From e612e317409e8a9d74cf05db0bd733403305f459 Mon Sep 17 00:00:00 2001 From: Kamalalochana Subbaiah Date: Thu, 31 Aug 2017 14:02:06 -0700 Subject: [PATCH] Updated CRC32 Power Optimization Changes Summary: Support for PowerPC Architecture Detecting AltiVec Support Closes https://github.com/facebook/rocksdb/pull/2716 Differential Revision: D5606836 Pulled By: siying fbshipit-source-id: 720262453b1546e5fdbbc668eff56848164113f3 --- Makefile | 100 +++- db/db_impl.cc | 4 +- db/db_test_util.cc | 7 +- src.mk | 10 + tools/db_stress.cc | 16 +- util/crc32c.cc | 84 +++- util/crc32c.h | 3 +- util/crc32c_ppc.c | 95 ++++ util/crc32c_ppc.h | 23 + util/crc32c_ppc_asm.S | 753 ++++++++++++++++++++++++++++++ util/crc32c_ppc_constants.h | 893 ++++++++++++++++++++++++++++++++++++ util/ppc-opcode.h | 31 ++ 12 files changed, 1993 insertions(+), 26 deletions(-) create mode 100644 util/crc32c_ppc.c create mode 100644 util/crc32c_ppc.h create mode 100644 util/crc32c_ppc_asm.S create mode 100644 util/crc32c_ppc_constants.h create mode 100644 util/ppc-opcode.h diff --git a/Makefile b/Makefile index 7c3b055ea..12a25cf6a 100644 --- a/Makefile +++ b/Makefile @@ -96,6 +96,18 @@ OPT += -momit-leaf-frame-pointer endif endif +ifeq (,$(shell $(CXX) -fsyntax-only -maltivec -xc /dev/null 2>&1)) +CXXFLAGS += -DHAS_ALTIVEC +CFLAGS += -DHAS_ALTIVEC +HAS_ALTIVEC=1 +endif + +ifeq (,$(shell $(CXX) -fsyntax-only -mcpu=power8 -xc /dev/null 2>&1)) +CXXFLAGS += -DHAVE_POWER8 +CFLAGS += -DHAVE_POWER8 +HAVE_POWER8=1 +endif + # if we're compiling for release, compile without debug code (-DNDEBUG) and # don't treat warnings as errors ifeq ($(DEBUG_LEVEL),0) @@ -322,6 +334,12 @@ util/build_version.cc: FORCE endif LIBOBJECTS = $(LIB_SOURCES:.cc=.o) +ifeq ($(HAVE_POWER8),1) +LIB_CC_OBJECTS = $(LIB_SOURCES:.cc=.o) +LIBOBJECTS += $(LIB_SOURCES_C:.c=.o) +LIBOBJECTS += $(LIB_SOURCES_ASM:.S=.o) +endif + LIBOBJECTS += $(TOOL_LIB_SOURCES:.cc=.o) MOCKOBJECTS = $(MOCK_LIB_SOURCES:.cc=.o) @@ -573,14 +591,35 @@ $(SHARED2): $(SHARED4) $(SHARED3): $(SHARED4) ln -fs $(SHARED4) $(SHARED3) endif - +ifeq ($(HAVE_POWER8),1) +SHARED_C_OBJECTS = $(LIB_SOURCES_C:.c=.o) +SHARED_ASM_OBJECTS = $(LIB_SOURCES_ASM:.S=.o) +SHARED_C_LIBOBJECTS = $(patsubst %.o,shared-objects/%.o,$(SHARED_C_OBJECTS)) +SHARED_ASM_LIBOBJECTS = $(patsubst %.o,shared-objects/%.o,$(SHARED_ASM_OBJECTS)) +shared_libobjects = $(patsubst %,shared-objects/%,$(LIB_CC_OBJECTS)) +else shared_libobjects = $(patsubst %,shared-objects/%,$(LIBOBJECTS)) -CLEAN_FILES += shared-objects +endif +CLEAN_FILES += shared-objects +shared_all_libobjects = $(shared_libobjects) + +ifeq ($(HAVE_POWER8),1) +shared-ppc-objects = $(SHARED_C_LIBOBJECTS) $(SHARED_ASM_LIBOBJECTS) + +shared-objects/util/crc32c_ppc.o: util/crc32c_ppc.c + $(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@ + +shared-objects/util/crc32c_ppc_asm.o: util/crc32c_ppc_asm.S + $(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@ +endif $(shared_libobjects): shared-objects/%.o: %.cc $(AM_V_CC)mkdir -p $(@D) && $(CXX) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) -c $< -o $@ -$(SHARED4): $(shared_libobjects) +ifeq ($(HAVE_POWER8),1) +shared_all_libobjects = $(shared_libobjects) $(shared-ppc-objects) +endif +$(SHARED4): $(shared_all_libobjects) $(CXX) $(PLATFORM_SHARED_LDFLAGS)$(SHARED3) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) $(shared_libobjects) $(LDFLAGS) -o $@ endif # PLATFORM_SHARED_EXT @@ -1665,13 +1704,36 @@ rocksdbjavastaticpublishcentral: mvn gpg:sign-and-deploy-file -Durl=https://oss.sonatype.org/service/local/staging/deploy/maven2/ -DrepositoryId=sonatype-nexus-staging -DpomFile=java/rocksjni.pom -Dfile=java/target/rocksdbjni-$(ROCKSDB_MAJOR).$(ROCKSDB_MINOR).$(ROCKSDB_PATCH).jar # A version of each $(LIBOBJECTS) compiled with -fPIC +ifeq ($(HAVE_POWER8),1) +JAVA_CC_OBJECTS = $(SHARED_CC_OBJECTS) +JAVA_C_OBJECTS = $(SHARED_C_OBJECTS) +JAVA_ASM_OBJECTS = $(SHARED_ASM_OBJECTS) + +JAVA_C_LIBOBJECTS = $(patsubst %.c.o,jl/%.c.o,$(JAVA_C_OBJECTS)) +JAVA_ASM_LIBOBJECTS = $(patsubst %.S.o,jl/%.S.o,$(JAVA_ASM_OBJECTS)) +endif + java_libobjects = $(patsubst %,jl/%,$(LIBOBJECTS)) CLEAN_FILES += jl +java_all_libobjects = $(java_libobjects) + +ifeq ($(HAVE_POWER8),1) +java_ppc_libobjects = $(JAVA_C_LIBOBJECTS) $(JAVA_ASM_LIBOBJECTS) + +jl/crc32c_ppc.o: util/crc32c_ppc.c + $(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@ + +jl/crc32c_ppc_asm.o: util/crc32c_ppc_asm.S + $(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@ +java_all_libobjects += $(java_ppc_libobjects) +endif $(java_libobjects): jl/%.o: %.cc $(AM_V_CC)mkdir -p $(@D) && $(CXX) $(CXXFLAGS) -fPIC -c $< -o $@ $(COVERAGEFLAGS) -rocksdbjava: $(java_libobjects) + + +rocksdbjava: $(java_all_libobjects) $(AM_V_GEN)cd java;$(MAKE) javalib; $(AM_V_at)rm -f ./java/target/$(ROCKSDBJNILIB) $(AM_V_at)$(CXX) $(CXXFLAGS) -I./java/. $(JAVA_INCLUDE) -shared -fPIC -o ./java/target/$(ROCKSDBJNILIB) $(JNI_NATIVE_SOURCES) $(java_libobjects) $(JAVA_LDFLAGS) $(COVERAGEFLAGS) @@ -1726,30 +1788,54 @@ IOSVERSION=$(shell defaults read $(PLATFORMSROOT)/iPhoneOS.platform/version CFBu lipo ios-x86/$@ ios-arm/$@ -create -output $@ else +ifeq ($(HAVE_POWER8),1) +util/crc32c_ppc.o: util/crc32c_ppc.c + $(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@ + +util/crc32c_ppc_asm.o: util/crc32c_ppc_asm.S + $(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@ +endif .cc.o: $(AM_V_CC)$(CXX) $(CXXFLAGS) -c $< -o $@ $(COVERAGEFLAGS) .c.o: $(AM_V_CC)$(CC) $(CFLAGS) -c $< -o $@ endif - # --------------------------------------------------------------------------- # Source files dependencies detection # --------------------------------------------------------------------------- all_sources = $(LIB_SOURCES) $(MAIN_SOURCES) $(MOCK_LIB_SOURCES) $(TOOL_LIB_SOURCES) $(BENCH_LIB_SOURCES) $(TEST_LIB_SOURCES) $(EXP_LIB_SOURCES) -DEPFILES = $(all_sources:.cc=.d) +DEPFILES = $(all_sources:.cc=.cc.d) # Add proper dependency support so changing a .h file forces a .cc file to # rebuild. # The .d file indicates .cc file's dependencies on .h files. We