Rename chunkset_avx to chunkset_avx2

diff --git a/CMakeLists.txt b/CMakeLists.txt
index d5f66d6dc8acaf43c8cd17a83867fabc2d7b1bf6..6fc12f5bcff3fff9e1dbe72d296a29ea54c91159 100644 (file)
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -762,8 +762,8 @@ if(WITH_OPTIM)
                 add_definitions(-DX86_AVX2)
                 set(AVX2_SRCS ${ARCHDIR}/slide_hash_avx2.c)
                 add_feature_info(AVX2_SLIDEHASH 1 "Support AVX2 optimized slide_hash, using \"${AVX2FLAG}\"")
-                list(APPEND AVX2_SRCS ${ARCHDIR}/chunkset_avx.c)
-                add_feature_info(AVX_CHUNKSET 1 "Support AVX optimized chunkset, using \"${AVX2FLAG}\"")
+                list(APPEND AVX2_SRCS ${ARCHDIR}/chunkset_avx2.c)
+                add_feature_info(AVX2_CHUNKSET 1 "Support AVX2 optimized chunkset, using \"${AVX2FLAG}\"")
                 list(APPEND AVX2_SRCS ${ARCHDIR}/compare256_avx2.c)
                 add_feature_info(AVX2_COMPARE256 1 "Support AVX2 optimized compare256, using \"${AVX2FLAG}\"")
                 list(APPEND AVX2_SRCS ${ARCHDIR}/adler32_avx2.c)

diff --git a/arch/x86/Makefile.in b/arch/x86/Makefile.in
index 5fd51929ce28761d423fbbc39ea6493b3a2beb20..f0478bfdcc63245d837127b2b739a2a67b7c7ba6 100644 (file)
--- a/arch/x86/Makefile.in
+++ b/arch/x86/Makefile.in
@@ -30,7 +30,7 @@ all: \
        adler32_avx512_vnni.o adler32_avx512_vnni.lo \
        adler32_sse42.o adler32_sse42.lo \
        adler32_ssse3.o adler32_ssse3.lo \
-       chunkset_avx.o chunkset_avx.lo \
+       chunkset_avx2.o chunkset_avx2.lo \
        chunkset_sse2.o chunkset_sse2.lo \
        chunkset_ssse3.o chunkset_ssse3.lo \
        compare256_avx2.o compare256_avx2.lo \
@@ -47,11 +47,11 @@ x86_features.o:
 x86_features.lo:
        $(CC) $(SFLAGS) $(XSAVEFLAG) $(INCLUDES) -c -o $@ $(SRCDIR)/x86_features.c
 
-chunkset_avx.o:
-       $(CC) $(CFLAGS) $(AVX2FLAG) $(NOLTOFLAG) $(INCLUDES) -c -o $@ $(SRCDIR)/chunkset_avx.c
+chunkset_avx2.o:
+       $(CC) $(CFLAGS) $(AVX2FLAG) $(NOLTOFLAG) $(INCLUDES) -c -o $@ $(SRCDIR)/chunkset_avx2.c
 
-chunkset_avx.lo:
-       $(CC) $(SFLAGS) $(AVX2FLAG) $(NOLTOFLAG) -DPIC $(INCLUDES) -c -o $@ $(SRCDIR)/chunkset_avx.c
+chunkset_avx2.lo:
+       $(CC) $(SFLAGS) $(AVX2FLAG) $(NOLTOFLAG) -DPIC $(INCLUDES) -c -o $@ $(SRCDIR)/chunkset_avx2.c
 
 chunkset_sse2.o:
        $(CC) $(CFLAGS) $(SSE2FLAG) $(NOLTOFLAG) $(INCLUDES) -c -o $@ $(SRCDIR)/chunkset_sse2.c

