edx,%eax         // eax = frac
-	movl _dc_colormap,%edx  // edx = lighting/special effects LUT
-   	shll $9, %eax
-	leal (%ecx, %ecx), %edi
-   	movl _dc_source,%ebp    // ebp = source ptr
-	movl %edi, 0(%esp)     // Start moving frac and fracstep to MMX regs
-
-	imull $0x12345678, %ebx  // ebx = negative offset to pixel
-.globl rdc8mwidth5
-rdc8mwidth5:  // DeadBeef = -SCREENWIDTH
-
-	movl %edi, 4(%esp)
-	leal (%eax, %ecx), %edi
-	movq 0(%esp), %mm1     // fracstep:fracstep in mm1
-	movl %eax, 0(%esp)
-	shrl $25, %eax
-	movl %edi, 4(%esp)
-	movzbl (%ebp, %eax), %eax
-	movq 0(%esp), %mm0     // frac:frac in mm0
-
-	paddd %mm1, %mm0
-	shrl $25, %edi
-	movq %mm0, %mm2
-	psrld $25, %mm2         // texture index in mm2
-	paddd %mm1, %mm0
-	movq %mm2, 0(%esp)
-
-.globl rdc8mloop
-rdc8mloop:                      		// The main loop
-	movq %mm0, %mm2                    // move 4-5 to temp reg
-	movzbl (%ebp, %edi), %edi 		// read 1
-
-	psrld $25, %mm2 			// shift 4-5
-	movb (%edx,%eax), %cl 		// lookup 0
-
-	movl 0(%esp), %eax 			// load 2
-	addl $0x12345678, %ebx 		// counter
-.globl rdc8mwidth2
-rdc8mwidth2:  // DeadBeef = 2*SCREENWIDTH
-
-	movb %cl, (%esi, %ebx)		// write 0
-	movb (%edx,%edi), %ch 		// lookup 1
-
-	movb %ch, 0x12345678(%esi, %ebx) 	// write 1
-.globl rdc8mwidth1
-rdc8mwidth1:  // DeadBeef = SCREENWIDTH
-	movl 4(%esp), %edi			// load 3
-
-	paddd %mm1, %mm0 			// frac 6-7
-	movzbl (%ebp, %eax), %eax 		// lookup 2
-
-	movq %mm2, 0(%esp) 		     // store texture index 4-5
-	jl rdc8mloop
-
-	jnz rdc8mno_odd
-	movb (%edx,%eax), %cl  // write the last odd pixel
-	movb %cl, 0x12345678(%esi)
-.globl rdc8mwidth4
-rdc8mwidth4:  // DeadBeef = 2*SCREENWIDTH
-rdc8mno_odd:
-
-.globl rdc8mdone
-rdc8mdone:
-        emms
-
-        addl _mmxcomm, %esp
-        popl %edi
-	popl %esi
-        popl %ebx
-	popl %ebp
-        ret
-
-// Need some extra space to align run-time
-.globl R_DrawColumn_8_K6_MMX_end
-R_DrawColumn_8_K6_MMX_end:
-nop;nop;nop;nop;nop;nop;nop;nop;
-nop;nop;nop;nop;nop;nop;nop;nop;
-nop;nop;nop;nop;nop;nop;nop;nop;
-nop;nop;nop;nop;nop;nop;nop;
diff --git a/src/tmap_mmx.nas b/src/tmap_mmx.nas
deleted file mode 100644
index 15b97499d..000000000
--- a/src/tmap_mmx.nas
+++ /dev/null
@@ -1,674 +0,0 @@
-;; SONIC ROBO BLAST 2
-;;-----------------------------------------------------------------------------
-;; Copyright (C) 1998-2000 by DOSDOOM.
-;; Copyright (C) 2010-2020 by Sonic Team Junior.
-;;
-;; This program is free software distributed under the
-;; terms of the GNU General Public License, version 2.
-;; See the 'LICENSE' file for more details.
-;;-----------------------------------------------------------------------------
-;; FILE:
-;;      tmap_mmx.nas
-;; DESCRIPTION:
-;;      Assembler optimised rendering code for software mode, using SIMD
-;;      instructions.
-;;      Draw wall columns.
-
-
-[BITS 32]
-
-%define FRACBITS 16
-%define TRANSPARENTPIXEL 255
-
-%ifdef LINUX
-%macro cextern 1
-[extern %1]
-%endmacro
-
-%macro cglobal 1
-[global %1]
-%endmacro
-
-%else
-%macro cextern 1
-%define %1 _%1
-[extern %1]
-%endmacro
-
-%macro cglobal 1
-%define %1 _%1
-[global %1]
-%endmacro
-
-%endif
-
-
-; The viddef_s structure. We only need the width field.
-struc viddef_s
-		resb 12
-.width: resb 4
-		resb 44
-endstruc
-
-
-;; externs
-;; columns
-cextern dc_colormap
-cextern dc_x
-cextern dc_yl
-cextern dc_yh
-cextern dc_iscale
-cextern dc_texturemid
-cextern dc_texheight
-cextern dc_source
-cextern dc_hires
-cextern centery
-cextern centeryfrac
-cextern dc_transmap
-
-cextern R_DrawColumn_8_ASM
-cextern R_Draw2sMultiPatchColumn_8_ASM
-
-;; spans
-cextern nflatshiftup
-cextern nflatxshift
-cextern nflatyshift
-cextern nflatmask
-cextern ds_xfrac
-cextern ds_yfrac
-cextern ds_xstep
-cextern ds_ystep
-cextern ds_x1
-cextern ds_x2
-cextern ds_y
-cextern ds_source
-cextern ds_colormap
-
-cextern ylookup
-cextern columnofs
-cextern vid
-
-[SECTION .data]
-
-nflatmask64		dq		0
-
-
-[SECTION .text]
-
-;;----------------------------------------------------------------------
-;;
-;; R_DrawColumn : 8bpp column drawer
-;;
-;; MMX column drawer.