generate such # dependency by g++'s -MM option, whose output is a make dependency rule. -$(DEPFILES): %.d: %.cc +%.cc.d: %.cc @$(CXX) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) \ -MM -MT'$@' -MT'$(<:.cc=.o)' "$<" -o '$@' +ifeq ($(HAVE_POWER8),1) +DEPFILES_C = $(LIB_SOURCES_C:.c=.c.d) +DEPFILES_ASM = $(LIB_SOURCES_ASM:.S=.S.d) + +%.c.d: %.c + @$(CXX) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) \ + -MM -MT'$@' -MT'$(<:.c=.o)' "$<" -o '$@' + +%.S.d: %.S + @$(CXX) $(CXXFLAGS) $(PLATFORM_SHARED_CFLAGS) \ + -MM -MT'$@' -MT'$(<:.S=.o)' "$<" -o '$@' + +$(DEPFILES_C): %.c.d + +$(DEPFILES_ASM): %.S.d +depend: $(DEPFILES) $(DEPFILES_C) $(DEPFILES_ASM) +else depend: $(DEPFILES) +endif # if the make goal is either "clean" or "format", we shouldn't # try to import the *.d files. diff --git a/db/db_impl.cc b/db/db_impl.cc index 502d55f7d..0652ce0c0 100644 --- a/db/db_impl.cc +++ b/db/db_impl.cc @@ -133,8 +133,8 @@ void DumpSupportInfo(Logger* logger) { ROCKS_LOG_HEADER(logger, "\tZSTDNotFinal supported: %d", ZSTDNotFinal_Supported()); ROCKS_LOG_HEADER(logger, "\tZSTD supported: %d", ZSTD_Supported()); - ROCKS_LOG_HEADER(logger, "Fast CRC32 supported: %d", - crc32c::IsFastCrc32Supported()); + ROCKS_LOG_HEADER(logger, "Fast CRC32 supported: %s", + crc32c::IsFastCrc32Supported().c_str()); } int64_t kDefaultLowPriThrottledRate = 2 * 1024 * 1024; diff --git a/db/db_test_util.cc b/db/db_test_util.cc index c4d465ba1..79212353a 100644 --- a/db/db_test_util.cc +++ b/db/db_test_util.cc @@ -288,12 +288,11 @@ Options DBTestBase::GetOptions( Options options = default_options; BlockBasedTableOptions table_options; bool set_block_based_table_factory = true; -#if !defined(OS_MACOSX) && !defined(OS_WIN) && !defined(OS_SOLARIS) && \ - !defined(OS_AIX) +#if !defined(OS_MACOSX) && !defined(OS_WIN) && !defined(OS_SOLARIS) && \ + !defined(OS_AIX) rocksdb::SyncPoint::GetInstance()->ClearCallBack( "NewRandomAccessFile:O_DIRECT"); - rocksdb::SyncPoint::GetInstance()->ClearCallBack( - "NewWritableFile:O_DIRECT"); + rocksdb::SyncPoint::GetInstance()->ClearCallBack("NewWritableFile:O_DIRECT"); #endif bool can_allow_mmap = IsMemoryMappedAccessSupported(); diff --git a/src.mk b/src.mk index 443070ea2..a81d1d731 100644 --- a/src.mk +++ b/src.mk @@ -205,6 +205,16 @@ LIB_SOURCES = \ utilities/write_batch_with_index/write_batch_with_index.cc \ utilities/write_batch_with_index/write_batch_with_index_internal.cc \ +ifeq (,$(shell $(CXX) -fsyntax-only -maltivec -xc /dev/null 2>&1)) +LIB_SOURCES_ASM =\ + util/crc32c_ppc_asm.S +LIB_SOURCES_C = \ + util/crc32c_ppc.c +else +LIB_SOURCES_ASM = +LIB_SOURCES_C = +endif + TOOL_LIB_SOURCES = \ tools/ldb_cmd.cc \ tools/ldb_tool.cc \ diff --git a/tools/db_stress.cc b/tools/db_stress.cc index e5361233b..6de5eb7fb 100644 --- a/tools/db_stress.cc +++ b/tools/db_stress.cc @@ -2422,15 +2422,15 @@ int main(int argc, char** argv) { #if !defined(NDEBUG) && !defined(OS_MACOSX) && !defined(OS_WIN) && \ !defined(OS_SOLARIS) && !defined(OS_AIX) rocksdb::SyncPoint::GetInstance()->SetCallBack( - "NewWritableFile:O_DIRECT", [&](void* arg) { - int* val = static_cast(arg); - *val &= ~O_DIRECT; - }); + "NewWritableFile:O_DIRECT", [&](void* arg) { + int* val = static_cast(arg); + *val &= ~O_DIRECT; + }); rocksdb::SyncPoint::GetInstance()->SetCallBack( - "NewRandomAccessFile:O_DIRECT", [&](void* arg) { - int* val = static_cast(arg); - *val &= ~O_DIRECT; - }); + "NewRandomAccessFile:O_DIRECT", [&](void* arg) { + int* val = static_cast(arg); + *val &= ~O_DIRECT; + }); rocksdb::SyncPoint::GetInstance()->EnableProcessing(); #endif diff --git a/util/crc32c.cc b/util/crc32c.cc index ae36f8230..d44a8606a 100644 --- a/util/crc32c.cc +++ b/util/crc32c.cc @@ -18,9 +18,32 @@ #endif #include "util/coding.h" +#ifdef __powerpc64__ +#include "util/crc32c_ppc.h" +#include "util/crc32c_ppc_constants.h" + +#if __linux__ +#include + +#ifndef PPC_FEATURE2_VEC_CRYPTO +#define PPC_FEATURE2_VEC_CRYPTO 0x02000000 +#endif + +#ifndef AT_HWCAP2 +#define AT_HWCAP2 26 +#endif + +#endif /* __linux__ */ + +#endif + namespace rocksdb { namespace crc32c { +#ifdef __powerpc64__ +static int arch_ppc_crc32 = 0; +#endif + static const uint32_t table0_[256] = { 0x00000000, 0xf26b8303, 0xe13b70f7, 0x1350f3f4, 0xc79a971f, 0x35f1141c, 0x26a1e7e8, 0xd4ca64eb, @@ -371,6 +394,7 @@ uint32_t ExtendImpl(uint32_t crc, const char* buf, size_t size) { } // Detect if SS42 or not. +#ifndef HAVE_POWER8 static bool isSSE42() { #ifndef HAVE_SSE42 return false; @@ -387,15 +411,67 @@ static bool isSSE42() { return false; #endif } +#endif typedef uint32_t (*Function)(uint32_t, const char*, size_t); -static inline Function Choose_Extend() { - return isSSE42() ? ExtendImpl : ExtendImpl; +#if defined(HAVE_POWER8) && defined(HAS_ALTIVEC) +uint32_t ExtendPPCImpl(uint32_t crc, const char *buf, size_t size) { + return crc32c_ppc(crc, (const unsigned char *)buf, size); } -bool IsFastCrc32Supported() { - return isSSE42(); +#if __linux__ +static int arch_ppc_probe(void) { + arch_ppc_crc32 = 0; + +#if defined(__powerpc64__) + if (getauxval(AT_HWCAP2) & PPC_FEATURE2_VEC_CRYPTO) arch_ppc_crc32 = 1; +#endif /* __powerpc64__ */ + + return arch_ppc_crc32; +} +#endif // __linux__ + +static bool isAltiVec() { + if (arch_ppc_probe()) { + return true; + } else { + return false; + } +} +#endif + +static inline Function Choose_Extend() { +#ifndef HAVE_POWER8 + return isSSE42() ? ExtendImpl : ExtendImpl; +#else + return isAltiVec() ? ExtendPPCImpl : ExtendImpl; +#endif +} + +std::string IsFastCrc32Supported() { + bool has_fast_crc = false; + std::string fast_zero_msg; + std::string arch; +#ifdef HAVE_POWER8 +#ifdef HAS_ALTIVEC + if (arch_ppc_probe()) { + has_fast_crc = true; + arch = "PPC"; + } +#else + has_fast_crc = false; + arch = "PPC"; +#endif +#else + has_fast_crc = isSSE42(); + arch = "x86"; +#endif + if (has_fast_crc){ + fast_zero_msg.append("Supported on " + arch); + } + fast_zero_msg.append("Not supported on " + arch); + return fast_zero_msg; } Function ChosenExtend = Choose_Extend(); diff --git a/util/crc32c.h