in hpel search, merge two 16x16 mc calls into one 16x17. 15% faster hpel, .3% overall.

[x264] / Makefile

diff --git a/Makefile b/Makefile
index e268e921a8282b09ba8a3d642355a67dfa245fd2..8e81ba04f188a230a28fbbae6e350865bb8a668e 100644 (file)
--- a/Makefile
+++ b/Makefile
@@ -19,7 +19,7 @@ endif
 
 # MMX/SSE optims
 ifeq ($(ARCH),X86)
-SRCS   += common/i386/mc-c.c common/i386/dct-c.c common/i386/predict.c
+SRCS   += common/i386/mc-c.c common/i386/predict-c.c
 ASMSRC  = common/i386/dct-a.asm common/i386/cpu-a.asm \
           common/i386/pixel-a.asm common/i386/mc-a.asm \
           common/i386/mc-a2.asm common/i386/predict-a.asm \
@@ -31,7 +31,7 @@ endif
 
 # MMX/SSE optims
 ifeq ($(ARCH),X86_64)
-SRCS   += common/i386/mc-c.c common/i386/dct-c.c common/i386/predict.c
+SRCS   += common/i386/mc-c.c common/i386/predict-c.c
 ASMSRC  = common/amd64/dct-a.asm common/amd64/cpu-a.asm \
           common/amd64/pixel-a.asm common/amd64/mc-a.asm \
           common/amd64/mc-a2.asm common/amd64/predict-a.asm \
@@ -43,7 +43,10 @@ endif
 
 # AltiVec optims
 ifeq ($(ARCH),PPC)
-SRCS += common/ppc/mc.c common/ppc/pixel.c common/ppc/dct.c
+ALTIVECSRC += common/ppc/mc.c common/ppc/pixel.c common/ppc/dct.c \
+              common/ppc/quant.c
+SRCS += $(ALTIVECSRC)
+$(ALTIVECSRC:%.c=%.o): CFLAGS += $(ALTIVECFLAGS)
 endif
 
 # VIS optims
@@ -60,7 +63,9 @@ OBJS = $(SRCS:%.c=%.o)
 OBJCLI = $(SRCCLI:%.c=%.o)
 DEP  = depend
 
-.PHONY: default fprofiled clean distclean install uninstall
+.PHONY: all default fprofiled clean distclean install install-gtk uninstall dox test testclean
+all: default
+
 default: $(DEP) x264$(EXE)
 
 libx264.a: .depend $(OBJS) $(OBJASM)
@@ -73,8 +78,8 @@ $(SONAME): .depend $(OBJS) $(OBJASM)
 x264$(EXE): $(OBJCLI) libx264.a 
        $(CC) -o $@ $+ $(LDFLAGS)
 
-x264vfw.dll: libx264.a $(wildcard vfw/*.c vfw/*.h)
-       make -C vfw/build/cygwin
+libx264gtk.a: muxers.o libx264.a
+       $(MAKE) -C gtk
 
 checkasm: tools/checkasm.o libx264.a
        $(CC) -o $@ $+ $(LDFLAGS)
@@ -83,11 +88,13 @@ common/amd64/*.o: common/amd64/amd64inc.asm
 common/i386/*.o: common/i386/i386inc.asm
 %.o: %.asm
        $(AS) $(ASFLAGS) -o $@ $<
+# delete local/anonymous symbols, so they don't show up in oprofile
+       -@ strip -x $@
 
 .depend: config.mak
        rm -f .depend
 # Hacky - because gcc 2.9x doesn't have -MT
-       $(foreach SRC, $(SRCS) $(SRCCLI), ( echo -n "`dirname $(SRC)`/" && $(CC) $(CFLAGS) $(SRC) -MM -g0 ) 1>> .depend;)
+       $(foreach SRC, $(SRCS) $(SRCCLI), ( $(ECHON) "`dirname $(SRC)`/" && $(CC) $(CFLAGS) $(ALTIVECFLAGS) $(SRC) -MM -g0 ) 1>> .depend;)
 
 config.mak: $(wildcard .svn/entries */.svn/entries */*/.svn/entries)
        ./configure $(CONFIGURE_ARGS)
@@ -101,23 +108,23 @@ SRC2 = $(SRCS) $(SRCCLI)
 # These should cover most of the important codepaths
 OPT0 = --crf 30 -b1 -m1 -r1 --me dia --no-cabac
 OPT1 = --crf 18 -b2 -m3 -r3 --me hex -8 --cqm jvt --direct spatial
-OPT2 = --crf 24 -b3 -m6 -r6 --me umh -8 -w -t1 -A all --b-pyramid --b-rdo --mixed-refs
+OPT2 = --crf 24 -b3 -m7 -r5 --me umh -8 -w -t1 -A all --b-pyramid --b-rdo --mixed-refs --direct auto
 