-;;
-;;----------------------------------------------------------------------
-;; eax = accumulator
-;; ebx = colormap
-;; ecx = count
-;; edx = accumulator
-;; esi = source
-;; edi = dest
-;; ebp = vid.width
-;; mm0 = accumulator
-;; mm1 = heightmask, twice
-;; mm2 = 2 * fracstep, twice
-;; mm3 = pair of consecutive fracs
-;;----------------------------------------------------------------------
-
-
-cglobal R_DrawColumn_8_MMX
-R_DrawColumn_8_MMX:
-		push		ebp						;; preserve caller's stack frame pointer
-		push		esi						;; preserve register variables
-		push		edi
-		push		ebx
-
-;;
-;; Our algorithm requires that the texture height be a power of two.
-;; If not, fall back to the non-MMX drawer.
-;;
-.texheightcheck:
-		mov			edx, [dc_texheight]
-		sub			edx, 1					;; edx = heightmask
-		test		edx, [dc_texheight]
-		jnz			near .usenonMMX
-
-		mov			ebp, edx				;; Keep a copy of heightmask in a
-											;; GPR for the time being.
-
-;;
-;; Fill mm1 with heightmask
-;;
-		movd		mm1, edx				;; low dword = heightmask
-		punpckldq	mm1, mm1				;; copy low dword to high dword
-
-;;
-;; dest = ylookup[dc_yl] + columnofs[dc_x];
-;;
-		mov			eax, [dc_yl]
-		mov			edi, [ylookup+eax*4]
-		mov			ebx, [dc_x]
-		add			edi, [columnofs+ebx*4]	;; edi = dest
-
-
-;;
-;; pixelcount = yh - yl + 1
-;;
-		mov			ecx, [dc_yh]
-		add			ecx, 1
-		sub			ecx, eax				;; pixel count
-		jle			near .done				;; nothing to scale
-
-;;
-;; fracstep = dc_iscale;
-;;
-		movd		mm2, [dc_iscale]		;; fracstep in low dword
-		punpckldq	mm2, mm2				;; copy to high dword
-
-		mov			ebx, [dc_colormap]
-		mov			esi, [dc_source]
-
-;;
-;; frac = (dc_texturemid + FixedMul((dc_yl << FRACBITS) - centeryfrac, fracstep));
-;;
-											;; eax == dc_yl already
-		shl			eax, FRACBITS
-		sub			eax, [centeryfrac]
-		imul		dword [dc_iscale]
-		shrd		eax, edx, FRACBITS
-		add			eax, [dc_texturemid]
-
-;;
-;; if (dc_hires) frac = 0;
-;;
-		test		byte [dc_hires], 0x01
-		jz			.mod2
-		xor			eax, eax
-
-
-;;
-;; Do mod-2 pixel.
-;;
-.mod2:
-		test		ecx, 1
-		jz			.pairprepare
-		mov			edx, eax				;; edx = frac
-		add			eax, [dc_iscale]		;; eax += fracstep
-		sar			edx, FRACBITS
-		and			edx, ebp				;; edx &= heightmask
-		movzx		edx, byte [esi + edx]
-		movzx		edx, byte [ebx + edx]
-		mov			[edi], dl
-
-		add			edi, [vid + viddef_s.width]
-		sub			ecx, 1
-		jz			.done
-
-.pairprepare:
-;;
-;; Prepare for the main loop.
-;;
-		movd		mm3, eax				;; Low dword = frac
-		movq		mm4, mm3				;; Copy to intermediate register
-		paddd		mm4, mm2				;; dwords of mm4 += fracstep
-		punpckldq	mm3, mm4				;; Low dword = first frac, high = second
-		pslld		mm2, 1					;; fracstep *= 2
-
-;;
-;; ebp = vid.width
-;;
-		mov			ebp, [vid + viddef_s.width]
-
-		align		16
-.pairloop:
-		movq		mm0, mm3				;; 3B 1u.
-		psrad		mm0, FRACBITS			;; 4B 1u.
-		pand		mm0, mm1				;; 3B 1u. frac &= heightmask
-		paddd		mm3, mm2				;; 3B 1u. frac += fracstep
-
-		movd		eax, mm0				;; 3B 1u. Get first frac
-;; IFETCH boundary
-		movzx		eax, byte [esi + eax]	;; 4B 1u. Texture map
-		movzx		eax, byte [ebx + eax]	;; 4B 1u. Colormap
-
-		punpckhdq	mm0, mm0				;; 3B 1(2)u. low dword = high dword
-		movd		edx, mm0				;; 3B 1u. Get second frac
-		mov			[edi], al				;; 2B 1(2)u. First pixel
-;; IFETCH boundary
-
-		movzx		edx, byte [esi + edx]	;; 4B 1u. Texture map
-		movzx		edx, byte [ebx + edx]	;; 4B 1u. Colormap
-		mov			[edi + 1*ebp], dl		;; 3B 1(2)u. Second pixel
-
-		lea			edi, [edi + 2*ebp]		;; 3B 1u. edi += 2 * vid.width
-;; IFETCH boundary
-		sub			ecx, 2					;; 3B 1u. count -= 2
-		jnz			.pairloop				;; 2B 1u. if(count != 0) goto .pairloop
-
-
-.done:
-;;
-;; Clear MMX state, or else FPU operations will go badly awry.