b/util/crc32c.h index 984852969..faee5d54b 100644 --- a/util/crc32c.h +++ b/util/crc32c.h @@ -10,11 +10,12 @@ #pragma once #include #include +#include namespace rocksdb { namespace crc32c { -extern bool IsFastCrc32Supported(); +extern std::string IsFastCrc32Supported(); // Return the crc32c of concat(A, data[0,n-1]) where init_crc is the // crc32c of some string A. Extend() is often used to maintain the diff --git a/util/crc32c_ppc.c b/util/crc32c_ppc.c new file mode 100644 index 000000000..3c517c88c --- /dev/null +++ b/util/crc32c_ppc.c @@ -0,0 +1,95 @@ +// Copyright (c) 2017 International Business Machines Corp. +// All rights reserved. +// This source code is licensed under the BSD-style license found in the +// LICENSE file in the root directory of this source tree. An additional grant +// of patent rights can be found in the PATENTS file in the same directory. +// This source code is also licensed under the GPLv2 license found in the +// COPYING file in the root directory of this source tree. + +#define CRC_TABLE +#include +#include +#include +#include "util/crc32c_ppc_constants.h" + +#define VMX_ALIGN 16 +#define VMX_ALIGN_MASK (VMX_ALIGN - 1) + +#ifdef REFLECT +static unsigned int crc32_align(unsigned int crc, unsigned char const *p, + unsigned long len) { + while (len--) crc = crc_table[(crc ^ *p++) & 0xff] ^ (crc >> 8); + return crc; +} +#endif + +#ifdef HAVE_POWER8 +unsigned int __crc32_vpmsum(unsigned int crc, unsigned char const *p, + unsigned long len); + +static uint32_t crc32_vpmsum(uint32_t crc, unsigned char const *data, + unsigned len) { + unsigned int prealign; + unsigned int tail; + +#ifdef CRC_XOR + crc ^= 0xffffffff; +#endif + + if (len < VMX_ALIGN + VMX_ALIGN_MASK) { + crc = crc32_align(crc, data, (unsigned long)len); + goto out; + } + + if ((unsigned long)data & VMX_ALIGN_MASK) { + prealign = VMX_ALIGN - ((unsigned long)data & VMX_ALIGN_MASK); + crc = crc32_align(crc, data, prealign); + len -= prealign; + data += prealign; + } + + crc = __crc32_vpmsum(crc, data, (unsigned long)len & ~VMX_ALIGN_MASK); + + tail = len & VMX_ALIGN_MASK; + if (tail) { + data += len & ~VMX_ALIGN_MASK; + crc = crc32_align(crc, data, tail); + } + +out: +#ifdef CRC_XOR + crc ^= 0xffffffff; +#endif + + return crc; +} + +/* This wrapper function works around the fact that crc32_vpmsum + * does not gracefully handle the case where the data pointer is NULL. There + * may be room for performance improvement here. + */ +uint32_t crc32c_ppc(uint32_t crc, unsigned char const *data, unsigned len) { + unsigned char *buf2; + + if (!data) { + buf2 = (unsigned char *)malloc(len); + bzero(buf2, len); + crc = crc32_vpmsum(crc, buf2, len); + free(buf2); + } else { + crc = crc32_vpmsum(crc, data, (unsigned long)len); + } + return crc; +} + +#else /* HAVE_POWER8 */ + +/* This symbol has to exist on non-ppc architectures (and on legacy + * ppc systems using power7 or below) in order to compile properly + * there, even though it won't be called. + */ +uint32_t crc32c_ppc(uint32_t crc, unsigned char const *data, unsigned len) { + return 0; +} + +#endif /* HAVE_POWER8 */ diff --git a/util/crc32c_ppc.h b/util/crc32c_ppc.h new file mode 100644 index 000000000..b52ad9b2a --- /dev/null +++ b/util/crc32c_ppc.h @@ -0,0 +1,23 @@ +// Copyright (c) 2017 International Business Machines Corp. +// All rights reserved. +// This source code is licensed under the BSD-style license found in the +// LICENSE file in the root directory of this source tree. An additional grant +// of patent rights can be found in the PATENTS file in the same directory. +// This source code is also licensed under the GPLv2 license found in the +// COPYING file in the root directory of this source tree. + +#ifndef CRC32C_PPC_H +#define CRC32C_PPC_H + +#ifdef __cplusplus +extern "C" { +#endif + +extern uint32_t crc32c_ppc(uint32_t crc, unsigned char const *buffer, + unsigned len); + +#ifdef __cplusplus +} +#endif + +#endif diff --git a/util/crc32c_ppc_asm.S b/util/crc32c_ppc_asm.S new file mode 100644 index 000000000..6de797973 --- /dev/null +++ b/util/crc32c_ppc_asm.S @@ -0,0 +1,753 @@ +// Copyright (c) 2015 Anton Blanchard , IBM +// Copyright (c) 2017 International Business Machines Corp. +// All rights reserved. +// This source code is licensed under the BSD-style license found in the +// LICENSE file in the root directory of this source tree. An additional grant +// of patent rights can be found in the PATENTS file in the same directory. +// This source code is also licensed under the GPLv2 license found in the +// COPYING file in the root directory of this source tree. + +#include +#include "ppc-opcode.h" + +#undef toc + +#ifndef r1 +#define r1 1 +#endif + +#ifndef r2 +#define r2 2 +#endif + + .section .rodata +.balign 16 + +.byteswap_constant: + /* byte reverse permute constant */ + .octa 0x0F0E0D0C0B0A09080706050403020100 + +#define __ASSEMBLY__ +#include "crc32c_ppc_constants.h" + + .text + +#if defined(__BIG_ENDIAN__) && defined(REFLECT) +#define BYTESWAP_DATA +#elif defined(__LITTLE_ENDIAN__) && !defined(REFLECT) +#define BYTESWAP_DATA +#else +#undef BYTESWAP_DATA +#endif + +#define off16 r25 +#define off32 r26 +#define off48 r27 +#define off64 r28 +#define off80 r29 +#define off96 r30 +#define off112 r31 + +#define const1 v24 +#define const2 v25 + +#define byteswap v26 +#define mask_32bit v27 +#define mask_64bit v28 +#define zeroes v29 + +#ifdef BYTESWAP_DATA +#define VPERM(A, B, C, D) vperm A, B, C, D +#else +#define VPERM(A, B, C, D) +#endif + +/* unsigned int __crc32_vpmsum(unsigned int crc, void *p, unsigned long len) */ +FUNC_START(__crc32_vpmsum) + std r31,-8(r1) + std r30,-16(r1) + std r29,-24(r1) + std r28,-32(r1) + std r27,-40(r1) + std r26,-48(r1) + std r25,-56(r1) + + li off16,16 + li off32,32 + li off48,48 + li off64,64 + li off80,80 + li off96,96 + li off112,112 + li r0,0 + + /* Enough room for saving 10 non volatile VMX registers */ + subi r6,r1,56+10*16 + subi r7,r1,56+2*16 + + stvx v20,0,r6 + stvx v21,off16,r6 + stvx v22,off32,r6 + stvx v23,off48,r6 + stvx v24,off64,r6 + stvx v25,off80,r6 + stvx v26,off96,r6 + stvx v27,off112,r6 + stvx v28,0,r7 + stvx v29,off16,r7 + + mr r10,r3 + + vxor zeroes,zeroes,zeroes + vspltisw v0,-1 + + vsldoi mask_32bit,zeroes,v0,4 + vsldoi mask_64bit,zeroes,v0,8 + + /* Get the initial value into v8 */ + vxor v8,v8,v8 + MTVRD(v8, r3) +#ifdef REFLECT + vsldoi v8,zeroes,v8,8 /* shift into bottom 32 bits */ +#else + vsldoi v8,v8,zeroes,4 /* shift into top 32 bits */ +#endif + +#ifdef BYTESWAP_DATA + addis r3,r2,.byteswap_constant@toc@ha + addi r3,r3,.byteswap_constant@toc@l + + lvx byteswap,0,r3 + addi r3,r3,16 +#endif + + cmpdi r5,256 + blt .Lshort + + rldicr r6,r5,0,56 + + /* Checksum in blocks of MAX_SIZE */ +1: lis r7,MAX_SIZE@h + ori r7,r7,MAX_SIZE@l + mr r9,r7 + cmpd r6,r7 + bgt 2f + mr r7,r6 +2: subf r6,r7,r6 + + /* our main loop does 128 bytes at a time */ + srdi r7,r7,7 + + /* + * Work out the offset into the constants table to start at. Each + * constant is 16 bytes, and it is used against 128 bytes of input + * data - 128 / 16 = 8 + */ + sldi r8,r7,4 + srdi r9,r9,3 + subf r8,r8,r9 + + /* We reduce our final 128 bytes in a separate step */ + addi r7,r7,-1 + mtctr r7 + + addis r3,r2,.constants@toc@ha + addi r3,r3,.constants@toc@l + + /* Find the start of our constants */ + add r3,r3,r8 + + /* zero v0-v7 which will contain our checksums */ + vxor v0,v0,v0 + vxor v1,v1,v1 + vxor v2,v2,v2 + vxor v3,v3,v3 + vxor v4,v4,v4 + vxor v5,v5,v5 + vxor v6,v6,v6 + vxor v7,v7,v7 + + lvx const1,0,r3 + + /* + * If we are looping back to consume more data we use the values + * already in v16-v23. + */ + cmpdi r0,1 + beq 2f + + /* First warm up pass */ + lvx v16,0,r4 + lvx v17,off16,r4 + VPERM(v16,v16,v16,byteswap) + VPERM(v17,v17,v17,byteswap) + lvx v18,off32,r4 + lvx v19,off48,r4 + VPERM(v18,v18,v18,byteswap) + VPERM(v19,v19,v19,byteswap) + lvx v20,off64,r4 + lvx v21,off80,r4 + VPERM(v20,v20,v20,byteswap) + VPERM(v21,v21,v21,byteswap) + lvx v22,off96,r4 + lvx v23,off112,r4 + VPERM(v22,v22,v22,byteswap) + VPERM(v23,v23,v23,byteswap) + addi r4,r4,8*16 + + /* xor in initial value */ + vxor v16,v16,v8 + +2: bdz .Lfirst_warm_up_done + + addi r3,r3,16 + lvx const2,0,r3 + + /* Second warm up pass */ + VPMSUMD(v8,v16,const1) + lvx v16,0,r4 + VPERM(v16,v16,v16,byteswap) + ori r2,r2,0 + + VPMSUMD(v9,v17,const1) + lvx v17,off16,r4 + VPERM(v17,v17,v17,byteswap) + ori r2,r2,0 + + VPMSUMD(v10,v18,const1) + lvx v18,off32,r4 + VPERM(v18,v18,v18,byteswap) + ori r2,r2,0 + + VPMSUMD(v11,v19,const1) + lvx v19,off48,r4 + VPERM(v19,v19,v19,byteswap) + ori r2,r2,0 + + VPMSUMD(v12,v20,const1) + lvx v20,off64,r4 + VPERM(v20,v20,v20,byteswap) + ori r2,r2,0 + + VPMSUMD(v13,v21,const1) + lvx v21,off80,r4 + VPERM(v21,v21,v21,byteswap) + ori r2,r2,0 + + VPMSUMD(v14,v22,const1) + lvx v22,off96,r4 + VPERM(v22,v22,v22,byteswap) + ori r2,r2,0 + + VPMSUMD(v15,v23,const1) + lvx v23,off112,r4 + VPERM(v23,v23,v23,byteswap) + + addi r4,r4,8*16 + + bdz .Lfirst_cool_down + + /* + * main loop. We modulo schedule it such that it takes three iterations + * to complete - first iteration load, second iteration vpmsum, third + * iteration xor. + */ + .balign 16 +4: lvx const1,0,r3 + addi r3,r3,16 + ori r2,r2,0 + + vxor v0,v0,v8 + VPMSUMD(v8,v16,const2) + lvx v16,0,r4 + VPERM(v16,v16,v16,byteswap) + ori r2,r2,0 + + vxor v1,v1,v9 + VPMSUMD(v9,v17,const2) + lvx v17,off16,r4 + VPERM(v17,v17,v17,byteswap) + ori r2,r2,0 + + vxor v2,v2,v10 + VPMSUMD(v10,v18,const2) + lvx v18,off32,r4 + VPERM(v18,v18,v18,byteswap) + ori r2,r2,0 + + vxor v3,v3,v11 + VPMSUMD(v11,v19,const2) + lvx v19,off48,r4 + VPERM(v19,v19,v19,byteswap) + lvx const2,0,r3 + ori r2,r2,0 + + vxor v4,v4,v12 + VPMSUMD(v12,v20,const1) + lvx v20,off64,r4 + VPERM(v20,v20,v20,byteswap) + ori r2,r2,0 + + vxor v5,v5,v13 + VPMSUMD(v13,v21,const1) + lvx v21,off80,r4 + VPERM(v21,v21,v21,byteswap) + ori r2,r2,0 + + vxor v6,v6,v14 + VPMSUMD(v14,v22,const1) + lvx v22,off96,r4 + VPERM(v22,v22,v22,byteswap) + ori r2,r2,0 + + vxor v7,v7,v15 + VPMSUMD(v15,v23,const1) + lvx v23,off112,r4 + VPERM(v23,v23,v23,byteswap) + + addi r4,r4,8*16 + + bdnz 4b + +.Lfirst_cool_down: + /* First cool down pass */ + lvx const1,0,r3 + addi r3,r3,16 + + vxor v0,v0,v8 + VPMSUMD(v8,v16,const1) + ori r2,r2,0 + + vxor v1,v1,v9 + VPMSUMD(v9,v17,const1) + ori r2,r2,0 + + vxor v2,v2,v10 + VPMSUMD(v10,v18,const1) + ori r2,r2,0 + + vxor v3,v3,v11 + VPMSUMD(v11,v19,const1) + ori r2,r2,0 + + vxor v4,v4,v12 + VPMSUMD(v12,v20,const1) + ori r2,r2,0 + + vxor v5,v5,v13 + VPMSUMD(v13,v21,const1) + ori r2,r2,0 + + vxor v6,v6,v14 + VPMSUMD(v14,v22,const1) + ori r2,r2,0 + + vxor v7,v7,v15 + VPMSUMD(v15,v23,const1) + ori r2,r2,0 + +.Lsecond_cool_down: + /* Second cool down pass */ + vxor v0,v0,v8 + vxor v1,v1,v9 + vxor v2,v2,v10 + vxor v3,v3,v11 + vxor v4,v4,v12 + vxor v5,v5,v13 + vxor v6,v6,v14 + vxor v7,v7,v15 + +#ifdef REFLECT + /* + * vpmsumd produces a 96 bit result in the least significant bits + * of the register. Since we are bit reflected we have to shift it + * left 32 bits so it occupies the least significant bits in the + * bit reflected domain. + */ + vsldoi v0,v0,zeroes,4 + vsldoi v1,v1,zeroes,4 + vsldoi v2,v2,zeroes,4 + vsldoi v3,v3,zeroes,4 + vsldoi v4,v4,zeroes,4 + vsldoi v5,v5,zeroes,4 + vsldoi v6,v6,zeroes,4 + vsldoi v7,v7,zeroes,4 +#endif + + /* xor with last 1024 bits */ + lvx v8,0,r4 + lvx v9,off16,r4 + VPERM(v8,v8,v8,byteswap) + VPERM(v9,v9,v9,byteswap) + lvx v10,off32,r4 + lvx v11,off48,r4 + VPERM(v10,v10,v10,byteswap) + VPERM(v11,v11,v11,byteswap) + lvx v12,off64,r4 + lvx v13,off80,r4 + VPERM(v12,v12,v12,byteswap) + VPERM(v13,v13,v13,byteswap) + lvx v14,off96,r4 + lvx v15,off112,r4 + VPERM(v14,v14,v14,byteswap) + VPERM(v15,v15,v15,byteswap) + + addi r4,r4,8*16 + + vxor v16,v0,v8 + vxor v17,v1,v9 + vxor v18,v2,v10 + vxor v19,v3,v11 + vxor v20,v4,v12 + vxor v21,v5,v13 + vxor v22,v6,v14 + vxor v23,v7,v15 + + li r0,1 + cmpdi r6,0 + addi r6,r6,128 + bne 1b + + /* Work out how many bytes we have left */ + andi. r5,r5,127 + + /* Calculate where in the constant table we need to start */ + subfic r6,r5,128 + add r3,r3,r6 + + /* How many 16 byte chunks are in the tail */ + srdi r7,r5,4 + mtctr r7 + + /* + * Reduce the previously calculated 1024 bits to 64 bits, shifting + * 32 bits to include the trailing 32 bits of zeros + */ + lvx v0,0,r3 + lvx v1,off16,r3 + lvx v2,off32,r3 + lvx v3,off48,r3 + lvx v4,off64,r3 + lvx v5,off80,r3 + lvx v6,off96,r3 + lvx v7,off112,r3 + addi r3,r3,8*16 + + VPMSUMW(v0,v16,v0) + VPMSUMW(v1,v17,v1) + VPMSUMW(v2,v18,v2) + VPMSUMW(v3,v19,v3) + VPMSUMW(v4,v20,v4) + VPMSUMW(v5,v21,v5) + VPMSUMW(v6,v22,v6) + VPMSUMW(v7,v23,v7) + + /* Now reduce the tail (0 - 112 bytes) */ + cmpdi r7,0 + beq 1f + + lvx v16,0,r4 + lvx v17,0,r3 + VPERM(v16,v16,v16,byteswap) + VPMSUMW(v16,v16,v17) + vxor v0,v0,v16 + bdz 1f + + lvx v16,off16,r4 + lvx v17,off16,r3 + VPERM(v16,v16,v16,byteswap) + VPMSUMW(v16,v16,v17) + vxor v0,v0,v16 + bdz 1f + + lvx v16,off32,r4 + lvx v17,off32,r3 + VPERM(v16,v16,v16,byteswap) + VPMSUMW(v16,v16,v17) + vxor v0,v0,v16 + bdz 1f + + lvx v16,off48,r4 + lvx v17,off48,r3 + VPERM(v16,v16,v16,byteswap) + VPMSUMW(v16,v16,v17) + vxor v0,v0,v16 + bdz 1f + + lvx v16,off64,r4 + lvx v17,off64,r3 + VPERM(v16,v16,v16,byteswap) + VPMSUMW(v16,v16,v17) + vxor v0,v0,v16 + bdz 1f + + lvx v16,off80,r4 + lvx v17,off80,r3 + VPERM(v16,v16,v16,byteswap) + VPMSUMW(v16,v16,v17) + vxor v0,v0,v16 + bdz 1f + + lvx v16,off96,r4 + lvx v17,off96,r3 + VPERM(v16,v16,v16,byteswap) + VPMSUMW(v16,v16,v17) + vxor v0,v0,v16 + + /* Now xor all the parallel chunks together */ +1: vxor v0,v0,v1 + vxor v2,v2,v3 + vxor v4,v4,v5 + vxor v6,v6,v7 + + vxor v0,v0,v2 + vxor v4,v4,v6 + + vxor v0,v0,v4 + +.Lbarrett_reduction: + /* Barrett constants */ + addis r3,r2,.barrett_constants@toc@ha + addi r3,r3,.barrett_constants@toc@l + + lvx const1,0,r3 + lvx const2,off16,r3 + + vsldoi v1,v0,v0,8 + vxor v0,v0,v1 /* xor two 64 bit results together */ + +#ifdef REFLECT + /* shift left one bit */ + vspltisb v1,1 + vsl v0,v0,v1 +#endif + + vand v0,v0,mask_64bit + +#ifndef REFLECT + /* + * Now for the Barrett reduction algorithm. The idea is to calculate q, + * the multiple of our polynomial that we need to subtract. By + * doing the computation 2x bits higher (ie 64 bits) and shifting the + * result back down 2x bits, we round down to the nearest multiple. + */ + VPMSUMD(v1,v0,const1) /* ma */ + vsldoi v1,zeroes,v1,8 /* q = floor(ma/(2^64)) */ + VPMSUMD(v1,v1,const2) /* qn */ + vxor v0,v0,v1 /* a - qn, subtraction is xor in GF(2) */ + + /* + * Get the result into r3. We need to shift it left 8 bytes: + * V0 [ 0 1 2 X ] + * V0 [ 0 X 2 3 ] + */ + vsldoi v0,v0,zeroes,8 /* shift result into top 64 bits */ +#else + /* + * The reflected version of Barrett reduction. Instead of bit + * reflecting our data (which is expensive to do), we bit reflect our + * constants and our algorithm, which means the intermediate data in + * our vector registers goes from 0-63 instead of 63-0. We can reflect + * the algorithm because we don't carry in mod 2 arithmetic. + */ + vand v1,v0,mask_32bit /* bottom 32 bits of a */ + VPMSUMD(v1,v1,const1) /* ma */ + vand v1,v1,mask_32bit /* bottom 32bits of ma */ + VPMSUMD(v1,v1,const2) /* qn */ + vxor v0,v0,v1 /* a - qn, subtraction is xor in GF(2) */ + + /* + * Since we are bit reflected, the result (ie the low 32 bits) is in + * the high 32 bits. We just need to shift it left 4 bytes + * V0 [ 0 1 X 3 ] + * V0 [ 0 X 2 3 ] + */ + vsldoi v0,v0,zeroes,4 /* shift result into top 64 bits of */ +#endif + + /* Get it into r3 */ + MFVRD(r3, v0) + +.Lout: + subi r6,r1,56+10*16 + subi r7,r1,56+2*16 + + lvx v20,0,r6 + lvx v21,off16,r6 + lvx v22,off32,r6 + lvx v23,off48,r6 + lvx v24,off64,r6 + lvx v25,off80,r6 + lvx v26,off96,r6 + lvx v27,off112,r6 + lvx v28,0,r7 + lvx v29,off16,r7 + + ld r31,-8(r1) + ld r30,-16(r1) + ld r29,-24(r1) + ld r28,-32(r1) + ld r27,-40(r1) + ld r26,-48(r1) + ld r25,-56(r1) + + blr + +.Lfirst_warm_up_done: + lvx const1,0,r3 + addi r3,r3,16 + + VPMSUMD(v8,v16,const1) + VPMSUMD(v9,v17,const1) + VPMSUMD(v10,v18,const1) + VPMSUMD(v11,v19,const1) + VPMSUMD(v12,v20,const1) + VPMSUMD(v13,v21,const1) + VPMSUMD(v14,v22,const1) + VPMSUMD(v15,v23,const1) + + b .Lsecond_cool_down + +.Lshort: + cmpdi r5,0 + beq .Lzero + + addis r3,r2,.short_constants@toc@ha + addi r3,r3,.short_constants@toc@l + + /* Calculate where in the constant table we need to start */ + subfic r6,r5,256 + add r3,r3,r6 + + /* How many 16 byte chunks? */ + srdi r7,r5,4 + mtctr r7 + + vxor v19,v19,v19 + vxor v20,v20,v20 + + lvx v0,0,r4 + lvx v16,0,r3 + VPERM(v0,v0,v16,byteswap) + vxor v0,v0,v8 /* xor in initial value */ + VPMSUMW(v0,v0,v16) + bdz .Lv0 + + lvx v1,off16,r4 + lvx v17,off16,r3 + VPERM(v1,v1,v17,byteswap) + VPMSUMW(v1,v1,v17) + bdz .Lv1 + + lvx v2,off32,r4 + lvx v16,off32,r3 + VPERM(v2,v2,v16,byteswap) + VPMSUMW(v2,v2,v16) + bdz .Lv2 + + lvx v3,off48,r4 + lvx v17,off48,r3 + VPERM(v3,v3,v17,byteswap) + VPMSUMW(v3,v3,v17) + bdz .Lv3 + + lvx v4,off64,r4 + lvx v16,off64,r3 + VPERM(v4,v4,v16,byteswap) + VPMSUMW(v4,v4,v16) + bdz .Lv4 + + lvx v5,off80,r4 + lvx v17,off80,r3 + VPERM(v5,v5,v17,byteswap) + VPMSUMW(v5,v5,v17) + bdz .Lv5 + + lvx v6,off96,r4 + lvx v16,off96,r3 + VPERM(v6,v6,v16,byteswap) + VPMSUMW(v6,v6,v16) + bdz .Lv6 + + lvx v7,off112,r4 + lvx v17,off112,r3 + VPERM(v7,v7,v17,byteswap) + VPMSUMW(v7,v7,v17) + bdz .Lv7 + + addi r3,r3,128 + addi r4,r4,128 + + lvx v8,0,r4 + lvx v16,0,r3 + VPERM(v8,v8,v16,byteswap) + VPMSUMW(v8,v8,v16) + bdz .Lv8 + + lvx v9,off16,r4 + lvx v17,off16,r3 + VPERM(v9,v9,v17,byteswap) + VPMSUMW(v9,v9,v17) + bdz .Lv9 + + lvx v10,off32,r4 + lvx v16,off32,r3 + VPERM(v10,v10,v16,byteswap) + VPMSUMW(v10,v10,v16) + bdz .Lv10 + + lvx v11,off48,r4 + lvx v17,off48,r3 + VPERM(v11,v11,v17,byteswap) + VPMSUMW(v11,v11,v17) + bdz .Lv11 + + lvx v12,off64,r4 + lvx v16,off64,r3 + VPERM(v12,v12,v16,byteswap) + VPMSUMW(v12,v12,v16) + bdz .Lv12 + + lvx v13,off80,r4 + lvx v17,off80,r3 + VPERM(v13,v13,v17,byteswap) + VPMSUMW(v13,v13,v17) + bdz .Lv13 + + lvx v14,off96,r4 + lvx v16,off96,r3 + VPERM(v14,v14,v16,byteswap) + VPMSUMW(v14,v14,v16) + bdz .Lv14 + + lvx