diff --git a/arch/x86/chunkset_avx.c b/arch/x86/chunkset_avx2.c
similarity index 93%
rename from arch/x86/chunkset_avx.c
rename to arch/x86/chunkset_avx2.c
index abcbb474f5183ce23a810e2a22f5eef02036cc84..f309878b351ec5bec6b49c1cd1ec18b98c1b15c5 100644 (file)
--- a/arch/x86/chunkset_avx.c
+++ b/arch/x86/chunkset_avx2.c
@@ -1,4 +1,4 @@
-/* chunkset_avx.c -- AVX inline functions to copy small data chunks.
+/* chunkset_avx2.c -- AVX2 inline functions to copy small data chunks.
  * For conditions of distribution and use, see copyright notice in zlib.h
  */
 #include "zbuild.h"
@@ -120,15 +120,15 @@ static inline chunk_t GET_CHUNK_MAG(uint8_t *buf, uint32_t *chunk_rem, uint32_t
     return ret_vec;
 }
 
-#define CHUNKSIZE        chunksize_avx
-#define CHUNKCOPY        chunkcopy_avx
-#define CHUNKUNROLL      chunkunroll_avx
-#define CHUNKMEMSET      chunkmemset_avx
-#define CHUNKMEMSET_SAFE chunkmemset_safe_avx
+#define CHUNKSIZE        chunksize_avx2
+#define CHUNKCOPY        chunkcopy_avx2
+#define CHUNKUNROLL      chunkunroll_avx2
+#define CHUNKMEMSET      chunkmemset_avx2
+#define CHUNKMEMSET_SAFE chunkmemset_safe_avx2
 
 #include "chunkset_tpl.h"
 
-#define INFLATE_FAST     inflate_fast_avx
+#define INFLATE_FAST     inflate_fast_avx2
 
 #include "inffast_tpl.h"
 

diff --git a/configure b/configure
index 2867302b37bc834845b37f048477605be01f388f..1f3ef7a02995081a4d11b60d21b5133888fe7105 100755 (executable)
--- a/configure
+++ b/configure
@@ -1546,8 +1546,8 @@ case "${ARCH}" in
             if test ${HAVE_AVX2_INTRIN} -eq 1; then
                 CFLAGS="${CFLAGS} -DX86_AVX2"
                 SFLAGS="${SFLAGS} -DX86_AVX2"
-                ARCH_STATIC_OBJS="${ARCH_STATIC_OBJS} slide_hash_avx2.o chunkset_avx.o compare256_avx2.o adler32_avx2.o"
-                ARCH_SHARED_OBJS="${ARCH_SHARED_OBJS} slide_hash_avx2.lo chunkset_avx.lo compare256_avx2.lo adler32_avx2.lo"
+                ARCH_STATIC_OBJS="${ARCH_STATIC_OBJS} slide_hash_avx2.o chunkset_avx2.o compare256_avx2.o adler32_avx2.o"
+                ARCH_SHARED_OBJS="${ARCH_SHARED_OBJS} slide_hash_avx2.lo chunkset_avx2.lo compare256_avx2.lo adler32_avx2.lo"
             fi
 
             check_avx512_intrinsics

diff --git a/cpu_features.h b/cpu_features.h
index c742f32c66b84ef649cd4f56a307377a2b665c83..b8ec883a226f33bf8a07e966c10bf2a57e7c21ae 100644 (file)
--- a/cpu_features.h
+++ b/cpu_features.h
@@ -103,8 +103,8 @@ extern uint8_t* chunkmemset_safe_sse2(uint8_t *out, unsigned dist, unsigned len,
 extern uint8_t* chunkmemset_safe_ssse3(uint8_t *out, unsigned dist, unsigned len, unsigned left);
 #endif
 #ifdef X86_AVX2
-extern uint32_t chunksize_avx(void);
-extern uint8_t* chunkmemset_safe_avx(uint8_t *out, unsigned dist, unsigned len, unsigned left);
+extern uint32_t chunksize_avx2(void);
+extern uint8_t* chunkmemset_safe_avx2(uint8_t *out, unsigned dist, unsigned len, unsigned left);
 #endif
 #ifdef ARM_NEON
 extern uint32_t chunksize_neon(void);
@@ -130,7 +130,7 @@ extern void inflate_fast_sse2(PREFIX3(stream) *strm, uint32_t start);
 extern void inflate_fast_ssse3(PREFIX3(stream) *strm, uint32_t start);
 #endif
 #ifdef X86_AVX2
-extern void inflate_fast_avx(PREFIX3(stream) *strm, uint32_t start);
+extern void inflate_fast_avx2(PREFIX3(stream) *strm, uint32_t start);
 #endif
 #ifdef ARM_NEON
 extern void inflate_fast_neon(PREFIX3(stream) *strm, uint32_t start);