-;;
-		emms
-
-		pop			ebx
-		pop			edi
-		pop			esi
-		pop			ebp
-		ret
-
-.usenonMMX:
-		call		R_DrawColumn_8_ASM
-		jmp			.done
-
-
-;;----------------------------------------------------------------------
-;;
-;; R_Draw2sMultiPatchColumn : Like R_DrawColumn, but omits transparent
-;;                            pixels.
-;;
-;; MMX column drawer.
-;;
-;;----------------------------------------------------------------------
-;; eax = accumulator
-;; ebx = colormap
-;; ecx = count
-;; edx = accumulator
-;; esi = source
-;; edi = dest
-;; ebp = vid.width
-;; mm0 = accumulator
-;; mm1 = heightmask, twice
-;; mm2 = 2 * fracstep, twice
-;; mm3 = pair of consecutive fracs
-;;----------------------------------------------------------------------
-
-
-cglobal R_Draw2sMultiPatchColumn_8_MMX
-R_Draw2sMultiPatchColumn_8_MMX:
-		push		ebp						;; preserve caller's stack frame pointer
-		push		esi						;; preserve register variables
-		push		edi
-		push		ebx
-
-;;
-;; Our algorithm requires that the texture height be a power of two.
-;; If not, fall back to the non-MMX drawer.
-;;
-.texheightcheck:
-		mov			edx, [dc_texheight]
-		sub			edx, 1					;; edx = heightmask
-		test		edx, [dc_texheight]
-		jnz			near .usenonMMX
-
-		mov			ebp, edx				;; Keep a copy of heightmask in a
-											;; GPR for the time being.
-
-;;
-;; Fill mm1 with heightmask
-;;
-		movd		mm1, edx				;; low dword = heightmask
-		punpckldq	mm1, mm1				;; copy low dword to high dword
-
-;;
-;; dest = ylookup[dc_yl] + columnofs[dc_x];
-;;
-		mov			eax, [dc_yl]
-		mov			edi, [ylookup+eax*4]
-		mov			ebx, [dc_x]
-		add			edi, [columnofs+ebx*4]	;; edi = dest
-
-
-;;
-;; pixelcount = yh - yl + 1
-;;
-		mov			ecx, [dc_yh]
-		add			ecx, 1
-		sub			ecx, eax				;; pixel count
-		jle			near .done				;; nothing to scale
-;;
-;; fracstep = dc_iscale;
-;;
-		movd		mm2, [dc_iscale]		;; fracstep in low dword
-		punpckldq	mm2, mm2				;; copy to high dword
-
-		mov			ebx, [dc_colormap]
-		mov			esi, [dc_source]
-
-;;
-;; frac = (dc_texturemid + FixedMul((dc_yl << FRACBITS) - centeryfrac, fracstep));
-;;
-											;; eax == dc_yl already
-		shl			eax, FRACBITS
-		sub			eax, [centeryfrac]
-		imul		dword [dc_iscale]
-		shrd		eax, edx, FRACBITS
-		add			eax, [dc_texturemid]
-
-;;
-;; if (dc_hires) frac = 0;
-;;
-		test		byte [dc_hires], 0x01
-		jz			.mod2
-		xor			eax, eax
-
-
-;;
-;; Do mod-2 pixel.
-;;
-.mod2:
-		test		ecx, 1
-		jz			.pairprepare
-		mov			edx, eax				;; edx = frac
-		add			eax, [dc_iscale]		;; eax += fracstep
-		sar			edx, FRACBITS
-		and			edx, ebp				;; edx &= heightmask
-		movzx		edx, byte [esi + edx]
-		cmp			dl, TRANSPARENTPIXEL
-		je			.nextmod2
-		movzx		edx, byte [ebx + edx]
-		mov			[edi], dl
-
-.nextmod2:
-		add			edi, [vid + viddef_s.width]
-		sub			ecx, 1
-		jz			.done
-
-.pairprepare:
-;;
-;; Prepare for the main loop.
-;;
-		movd		mm3, eax				;; Low dword = frac
-		movq		mm4, mm3				;; Copy to intermediate register
-		paddd		mm4, mm2				;; dwords of mm4 += fracstep
-		punpckldq	mm3, mm4				;; Low dword = first frac, high = second
-		pslld		mm2, 1					;; fracstep *= 2
-
-;;
-;; ebp = vid.width
-;;
-		mov			ebp, [vid + viddef_s.width]
-
-		align		16
-.pairloop:
-		movq		mm0, mm3				;; 3B 1u.
-		psrad		mm0, FRACBITS			;; 4B 1u.
-		pand		mm0, mm1				;; 3B 1u. frac &= heightmask
-		paddd		mm3, mm2				;; 3B 1u. frac += fracstep
-
-		movd		eax, mm0				;; 3B 1u. Get first frac
-;; IFETCH boundary
-		movzx		eax, byte [esi + eax]	;; 4B 1u. Texture map
-		punpckhdq	mm0, mm0				;; 3B 1(2)u. low dword = high dword
-		movd		edx, mm0				;; 3B 1u. Get second frac
-		cmp			al, TRANSPARENTPIXEL	;; 2B 1u.