v15,off112,r4 + lvx v17,off112,r3 + VPERM(v15,v15,v17,byteswap) + VPMSUMW(v15,v15,v17) + +.Lv15: vxor v19,v19,v15 +.Lv14: vxor v20,v20,v14 +.Lv13: vxor v19,v19,v13 +.Lv12: vxor v20,v20,v12 +.Lv11: vxor v19,v19,v11 +.Lv10: vxor v20,v20,v10 +.Lv9: vxor v19,v19,v9 +.Lv8: vxor v20,v20,v8 +.Lv7: vxor v19,v19,v7 +.Lv6: vxor v20,v20,v6 +.Lv5: vxor v19,v19,v5 +.Lv4: vxor v20,v20,v4 +.Lv3: vxor v19,v19,v3 +.Lv2: vxor v20,v20,v2 +.Lv1: vxor v19,v19,v1 +.Lv0: vxor v20,v20,v0 + + vxor v0,v19,v20 + + b .Lbarrett_reduction + +.Lzero: + mr r3,r10 + b .Lout + +FUNC_END(__crc32_vpmsum) diff --git a/util/crc32c_ppc_constants.h b/util/crc32c_ppc_constants.h new file mode 100644 index 000000000..aa5ea98f8 --- /dev/null +++ b/util/crc32c_ppc_constants.h @@ -0,0 +1,893 @@ +// Copyright (C) 2015, 2017 International Business Machines Corp. +// All rights reserved. +// This source code is licensed under the BSD-style license found in the +// LICENSE file in the root directory of this source tree. An additional grant +// of patent rights can be found in the PATENTS file in the same directory. +// This source code is also licensed under the GPLv2 license found in the +// COPYING file in the root directory of this source tree. +#ifndef CRC32C_PPC_CONST_H +#define CRC32C_PPC_CONST_H +#define CRC 0x1edc6f41 +#define REFLECT +#define CRC_XOR + +#ifndef __ASSEMBLY__ +#ifdef CRC_TABLE +static const unsigned int crc_table[] = { + 0x00000000, 0xf26b8303, 0xe13b70f7, 0x1350f3f4, 0xc79a971f, 0x35f1141c, + 0x26a1e7e8, 0xd4ca64eb, 0x8ad958cf, 0x78b2dbcc, 0x6be22838, 0x9989ab3b, + 0x4d43cfd0, 0xbf284cd3, 0xac78bf27, 0x5e133c24, 0x105ec76f, 0xe235446c, + 0xf165b798, 0x030e349b, 0xd7c45070, 0x25afd373, 0x36ff2087, 0xc494a384, + 0x9a879fa0, 0x68ec1ca3, 0x7bbcef57, 0x89d76c54, 0x5d1d08bf, 0xaf768bbc, + 0xbc267848, 0x4e4dfb4b, 0x20bd8ede, 0xd2d60ddd, 0xc186fe29, 0x33ed7d2a, + 0xe72719c1, 0x154c9ac2, 0x061c6936, 0xf477ea35, 0xaa64d611, 0x580f5512, + 0x4b5fa6e6, 0xb93425e5, 0x6dfe410e, 0x9f95c20d, 0x8cc531f9, 0x7eaeb2fa, + 0x30e349b1, 0xc288cab2, 0xd1d83946, 0x23b3ba45, 0xf779deae, 0x05125dad, + 0x1642ae59, 0xe4292d5a, 0xba3a117e, 0x4851927d, 0x5b016189, 0xa96ae28a, + 0x7da08661, 0x8fcb0562, 0x9c9bf696, 0x6ef07595, 0x417b1dbc, 0xb3109ebf, + 0xa0406d4b, 0x522bee48, 0x86e18aa3, 0x748a09a0, 0x67dafa54, 0x95b17957, + 0xcba24573, 0x39c9c670, 0x2a993584, 0xd8f2b687, 0x0c38d26c, 0xfe53516f, + 0xed03a29b, 0x1f682198, 0x5125dad3, 0xa34e59d0, 0xb01eaa24, 0x42752927, + 0x96bf4dcc, 0x64d4cecf, 0x77843d3b, 0x85efbe38, 0xdbfc821c, 0x2997011f, + 0x3ac7f2eb, 0xc8ac71e8, 0x1c661503, 0xee0d9600, 0xfd5d65f4, 0x0f36e6f7, + 0x61c69362, 0x93ad1061, 0x80fde395, 0x72966096, 0xa65c047d, 0x5437877e, + 0x4767748a, 0xb50cf789, 0xeb1fcbad, 0x197448ae, 0x0a24bb5a, 0xf84f3859, + 0x2c855cb2, 0xdeeedfb1, 0xcdbe2c45, 0x3fd5af46, 0x7198540d, 0x83f3d70e, + 0x90a324fa, 0x62c8a7f9, 0xb602c312, 0x44694011, 0x5739b3e5, 0xa55230e6, + 0xfb410cc2, 0x092a8fc1, 0x1a7a7c35, 0xe811ff36, 0x3cdb9bdd, 0xceb018de, + 0xdde0eb2a, 0x2f8b6829, 0x82f63b78, 0x709db87b, 0x63cd4b8f, 0x91a6c88c, + 0x456cac67, 0xb7072f64, 0xa457dc90, 0x563c5f93, 0x082f63b7, 0xfa44e0b4, + 0xe9141340, 0x1b7f9043, 0xcfb5f4a8, 0x3dde77ab, 0x2e8e845f, 0xdce5075c, + 0x92a8fc17, 0x60c37f14, 0x73938ce0, 0x81f80fe3, 0x55326b08, 0xa759e80b, + 0xb4091bff, 0x466298fc, 0x1871a4d8, 0xea1a27db, 0xf94ad42f, 0x0b21572c, + 0xdfeb33c7, 0x2d80b0c4, 0x3ed04330, 0xccbbc033, 0xa24bb5a6, 0x502036a5, + 0x4370c551, 0xb11b4652, 0x65d122b9, 0x97baa1ba, 0x84ea524e, 0x7681d14d, + 0x2892ed69, 0xdaf96e6a, 0xc9a99d9e, 0x3bc21e9d, 0xef087a76, 0x1d63f975, + 0x0e330a81, 0xfc588982, 0xb21572c9, 0x407ef1ca, 0x532e023e, 0xa145813d, + 0x758fe5d6, 0x87e466d5, 0x94b49521, 0x66df1622, 0x38cc2a06, 0xcaa7a905, + 0xd9f75af1, 0x2b9cd9f2, 0xff56bd19, 0x0d3d3e1a, 0x1e6dcdee, 0xec064eed, + 0xc38d26c4, 0x31e6a5c7, 0x22b65633, 0xd0ddd530, 0x0417b1db, 0xf67c32d8, + 0xe52cc12c, 0x1747422f, 0x49547e0b, 0xbb3ffd08, 0xa86f0efc, 0x5a048dff, + 0x8ecee914, 0x7ca56a17, 0x6ff599e3, 0x9d9e1ae0, 0xd3d3e1ab, 0x21b862a8, + 0x32e8915c, 0xc083125f, 0x144976b4, 0xe622f5b7, 0xf5720643, 0x07198540, + 0x590ab964, 0xab613a67, 0xb831c993, 0x4a5a4a90, 0x9e902e7b, 0x6cfbad78, + 0x7fab5e8c, 0x8dc0dd8f, 0xe330a81a, 0x115b2b19, 0x020bd8ed, 0xf0605bee, + 0x24aa3f05, 0xd6c1bc06, 0xc5914ff2, 0x37faccf1, 0x69e9f0d5, 0x9b8273d6, + 0x88d28022, 0x7ab90321, 0xae7367ca, 0x5c18e4c9, 0x4f48173d, 0xbd23943e, + 0xf36e6f75, 0x0105ec76, 0x12551f82, 0xe03e9c81, 0x34f4f86a, 0xc69f7b69, + 0xd5cf889d, 0x27a40b9e, 0x79b737ba, 0x8bdcb4b9, 0x988c474d, 0x6ae7c44e, + 0xbe2da0a5, 0x4c4623a6, 0x5f16d052, 0xad7d5351, +}; + +#endif + +#else +#define MAX_SIZE 32768 +.constants : + + /* Reduce 262144 kbits to 1024 bits */ + /* x^261120 mod p(x)` << 1, x^261184 mod p(x)` << 1 */ + .octa 0x00000000b6ca9e20000000009c37c408 + + /* x^260096 mod p(x)` << 1, x^260160 mod p(x)` << 1 */ + .octa 0x00000000350249a800000001b51df26c + + /* x^259072 mod p(x)` << 1, x^259136 mod p(x)` << 1 */ + .octa 0x00000001862dac54000000000724b9d0 + + /* x^258048 mod p(x)` << 1, x^258112 mod p(x)` << 1 */ + .octa 0x00000001d87fb48c00000001c00532fe + + /* x^257024 mod p(x)` << 1, x^257088 mod p(x)` << 1 */ + .octa 0x00000001f39b699e00000000f05a9362 + + /* x^256000 mod p(x)` << 1, x^256064 mod p(x)` << 1 */ + .octa 0x0000000101da11b400000001e1007970 + + /* x^254976 mod p(x)` << 1, x^255040 mod p(x)` << 1 */ + .octa 0x00000001cab571e000000000a57366ee + + /* x^253952 mod p(x)` << 1, x^254016 mod p(x)` << 1 */ + .octa 0x00000000c7020cfe0000000192011284 + + /* x^252928 mod p(x)` << 1, x^252992 mod p(x)` << 1 */ + .octa 0x00000000cdaed1ae0000000162716d9a + + /* x^251904 mod p(x)` << 1, x^251968 mod p(x)` << 1 */ + .octa 0x00000001e804effc00000000cd97ecde + + /* x^250880 mod p(x)` << 1, x^250944 mod p(x)` << 1 */ + .octa 0x0000000077c3ea3a0000000058812bc0 + + /* x^249856 mod p(x)` << 1, x^249920 mod p(x)` << 1 */ + .octa 0x0000000068df31b40000000088b8c12e + + /* x^248832 mod p(x)` << 1, x^248896 mod p(x)` << 1 */ + .octa 0x00000000b059b6c200000001230b234c + + /* x^247808 mod p(x)` << 1, x^247872 mod