 ifeq (,$(VIDS))
 fprofiled:
        @echo 'usage: make fprofiled VIDS="infile1 infile2 ..."'
        @echo 'where infiles are anything that x264 understands,'
-       @echo 'i.e. YUV with resolution in the filename, or avisynth.'
+       @echo 'i.e. YUV with resolution in the filename, y4m, or avisynth.'
 else
 fprofiled:
-       make clean
+       $(MAKE) clean
        mv config.mak config.mak2
        sed -e 's/CFLAGS.*/& -fprofile-generate/; s/LDFLAGS.*/& -fprofile-generate/' config.mak2 > config.mak
-       make x264$(EXE)
+       $(MAKE) x264$(EXE)
        $(foreach V, $(VIDS), $(foreach I, 0 1 2, ./x264$(EXE) $(OPT$I) $(V) --progress -o $(DEVNULL) ;))
        rm -f $(SRC2:%.c=%.o)
        sed -e 's/CFLAGS.*/& -fprofile-use/; s/LDFLAGS.*/& -fprofile-use/' config.mak2 > config.mak
-       make
+       $(MAKE)
        rm -f $(SRC2:%.c=%.gcda) $(SRC2:%.c=%.gcno)
        mv config.mak2 config.mak
 endif
@@ -126,13 +133,14 @@ clean:
        rm -f $(OBJS) $(OBJASM) $(OBJCLI) $(SONAME) *.a x264 x264.exe .depend TAGS
        rm -f checkasm checkasm.exe tools/checkasm.o
        rm -f tools/avc2avi tools/avc2avi.exe tools/avc2avi.o
-       rm -rf vfw/build/cygwin/bin
        rm -f $(SRC2:%.c=%.gcda) $(SRC2:%.c=%.gcno)
        - sed -e 's/ *-fprofile-\(generate\|use\)//g' config.mak > config.mak2 && mv config.mak2 config.mak
-       make -C gtk clean
+       $(MAKE) -C gtk clean
 
 distclean: clean
-       rm -f config.mak config.h vfw/build/cygwin/config.mak gtk/config.mak x264.pc
+       rm -f config.mak config.h x264.pc
+       rm -rf test/
+       $(MAKE) -C gtk distclean
 
 install: x264 $(SONAME)
        install -d $(DESTDIR)$(bindir) $(DESTDIR)$(includedir)
@@ -142,15 +150,36 @@ install: x264 $(SONAME)
        install -m 644 x264.pc $(DESTDIR)$(libdir)/pkgconfig
        install x264 $(DESTDIR)$(bindir)
        ranlib $(DESTDIR)$(libdir)/libx264.a
-       $(if $(SONAME), ln -sf $(DESTDIR)$(libdir)/$(SONAME) $(DESTDIR)$(libdir)/libx264.so)
+       $(if $(SONAME), ln -sf $(SONAME) $(DESTDIR)$(libdir)/libx264.so)
        $(if $(SONAME), install -m 755 $(SONAME) $(DESTDIR)$(libdir))
 
+install-gtk: libx264gtk.a
+       $(MAKE) -C gtk install
+
 uninstall:
        rm -f $(DESTDIR)$(includedir)/x264.h $(DESTDIR)$(libdir)/libx264.a
        rm -f $(DESTDIR)$(bindir)/x264 $(DESTDIR)$(libdir)/pkgconfig/x264.pc
        $(if $(SONAME), rm -f $(DESTDIR)$(libdir)/$(SONAME) $(DESTDIR)$(libdir)/libx264.so)
+       $(MAKE) -C gtk uninstall
 
 etags: TAGS
 
 TAGS:
        etags $(SRCS)
+
+dox:
+       doxygen Doxyfile
+
+ifeq (,$(VIDS))
+test:
+       @echo 'usage: make test VIDS="infile1 infile2 ..."'
+       @echo 'where infiles are anything that x264 understands,'
+       @echo 'i.e. YUV with resolution in the filename, y4m, or avisynth.'
+else
+test:
+       perl tools/regression-test.pl --version=head,current --options='$(OPT0)' --options='$(OPT1)' --options='$(OPT2)' $(VIDS:%=--input=%)
+endif
+
+testclean:
+       rm -f test/*.log test/*.264
+       $(foreach DIR, $(wildcard test/x264-r*/), cd $(DIR) ; make clean ; cd ../.. ;)