-		je			.secondinpair			;; 2B 1u.
-;; IFETCH boundary
-		movzx		eax, byte [ebx + eax]	;; 4B 1u. Colormap
-		mov			[edi], al				;; 2B 1(2)u. First pixel
-
-.secondinpair:
-		movzx		edx, byte [esi + edx]	;; 4B 1u. Texture map
-		cmp			dl, TRANSPARENTPIXEL	;; 2B 1u.
-		je			.nextpair				;; 2B 1u.
-;; IFETCH boundary
-		movzx		edx, byte [ebx + edx]	;; 4B 1u. Colormap
-		mov			[edi + 1*ebp], dl		;; 3B 1(2)u. Second pixel
-
-.nextpair:
-		lea			edi, [edi + 2*ebp]		;; 3B 1u. edi += 2 * vid.width
-		sub			ecx, 2					;; 3B 1u. count -= 2
-		jnz			.pairloop				;; 2B 1u. if(count != 0) goto .pairloop
-
-
-.done:
-;;
-;; Clear MMX state, or else FPU operations will go badly awry.
-;;
-		emms
-
-		pop			ebx
-		pop			edi
-		pop			esi
-		pop			ebp
-		ret
-
-.usenonMMX:
-		call		R_Draw2sMultiPatchColumn_8_ASM
-		jmp			.done
-
-
-;;----------------------------------------------------------------------
-;;
-;; R_DrawSpan : 8bpp span drawer
-;;
-;; MMX span drawer.
-;;
-;;----------------------------------------------------------------------
-;; eax = accumulator
-;; ebx = colormap
-;; ecx = count
-;; edx = accumulator
-;; esi = source
-;; edi = dest
-;; ebp = two pixels
-;; mm0 = accumulator
-;; mm1 = xposition
-;; mm2 = yposition
-;; mm3 = 2 * xstep
-;; mm4 = 2 * ystep
-;; mm5 = nflatxshift
-;; mm6 = nflatyshift
-;; mm7 = accumulator
-;;----------------------------------------------------------------------
-
-cglobal R_DrawSpan_8_MMX
-R_DrawSpan_8_MMX:
-		push		ebp						;; preserve caller's stack frame pointer
-		push		esi						;; preserve register variables
-		push		edi
-		push		ebx
-
-;;
-;; esi = ds_source
-;; ebx = ds_colormap
-;;
-		mov			esi, [ds_source]
-		mov			ebx, [ds_colormap]
-
-;;
-;; edi = ylookup[ds_y] + columnofs[ds_x1]
-;;
-		mov			eax, [ds_y]
-		mov			edi, [ylookup + eax*4]
-		mov			edx, [ds_x1]
-		add			edi, [columnofs + edx*4]
-
-;;
-;; ecx = ds_x2 - ds_x1 + 1
-;;
-		mov			ecx, [ds_x2]
-		sub			ecx, edx
-		add			ecx, 1
-
-;;
-;; Needed for fracs and steps
-;;
-		movd		mm7, [nflatshiftup]
-
-;;
-;; mm3 = xstep
-;;
-		movd		mm3, [ds_xstep]
-		pslld		mm3, mm7
-		punpckldq	mm3, mm3
-
-;;
-;; mm4 = ystep
-;;
-		movd		mm4, [ds_ystep]
-		pslld		mm4, mm7
-		punpckldq	mm4, mm4
-
-;;
-;; mm1 = pair of consecutive xpositions
-;;
-		movd		mm1, [ds_xfrac]
-		pslld		mm1, mm7
-		movq		mm6, mm1
-		paddd		mm6, mm3
-		punpckldq	mm1, mm6
-
-;;
-;; mm2 = pair of consecutive ypositions
-;;
-		movd		mm2, [ds_yfrac]
-		pslld		mm2, mm7
-		movq		mm6, mm2
-		paddd		mm6, mm4
-		punpckldq	mm2, mm6
-
-;;
-;; mm5 = nflatxshift
-;; mm6 = nflatyshift
-;;
-		movd		mm5, [nflatxshift]
-		movd		mm6, [nflatyshift]
-
-;;
-;; Mask is in memory due to lack of registers.
-;;
-		mov			eax, [nflatmask]
-		mov			[nflatmask64], eax
-		mov			[nflatmask64 + 4], eax
-
-
-;;
-;; Go until we reach a dword boundary.
-;;
-.unaligned:
-		test		edi, 3
-		jz			.alignedprep
-.stragglers:
-		cmp			ecx, 0
-		je			.done					;; If ecx == 0, we're finished.
-
-;;
-;; eax = ((yposition >> nflatyshift) & nflatmask) | (xposition >> nflatxshift)
-;;
-		movq		mm0, mm1				;; mm0 = xposition
-		movq		mm7, mm2				;; mm7 = yposition
-		paddd		mm1, mm3				;; xposition += xstep (once!)
-		paddd		mm2, mm4				;; yposition += ystep (once!)