p(x)` << 1 */ + .octa 0x0000000145fb8ed800000001120b416e + + /* x^246784 mod p(x)` << 1, x^246848 mod p(x)` << 1 */ + .octa 0x00000000cbc0916800000001974aecb0 + + /* x^245760 mod p(x)` << 1, x^245824 mod p(x)` << 1 */ + .octa 0x000000005ceeedc2000000008ee3f226 + + /* x^244736 mod p(x)` << 1, x^244800 mod p(x)` << 1 */ + .octa 0x0000000047d74e8600000001089aba9a + + /* x^243712 mod p(x)` << 1, x^243776 mod p(x)` << 1 */ + .octa 0x00000001407e9e220000000065113872 + + /* x^242688 mod p(x)` << 1, x^242752 mod p(x)` << 1 */ + .octa 0x00000001da967bda000000005c07ec10 + + /* x^241664 mod p(x)` << 1, x^241728 mod p(x)` << 1 */ + .octa 0x000000006c8983680000000187590924 + + /* x^240640 mod p(x)` << 1, x^240704 mod p(x)` << 1 */ + .octa 0x00000000f2d14c9800000000e35da7c6 + + /* x^239616 mod p(x)` << 1, x^239680 mod p(x)` << 1 */ + .octa 0x00000001993c6ad4000000000415855a + + /* x^238592 mod p(x)` << 1, x^238656 mod p(x)` << 1 */ + .octa 0x000000014683d1ac0000000073617758 + + /* x^237568 mod p(x)` << 1, x^237632 mod p(x)` << 1 */ + .octa 0x00000001a7c93e6c0000000176021d28 + + /* x^236544 mod p(x)` << 1, x^236608 mod p(x)` << 1 */ + .octa 0x000000010211e90a00000001c358fd0a + + /* x^235520 mod p(x)` << 1, x^235584 mod p(x)` << 1 */ + .octa 0x000000001119403e00000001ff7a2c18 + + /* x^234496 mod p(x)` << 1, x^234560 mod p(x)` << 1 */ + .octa 0x000000001c3261aa00000000f2d9f7e4 + + /* x^233472 mod p(x)` << 1, x^233536 mod p(x)` << 1 */ + .octa 0x000000014e37a634000000016cf1f9c8 + + /* x^232448 mod p(x)` << 1, x^232512 mod p(x)` << 1 */ + .octa 0x0000000073786c0c000000010af9279a + + /* x^231424 mod p(x)` << 1, x^231488 mod p(x)` << 1 */ + .octa 0x000000011dc037f80000000004f101e8 + + /* x^230400 mod p(x)` << 1, x^230464 mod p(x)` << 1 */ + .octa 0x0000000031433dfc0000000070bcf184 + + /* x^229376 mod p(x)` << 1, x^229440 mod p(x)` << 1 */ + .octa 0x000000009cde8348000000000a8de642 + + /* x^228352 mod p(x)` << 1, x^228416 mod p(x)` << 1 */ + .octa 0x0000000038d3c2a60000000062ea130c + + /* x^227328 mod p(x)` << 1, x^227392 mod p(x)` << 1 */ + .octa 0x000000011b25f26000000001eb31cbb2 + + /* x^226304 mod p(x)` << 1, x^226368 mod p(x)` << 1 */ + .octa 0x000000001629e6f00000000170783448 + + /* x^225280 mod p(x)` << 1, x^225344 mod p(x)` << 1 */ + .octa 0x0000000160838b4c00000001a684b4c6 + + /* x^224256 mod p(x)` << 1, x^224320 mod p(x)` << 1 */ + .octa 0x000000007a44011c00000000253ca5b4 + + /* x^223232 mod p(x)` << 1, x^223296 mod p(x)` << 1 */ + .octa 0x00000000226f417a0000000057b4b1e2 + + /* x^222208 mod p(x)` << 1, x^222272 mod p(x)` << 1 */ + .octa 0x0000000045eb2eb400000000b6bd084c + + /* x^221184 mod p(x)` << 1, x^221248 mod p(x)` << 1 */ + .octa 0x000000014459d70c0000000123c2d592 + + /* x^220160 mod p(x)` << 1, x^220224 mod p(x)` << 1 */ + .octa 0x00000001d406ed8200000000159dafce + + /* x^219136 mod p(x)` << 1, x^219200 mod p(x)` << 1 */ + .octa 0x0000000160c8e1a80000000127e1a64e + + /* x^218112 mod p(x)` << 1, x^218176 mod p(x)` << 1 */ + .octa 0x0000000027ba80980000000056860754 + + /* x^217088 mod p(x)` << 1, x^217152 mod p(x)` << 1 */ + .octa 0x000000006d92d01800000001e661aae8 + + /* x^216064 mod p(x)` << 1, x^216128 mod p(x)` << 1 */ + .octa 0x000000012ed7e3f200000000f82c6166 + + /* x^215040 mod p(x)` << 1, x^215104 mod p(x)` << 1 */ + .octa 0x000000002dc8778800000000c4f9c7ae + + /* x^214016 mod p(x)` << 1, x^214080 mod p(x)` << 1 */ + .octa 0x0000000018240bb80000000074203d20 + + /* x^212992 mod p(x)` << 1, x^213056 mod p(x)` << 1 */ + .octa 0x000000001ad381580000000198173052 + + /* x^211968 mod p(x)` << 1, x^212032 mod p(x)` << 1 */ + .octa 0x00000001396b78f200000001ce8aba54 + + /* x^210944 mod p(x)` << 1, x^211008 mod p(x)` << 1 */ + .octa 0x000000011a68133400000001850d5d94 + + /* x^209920 mod p(x)` << 1, x^209984 mod p(x)` << 1 */ + .octa 0x000000012104732e00000001d609239c + + /* x^208896 mod p(x)` << 1, x^208960 mod p(x)` << 1 */ + .octa 0x00000000a140d90c000000001595f048 + + /* x^207872 mod p(x)` << 1, x^207936 mod p(x)` << 1 */ + .octa 0x00000001b7215eda0000000042ccee08 + + /* x^206848 mod p(x)` << 1, x^206912 mod p(x)` << 1 */ + .octa 0x00000001aaf1df3c000000010a389d74 + + /* x^205824 mod p(x)` << 1, x^205888 mod p(x)` << 1 */ + .octa 0x0000000029d15b8a000000012a840da6 + + /* x^204800 mod p(x)` << 1, x^204864 mod p(x)` << 1 */ + .octa 0x00000000f1a96922000000001d181c0c + + /* x^203776 mod p(x)` << 1, x^203840 mod p(x)` << 1 */ + .octa 0x00000001ac80d03c0000000068b7d1f6 + + /* x^202752 mod p(x)` << 1, x^202816 mod p(x)` << 1 */ + .octa 0x000000000f11d56a000000005b0f14fc + + /* x^201728 mod p(x)` << 1, x^201792 mod p(x)` << 1 */ + .octa 0x00000001f1c022a20000000179e9e730 + + /* x^200704 mod p(x)` << 1, x^200768 mod p(x)` << 1 */ + .octa 0x0000000173d00ae200000001ce1368d6 + + /* x^199680 mod p(x)` << 1, x^199744 mod p(x)` << 1 */ + .octa 0x00000001d4ffe4ac0000000112c3a84c + + /* x^198656 mod p(x)` << 1, x^198720 mod p(x)` << 1 */ + .octa 0x000000016edc5ae400000000de940fee + + /* x^197632 mod p(x)` << 1, x^197696 mod p(x)` << 1 */ + .octa 0x00000001f1a0214000000000fe896b7e + + /* x^196608 mod p(x)` << 1, x^196672 mod p(x)` << 1 */ + .octa 0x00000000ca0b28a000000001f797431c + + /* x^195584 mod p(x)` << 1, x^195648 mod p(x)` << 1 */ + .octa 0x00000001928e30a20000000053e989ba + + /* x^194560 mod p(x)` << 1, x^194624 mod p(x)` << 1 */ + .octa 0x0000000097b1b002000000003920cd16 + + /* x^193536 mod p(x)` << 1, x^193600 mod p(x)` << 1 */ + .octa 0x00000000b15bf90600000001e6f579b8 + + /* x^192512 mod p(x)` << 1, x^192576 mod p(x)` << 1 */ + .octa 0x00000000411c5d52000000007493cb0a + + /* x^191488 mod p(x)` << 1, x^191552 mod p(x)` << 1 */ + .octa 0x00000001c36f330000000001bdd376d8 + + /* x^190464 mod p(x)` << 1, x^190528 mod p(x)` << 1 */ + .octa 0x00000001119227e0000000016badfee6 + + /* x^189440 mod p(x)` << 1, x^189504 mod p(x)` << 1 */ + .octa 0x00000000114d47020000000071de5c58 + + /* x^188416 mod p(x)` << 1, x^188480 mod p(x)` << 1 */ + .octa 0x00000000458b5b9800000000453f317c + + /* x^187392 mod p(x)` << 1, x^187456 mod p(x)` << 1 */ + .octa 