diff --git a/functable.c b/functable.c
index f9e8163d3a954bd69db1b226548068f6d01d60d5..aa018c0d64dbda3d941814ffe20c5f65d0a665b5 100644 (file)
--- a/functable.c
+++ b/functable.c
@@ -107,9 +107,9 @@ static void init_functable(void) {
     if (cf.x86.has_avx2) {
         ft.adler32 = &adler32_avx2;
         ft.adler32_fold_copy = &adler32_fold_copy_avx2;
-        ft.chunkmemset_safe = &chunkmemset_safe_avx;
-        ft.chunksize = &chunksize_avx;
-        ft.inflate_fast = &inflate_fast_avx;
+        ft.chunkmemset_safe = &chunkmemset_safe_avx2;
+        ft.chunksize = &chunksize_avx2;
+        ft.inflate_fast = &inflate_fast_avx2;
         ft.slide_hash = &slide_hash_avx2;
 #  ifdef HAVE_BUILTIN_CTZ
         ft.compare256 = &compare256_avx2;

diff --git a/win32/Makefile.msc b/win32/Makefile.msc
index 8a01e31710c07a00e4da53118098c965db14be53..9ed26f283433968d23a2d79132a3eb9e88a5540c 100644 (file)
--- a/win32/Makefile.msc
+++ b/win32/Makefile.msc
@@ -54,7 +54,7 @@ OBJS = \
        adler32_ssse3.obj \
        adler32_fold.obj \
        chunkset.obj \
-       chunkset_avx.obj \
+       chunkset_avx2.obj \
        chunkset_sse2.obj \
        chunkset_ssse3.obj \
        compare256.obj \
@@ -201,7 +201,7 @@ gzwrite.obj: $(SRCDIR)/gzwrite.c $(SRCDIR)/zbuild.h $(SRCDIR)/gzguts.h $(SRCDIR)
 compress.obj: $(SRCDIR)/compress.c $(SRCDIR)/zbuild.h $(SRCDIR)/zlib$(SUFFIX).h
 uncompr.obj: $(SRCDIR)/uncompr.c $(SRCDIR)/zbuild.h $(SRCDIR)/zlib$(SUFFIX).h
 chunkset.obj: $(SRCDIR)/zbuild.h $(SRCDIR)/zutil.h
-chunkset_avx.obj: $(SRCDIR)/arch/x86/chunkset_avx.c $(SRCDIR)/zbuild.h $(SRCDIR)/zutil.h
+chunkset_avx2.obj: $(SRCDIR)/arch/x86/chunkset_avx2.c $(SRCDIR)/zbuild.h $(SRCDIR)/zutil.h
 chunkset_sse2.obj: $(SRCDIR)/arch/x86/chunkset_sse2.c $(SRCDIR)/zbuild.h $(SRCDIR)/zutil.h
 chunkset_ssse3.obj: $(SRCDIR)/arch/x86/chunkset_ssse3.c $(SRCDIR)/zbuild.h $(SRCDIR)/zutil.h
 cpu_features.obj: $(SRCDIR)/cpu_features.c $(SRCDIR)/zbuild.h $(SRCDIR)/zutil.h