-		psrld		mm0, mm5				;; shift
-		psrld		mm7, mm6				;; shift
-		pand		mm7, [nflatmask64]		;; mask
-		por			mm0, mm7				;; or x and y together
-
-		movd		eax, mm0				;; eax = index of first pixel
-		movzx		eax, byte [esi + eax]	;; al = source[eax]
-		movzx		eax, byte [ebx + eax]	;; al = colormap[al]
-
-		mov			[edi], al
-		add			edi, 1
-
-		sub			ecx, 1
-		jmp			.unaligned
-
-
-.alignedprep:
-;;
-;; We can double the steps now.
-;;
-		pslld		mm3, 1
-		pslld		mm4, 1
-
-
-;;
-;; Generate chunks of four pixels.
-;;
-.alignedloop:
-
-;;
-;; Make sure we have at least four pixels.
-;;
-		cmp			ecx, 4
-		jl			.prestragglers
-
-;;
-;; First two pixels.
-;;
-		movq		mm0, mm1				;; mm0 = xposition
-		movq		mm7, mm2				;; mm7 = yposition
-		paddd		mm1, mm3				;; xposition += xstep
-		paddd		mm2, mm4				;; yposition += ystep
-		psrld		mm0, mm5				;; shift
-		psrld		mm7, mm6				;; shift
-		pand		mm7, [nflatmask64]		;; mask
-		por			mm0, mm7				;; or x and y together
-
-		movd		eax, mm0				;; eax = index of first pixel
-		movzx		eax, byte [esi + eax]	;; al = source[eax]
-		movzx		ebp, byte [ebx + eax]	;; ebp = colormap[al]
-
-		punpckhdq	mm0, mm0				;; both dwords = high dword
-		movd		eax, mm0				;; eax = index of second pixel
-		movzx		eax, byte [esi + eax]	;; al = source[eax]
-		movzx		eax, byte [ebx + eax]	;; al = colormap[al]
-		shl			eax, 8					;; get pixel in right byte
-		or			ebp, eax				;; put pixel in ebp
-
-;;
-;; Next two pixels.
-;;
-		movq		mm0, mm1				;; mm0 = xposition
-		movq		mm7, mm2				;; mm7 = yposition
-		paddd		mm1, mm3				;; xposition += xstep
-		paddd		mm2, mm4				;; yposition += ystep
-		psrld		mm0, mm5				;; shift
-		psrld		mm7, mm6				;; shift
-		pand		mm7, [nflatmask64]		;; mask
-		por			mm0, mm7				;; or x and y together
-
-		movd		eax, mm0				;; eax = index of third pixel
-		movzx		eax, byte [esi + eax]	;; al = source[eax]
-		movzx		eax, byte [ebx + eax]	;; al = colormap[al]
-		shl			eax, 16					;; get pixel in right byte
-		or			ebp, eax				;; put pixel in ebp
-
-		punpckhdq	mm0, mm0				;; both dwords = high dword
-		movd		eax, mm0				;; eax = index of second pixel
-		movzx		eax, byte [esi + eax]	;; al = source[eax]
-		movzx		eax, byte [ebx + eax]	;; al = colormap[al]
-		shl			eax, 24					;; get pixel in right byte
-		or			ebp, eax				;; put pixel in ebp
-
-;;
-;; Write pixels.
-;;
-		mov			[edi], ebp
-		add			edi, 4
-
-		sub			ecx, 4
-		jmp			.alignedloop
-
-.prestragglers:
-;;
-;; Back to one step at a time.
-;;
-		psrad		mm3, 1
-		psrad		mm4, 1
-		jmp			.stragglers
-
-.done:
-;;
-;; Clear MMX state, or else FPU operations will go badly awry.
-;;
-		emms
-
-		pop			ebx
-		pop			edi
-		pop			esi
-		pop			ebp
-		ret
diff --git a/src/tmap_vc.nas b/src/tmap_vc.nas
deleted file mode 100644
index 49eb21a6d..000000000
--- a/src/tmap_vc.nas
+++ /dev/null
@@ -1,48 +0,0 @@
-;; SONIC ROBO BLAST 2
-;;-----------------------------------------------------------------------------
-;; Copyright (C) 1998-2000 by DooM Legacy Team.
-;; Copyright (C) 1999-2020 by Sonic Team Junior.
-;;
-;; This program is free software distributed under the
-;; terms of the GNU General Public License, version 2.
-;; See the 'LICENSE' file for more details.
-;;-----------------------------------------------------------------------------
-;; FILE:
-;;      tmap_vc.nas
-;; DESCRIPTION:
-;;      Assembler optimised math code for Visual C++.
-
-
-[BITS 32]
-
-%macro cglobal 1
-%define %1 _%1
-[global %1]
-%endmacro
-
-[SECTION .text write]
-
-;----------------------------------------------------------------------------
-;fixed_t FixedMul (fixed_t a, fixed_t b)
-;----------------------------------------------------------------------------
-cglobal FixedMul
-;       align   16
-FixedMul:
-        mov     eax,[esp+4]
-        imul    dword [esp+8]
-        shrd    eax,edx,16
-        ret
-
-;----------------------------------------------------------------------------
-;fixed_t FixedDiv2 (fixed_t a, fixed_t b);
-;----------------------------------------------------------------------------
-cglobal FixedDiv2
-;       align   16
-FixedDiv2:
-        mov     eax,[esp+4]
-        mov     edx,eax                 ;; these two instructions allow the next
-        sar     edx,31                  ;; two to pair, on the Pentium processor.
-        shld    edx,eax,16
-        sal     eax,16
-        idiv    dword [esp+8]
-        ret
diff --git a/src/v_video.c b/src/v_video.c
index 4141153e9..49d4bdc41 100644
--- a/src/v_video.c
+++ b/src/v_video.c
@@ -489,12 +489,6 @@ static void CV_palette_OnChange(void)
 	V_SetPalette(0);
 }
 
-#if defined (__GNUC__) && defined (__i386__) && !defined (NOASM) && !defined (__APPLE__) && !defined (NORUSEASM)
-void VID_BlitLinearScreen_ASM(const UINT8 *srcptr, UINT8 *destptr, INT32 width, INT32 height, size_t srcrowbytes,
-	size_t destrowbytes);
-#define HAVE_VIDCOPY
-#endif
-
 static void CV_constextsize_OnChange(void)
 {
 	if (!con_startup)
@@ -508,9 +502,6 @@ static void CV_constextsize_OnChange(void)
 void VID_BlitLinearScreen(const UINT8 *srcptr, UINT8 *destptr, INT32 width, INT32 height, size_t srcrowbytes,
 	size_t destrowbytes)
 {
-#ifdef HAVE_VIDCOPY
-    VID_BlitLinearScreen_ASM(srcptr,destptr,width,height,srcrowbytes,destrowbytes);
-#else
 	if ((srcrowbytes == destrowbytes) && (srcrowbytes == (size_t)width))
 		M_Memcpy(destptr, srcptr, srcrowbytes * height);
 	else
@@ -523,7 +514,6 @@ void VID_BlitLinearScreen(const UINT8 *srcptr, UINT8 *destptr, INT32 width, INT3
 			srcptr += srcrowbytes;
 		}
 	}
-#endif
 }
 
 void V_AdjustXYWithSnap(INT32 *x, INT32 *y, UINT32 options, INT32 dupx, INT32 dupy)
diff --git a/src/win32ce/win_main.c b/src/win32ce/win_main.c
index fe4147291..c56452048 100644
--- a/src/win32ce/win_main.c
+++ b/src/win32ce/win_main.c
@@ -426,16 +426,7 @@ static VOID     GetArgcArgv (LPSTR cmdline)
 
 static inline VOID MakeCodeWritable(VOID)
 {
-#ifdef USEASM
-	// Disable write-protection of code segment
-	DWORD OldRights;
-	BYTE *pBaseOfImage = (BYTE *)GetModuleHandle(NULL);
-	IMAGE_OPTIONAL_HEADER *pHeader = (IMAGE_OPTIONAL_HEADER *)
-		(pBaseOfImage + ((IMAGE_DOS_HEADER*)pBaseOfImage)->e_lfanew +
-		sizeof (IMAGE_NT_SIGNATURE) + sizeof (IMAGE_FILE_HEADER));
-	if (!VirtualProtect(pBaseOfImage+pHeader->BaseOfCode,pHeader->SizeOfCode,PAGE_EXECUTE_READWRITE,&OldRights))
-		I_Error("Could not make code writable\n");
-#endif
+	;
 }
 
 
diff --git a/tools/anglechk.c b/tools/anglechk.c
new file mode 100644
index 000000000..7f56abff7
--- /dev/null
+++ b/tools/anglechk.c
@@ -0,0 +1,364 @@
+/*
+ *      anglechk.c
+ *
+ *      Copyright 2007 Alam Arias <Alam.GBC@gmail.com>
+ *
+ *      This program is free software; you can redistribute it and/or modify
+ *      it under the terms of the GNU General Public License as published by
+ *      the Free Software Foundation; either version 2 of the License, or
+ *      (at your option) any later version.
+ *
+ *      This program is distributed in the hope that it will be useful,
+ *      but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *      MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *      GNU General Public License for more details.
+ *
+ *      You should have received a copy of the GNU General Public License
+ *      along with this program; if not, write to the Free Software
+ *      Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+ */
+
+#include <stdio.h>
+#ifdef _MSC_VER
+#include <assert.h>
+#endif
+#include "../src/tables.h"
+#define NO_M
+#include "../src/m_fixed.c"
+#define FIXEDPOINTCONV
+
+// With angle_t,
+// 360 deg = 2^32
+// 45 deg = 2^29
+// 1 deg = 2^29 / 45
+// To convert an angle to a fixed-point number of degrees, then, use
+//   fixed = angle * FRACUNIT / ((1<<29) / 45)
+// But, done literally like that, this will overflow.
+// It's mathematically equivalent to
+//   fixed = angle * (1<<FRACBITS) / (1<<29) * 45
+//   fixed = 45 * angle * (1<<(FRACBITS-29))
+//   fixed = 45 * angle * (1>>(29-FRACBITS))
+//   fixed = (angle>>(29-FRACBITS)) * 45
+
+#define ANGLE_TO_FIXED(a) (fixed_t)(((a)>>(29-FRACBITS))*45)
+#define FIXED_TO_ANGLE(x) (angle_t)(((x)/45)<<(29-FRACBITS))
+
+/*
+Old code that failed if FRACBITS was not 16.
+In particular, the use of FRACUNIT in the definition of ANGLE_F is completely
+wrong. The value wanted actually happens to be 65536 due to the definition
+of angle_t (it's specified so that 360 degrees = 2^32, to take advantage of
+modular arithmetic). That 65536 has nothing whatsoever to do with the setting
+of FRACUNIT.
+
+#define ANGF_D 8192
+#define ANGF_N 45
+#define ANGLE_F (ANGF_N*FRACUNIT/ANGF_D)
+#define FIXED_TO_ANGLE(x) (angle_t)FixedDiv(x, ANGLE_F) // angle_t = FixedDiv(fixed_t, ANGLE_F)
+#define ANGLE_TO_FIXED(x) FixedMul((fixed_t)(x), ANGLE_F) // fixed_t = FixedMul(angle_t, ANGLE_F)
+*/
+
+static fixed_t AngleFixed204(angle_t af)
+{
+	const fixed_t cfn = 180*FRACUNIT;
+	if (af == 0)
+		return 0;
+	else if (af > ANGLE_180)
+		return cfn + ANGLE_TO_FIXED(af - ANGLE_180);
+	else if (af < ANGLE_180)
+		return ANGLE_TO_FIXED(af);
+	else return cfn;
+}
+
+static inline angle_t FixedAngleC204(fixed_t fa, fixed_t factor)
+{
+#if 0 //FixedMod off?
+	const boolean neqda = da < 0, neqfa = fa < 0;
+	const fixed_t afactor = abs(factor);
+	angle_t ra = ANGLE_180;
+
+	if (factor == 0)
+		return FixedAngle204(fa);
+	else if (fa == 0)
+		return 0;
+
+	if (neqfactor)
+	{
+		const fixed_t maf = FixedDiv(afactor, ANGLE_F);
+		const fixed_t cfn = FixedMul(360*FRACUNIT, afactor), hcfn = (cfn/2);
+		const fixed_t fam = FixedMod(fa, cfn);
+
+		if (fam > hcfn)
+			ra = ANGLE_180 + (angle_t)FixedMul(fam - hcfn, maf);
+		else if (fam < hcfn)
+			ra = (angle_t)FixedMul(fam, maf);
+	}
+	else
+	{
+		const fixed_t maf = FixedMul(afactor, ANGLE_F);
+		const fixed_t cfn = FixedDiv(360*FRACUNIT, afactor), hcfn = (cfn/2);
+		const fixed_t fam = FixedMod(fa, cfn);
+
+		if (fam > hcfn)
+			ra = ANGLE_180 + (angle_t)FixedDiv(fam - hcfn, maf);
+		else if (fam < hcfn)
+			ra = (angle_t)FixedDiv(fam, maf);
+	}
+
+	if (neqfa)
+		ra = ANGLE_MAX-ra;
+
+	return ra;
+#else
+	if (factor == 0)
+		return FixedAngle(fa);
+	else if (factor > 0)
+		return (angle_t)((FIXED_TO_FLOAT(fa)/FIXED_TO_FLOAT(factor))*(ANGLE_45/45));
+	else
+		return (angle_t)((FIXED_TO_FLOAT(fa)*FIXED_TO_FLOAT(-factor))*(ANGLE_45/45));
+#endif
+}
+
+angle_t FixedAngle(fixed_t fa)
+{
+#if 0 //FixedMod off?
+	const boolean neqfa = fa < 0;
+	const fixed_t cfn = 180*FRACUNIT;
+	const fixed_t fam = FixedMod(fa, 360*FRACUNIT);
+	angle_t ra = ANGLE_180;
+
+	if (fa == 0)
+		return 0;
+
+	if (fam > cfn)
+		ra = ANGLE_180+FIXED_TO_ANGLE(fam-cfn);
+	else if (fam < cfn)
+		ra = FIXED_TO_ANGLE(fam);
+
+	if (neqfa)
+		ra = ANGLE_MAX-ra;
+
+	return ra;
+#else
+	return (angle_t)(FIXED_TO_FLOAT(fa)*(ANGLE_45/45));
+#endif
+}
+
+fixed_t AngleFixed205(angle_t af)
+{
+#ifdef FIXEDPOINTCONV
+	angle_t wa = ANGLE_180;
+	fixed_t wf = 180*FRACUNIT;
+	fixed_t rf = 0*FRACUNIT;
+	//const angle_t adj = 0x2000;
+
+	//if (af < adj) // too small to notice
+		//return rf;
+
+	while (af)
+	{
+		while (af < wa)
+		{
+			wa /= 2;
+			wf /= 2;
+		}
+		rf += wf;
+		af -= wa;
+	}
+
+	return rf;
+#else
+	const fixed_t cfn = 180*FRACUNIT;
+	if (af == 0)
+		return 0;
+	else if (af > ANGLE_180)
+		return cfn + ANGLE_TO_FIXED(af - ANGLE_180);
+	else if (af < ANGLE_180)
+		return ANGLE_TO_FIXED(af);
+	else return cfn;
+#endif
+}
+
+#ifdef FIXEDPOINTCONV
+static FUNCMATH angle_t AngleAdj(const fixed_t fa, const fixed_t wf,
+                                 angle_t ra)
+{
+	const angle_t adj = 0x77;
+	const boolean fan = fa < 0;
+	const fixed_t sl = FixedDiv(fa, wf*2);
+	const fixed_t lb = FixedRem(fa, wf*2);
+	const fixed_t lo = (wf*2)-lb;
+
+	if (ra == 0)
+	{
+		if (lb == 0)
+		{
+			ra = FixedMul(FRACUNIT/512, sl);
+			if (ra > FRACUNIT/64)
+				return ANGLE_MAX-ra+1;
+			return ra;
+		}
+		else if (lb > 0)
+			return ANGLE_MAX-FixedMul(lo*FRACUNIT, adj)+1;
+		else
+			return ANGLE_MAX-FixedMul(lo*FRACUNIT, adj)+1;
+	}
+
+	if (fan)
+		return ANGLE_MAX-ra+1;
+	else
+		return ra;
+}
+#endif
+
+angle_t FixedAngleC205(fixed_t fa, fixed_t factor)
+{
+#ifdef FIXEDPOINTCONV
+	angle_t wa = ANGLE_180;
+	fixed_t wf = 180*FRACUNIT;
+	angle_t ra = 0;
+	const fixed_t cfa = fa;
+	fixed_t cwf = wf;
+
+	if (fa == 0)
+		return 0;
+
+	if (factor == 0)
+		return FixedAngle(fa);
+	else if (factor > 0)
+		cwf = wf = FixedMul(wf, factor);
+	else if (factor < 0)
+		cwf = wf = FixedDiv(wf, -factor);
+
+	fa = abs(fa);
+
+	while (fa)
+	{
+		while (fa < wf)
+		{
+			wa /= 2;
+			wf /= 2;
+		}
+		ra = ra + wa;
+		fa = fa - wf;
+	}
+
+	return AngleAdj(cfa, cwf, ra);
+#else
+	if (factor == 0)
+		return FixedAngle(fa);
+
+	//fa = FixedMod(fa, 360*FRACUNIT);
+
+	if (factor > 0)
+		return (angle_t)((FIXED_TO_FLOAT(fa)/FIXED_TO_FLOAT(factor))*(ANGLE_45/45));
+	else
+		return (angle_t)((FIXED_TO_FLOAT(fa)*FIXED_TO_FLOAT(-factor))*(ANGLE_45/45));
+#endif
+}
+
+angle_t FixedAngle205(fixed_t fa)
+{
+#ifdef FIXEDPOINTCONV
+	angle_t wa = ANGLE_180;
+	fixed_t wf = 180*FRACUNIT;
+	angle_t ra = 0;
+	const fixed_t cfa = fa;
+	const fixed_t cwf = wf;
+
+	if (fa == 0)
+		return 0;
+
+	fa = abs(fa);
+
+	while (fa)
+	{
+		while (fa < wf)
+		{
+			wa /= 2;
+			wf /= 2;
+		}
+		ra = ra + wa;
+		fa = fa - wf;
+	}
+
+	return AngleAdj(cfa, cwf, ra);
+#else
+	//fa = FixedMod(fa, 360*FRACUNIT);
+
+	if (fa == 0)
+		return 0;
+
+	return (angle_t)(FIXED_TO_FLOAT(fa)*(ANGLE_45/45));
+#endif
+}
+
+int main(int argc, char** argv)
+{
+	fixed_t f, f204, f205;
+	INT64 a;
+	angle_t a204, a205;
+	fixed_t CF = 40*FRACUNIT;
+	int err = 0;
+	(void)argc;
+	(void)argv;
+
+	err = 0x29; //41
+
+	if (1)
+	for (a = 0; a < ANGLE_MAX; a += 0x1)
+	{
+		f204 = AngleFixed204((angle_t)a);
+		f205 = AngleFixed205((angle_t)a);
+		if (f204 != f205 && (abs(f204-f205) > err || f204 == 0 || f205 == 0))
+		{
+			printf("Angle: %u, %d, %d, %d\n", (angle_t)a, f204, f205, f204-f205);
+			//err = abs(f204-f205);
+		}
+	}
+
+	//err = FixedDiv(FRACUNIT, 120*FRACUNIT); // 547
+	err = FixedDiv(FRACUNIT,  62*FRACUNIT); //1059
+
+	if (1)
+	for (f = FRACUNIT*-720; f < FRACUNIT*720; f += 1)
+	{
+		a204 = FixedAngle(f);
+		a205 = FixedAngle205(f);
+		if (a204 != a205 && (abs(a204-a205) > err || a204 == 0 || a205 == 0))
+		{
+			printf("Fixed: %f, %u, %u, %d\n", FIXED_TO_FLOAT(f), a204, a205, a204-a205);
+			err = abs(a204-a205);
+		}
+	}
+
+	//err = FixedDiv(FRACUNIT, 316*FRACUNIT); //207
+	err = FixedDiv(FRACUNIT, 125*FRACUNIT); //526
+
+	if (1)
+	for (f = FixedMul(FRACUNIT*-720, CF); f < FixedMul(FRACUNIT*720, CF); f += FRACUNIT/16)
+	{
+		a204 = FixedAngleC204(f, CF);
+		a205 = FixedAngleC205(f, CF);
+		if (a204 != a205 && (abs(a204-a205) > err || a204 == 0 || a205 == 0))
+		{
+			printf("FixedC: %f, %u, %u, %d\n", FIXED_TO_FLOAT(f), a204, a205, a204-a205);
+			//err = abs(a204-a205);
+		}
+	}
+
+	return 0;
+}
+
+static void *cpu_cpy(void *dest, const void *src, size_t n)
+{
+	return memcpy(dest, src, n);
+}
+
+void *(*M_Memcpy)(void* dest, const void* src, size_t n) = cpu_cpy;
+
+void I_Error(const char *error, ...)
+{
+	(void)error;
+	exit(-1);
+}