0x000000012e31fb8e0000000121675cce + + /* x^186368 mod p(x)` << 1, x^186432 mod p(x)` << 1 */ + .octa 0x000000005cf619d800000001f409ee92 + + /* x^185344 mod p(x)` << 1, x^185408 mod p(x)` << 1 */ + .octa 0x0000000063f4d8b200000000f36b9c88 + + /* x^184320 mod p(x)` << 1, x^184384 mod p(x)` << 1 */ + .octa 0x000000004138dc8a0000000036b398f4 + + /* x^183296 mod p(x)` << 1, x^183360 mod p(x)` << 1 */ + .octa 0x00000001d29ee8e000000001748f9adc + + /* x^182272 mod p(x)` << 1, x^182336 mod p(x)` << 1 */ + .octa 0x000000006a08ace800000001be94ec00 + + /* x^181248 mod p(x)` << 1, x^181312 mod p(x)` << 1 */ + .octa 0x0000000127d4201000000000b74370d6 + + /* x^180224 mod p(x)` << 1, x^180288 mod p(x)` << 1 */ + .octa 0x0000000019d76b6200000001174d0b98 + + /* x^179200 mod p(x)` << 1, x^179264 mod p(x)` << 1 */ + .octa 0x00000001b1471f6e00000000befc06a4 + + /* x^178176 mod p(x)` << 1, x^178240 mod p(x)` << 1 */ + .octa 0x00000001f64c19cc00000001ae125288 + + /* x^177152 mod p(x)` << 1, x^177216 mod p(x)` << 1 */ + .octa 0x00000000003c0ea00000000095c19b34 + + /* x^176128 mod p(x)` << 1, x^176192 mod p(x)` << 1 */ + .octa 0x000000014d73abf600000001a78496f2 + + /* x^175104 mod p(x)` << 1, x^175168 mod p(x)` << 1 */ + .octa 0x00000001620eb84400000001ac5390a0 + + /* x^174080 mod p(x)` << 1, x^174144 mod p(x)` << 1 */ + .octa 0x0000000147655048000000002a80ed6e + + /* x^173056 mod p(x)` << 1, x^173120 mod p(x)` << 1 */ + .octa 0x0000000067b5077e00000001fa9b0128 + + /* x^172032 mod p(x)` << 1, x^172096 mod p(x)` << 1 */ + .octa 0x0000000010ffe20600000001ea94929e + + /* x^171008 mod p(x)` << 1, x^171072 mod p(x)` << 1 */ + .octa 0x000000000fee8f1e0000000125f4305c + + /* x^169984 mod p(x)` << 1, x^170048 mod p(x)` << 1 */ + .octa 0x00000001da26fbae00000001471e2002 + + /* x^168960 mod p(x)` << 1, x^169024 mod p(x)` << 1 */ + .octa 0x00000001b3a8bd880000000132d2253a + + /* x^167936 mod p(x)` << 1, x^168000 mod p(x)` << 1 */ + .octa 0x00000000e8f3898e00000000f26b3592 + + /* x^166912 mod p(x)` << 1, x^166976 mod p(x)` << 1 */ + .octa 0x00000000b0d0d28c00000000bc8b67b0 + + /* x^165888 mod p(x)` << 1, x^165952 mod p(x)` << 1 */ + .octa 0x0000000030f2a798000000013a826ef2 + + /* x^164864 mod p(x)` << 1, x^164928 mod p(x)` << 1 */ + .octa 0x000000000fba10020000000081482c84 + + /* x^163840 mod p(x)` << 1, x^163904 mod p(x)` << 1 */ + .octa 0x00000000bdb9bd7200000000e77307c2 + + /* x^162816 mod p(x)` << 1, x^162880 mod p(x)` << 1 */ + .octa 0x0000000075d3bf5a00000000d4a07ec8 + + /* x^161792 mod p(x)` << 1, x^161856 mod p(x)` << 1 */ + .octa 0x00000000ef1f98a00000000017102100 + + /* x^160768 mod p(x)` << 1, x^160832 mod p(x)` << 1 */ + .octa 0x00000000689c760200000000db406486 + + /* x^159744 mod p(x)` << 1, x^159808 mod p(x)` << 1 */ + .octa 0x000000016d5fa5fe0000000192db7f88 + + /* x^158720 mod p(x)` << 1, x^158784 mod p(x)` << 1 */ + .octa 0x00000001d0d2b9ca000000018bf67b1e + + /* x^157696 mod p(x)` << 1, x^157760 mod p(x)` << 1 */ + .octa 0x0000000041e7b470000000007c09163e + + /* x^156672 mod p(x)` << 1, x^156736 mod p(x)` << 1 */ + .octa 0x00000001cbb6495e000000000adac060 + + /* x^155648 mod p(x)` << 1, x^155712 mod p(x)` << 1 */ + .octa 0x000000010052a0b000000000bd8316ae + + /* x^154624 mod p(x)` << 1, x^154688 mod p(x)` << 1 */ + .octa 0x00000001d8effb5c000000019f09ab54 + + /* x^153600 mod p(x)` << 1, x^153664 mod p(x)` << 1 */ + .octa 0x00000001d969853c0000000125155542 + + /* x^152576 mod p(x)` << 1, x^152640 mod p(x)` << 1 */ + .octa 0x00000000523ccce2000000018fdb5882 + + /* x^151552 mod p(x)` << 1, x^151616 mod p(x)` << 1 */ + .octa 0x000000001e2436bc00000000e794b3f4 + + /* x^150528 mod p(x)` << 1, x^150592 mod p(x)` << 1 */ + .octa 0x00000000ddd1c3a2000000016f9bb022 + + /* x^149504 mod p(x)` << 1, x^149568 mod p(x)` << 1 */ + .octa 0x0000000019fcfe3800000000290c9978 + + /* x^148480 mod p(x)` << 1, x^148544 mod p(x)` << 1 */ + .octa 0x00000001ce95db640000000083c0f350 + + /* x^147456 mod p(x)` << 1, x^147520 mod p(x)` << 1 */ + .octa 0x00000000af5828060000000173ea6628 + + /* x^146432 mod p(x)` << 1, x^146496 mod p(x)` << 1 */ + .octa 0x00000001006388f600000001c8b4e00a + + /* x^145408 mod p(x)` << 1, x^145472 mod p(x)` << 1 */ + .octa 0x0000000179eca00a00000000de95d6aa + + /* x^144384 mod p(x)` << 1, x^144448 mod p(x)` << 1 */ + .octa 0x0000000122410a6a000000010b7f7248 + + /* x^143360 mod p(x)` << 1, x^143424 mod p(x)` << 1 */ + .octa 0x000000004288e87c00000001326e3a06 + + /* x^142336 mod p(x)` << 1, x^142400 mod p(x)` << 1 */ + .octa 0x000000016c5490da00000000bb62c2e6 + + /* x^141312 mod p(x)` << 1, x^141376 mod p(x)` << 1 */ + .octa 0x00000000d1c71f6e0000000156a4b2c2 + + /* x^140288 mod p(x)` << 1, x^140352 mod p(x)` << 1 */ + .octa 0x00000001b4ce08a6000000011dfe763a + + /* x^139264 mod p(x)` << 1, x^139328 mod p(x)` << 1 */ + .octa 0x00000001466ba60c000000007bcca8e2 + + /* x^138240 mod p(x)` << 1, x^138304 mod p(x)` << 1 */ + .octa 0x00000001f6c488a40000000186118faa + + /* x^137216 mod p(x)` << 1, x^137280 mod p(x)` << 1 */ + .octa 0x000000013bfb06820000000111a65a88 + + /* x^136192 mod p(x)` << 1, x^136256 mod p(x)` << 1 */ + .octa 0x00000000690e9e54000000003565e1c4 + + /* x^135168 mod p(x)` << 1, x^135232 mod p(x)` << 1 */ + .octa 0x00000000281346b6000000012ed02a82 + + /* x^134144 mod p(x)` << 1, x^134208 mod p(x)` << 1 */ + .octa 0x000000015646402400000000c486ecfc + + /* x^133120 mod p(x)` << 1, x^133184 mod p(x)` << 1 */ + .octa 0x000000016063a8dc0000000001b951b2 + + /* x^132096 mod p(x)` << 1, x^132160 mod p(x)` << 1 */ + .octa 0x0000000116a663620000000048143916 + + /* x^131072 mod p(x)` << 1, x^131136 mod p(x)` << 1 */ + .octa 0x000000017e8aa4d200000001dc2ae124 + + /* x^130048 mod p(x)` << 1, x^130112 mod p(x)` << 1 */ + .octa 0x00000001728eb10c00000001416c58d6 + + /* x^129024 mod p(x)` << 1, x^129088 mod p(x)` << 1 */ + .octa 0x00000001b08fd7fa00000000a479744a + + /* x^128000 mod p(x)` << 1, x^128064 mod p(x)` << 1 */ + .octa 0x00000001092a16e80000000096ca3a26 + + /* x^126976 mod p(x)` << 1, x^127040 mod p(x)` << 1 */ + .octa 0x00000000a505637c00000000ff223d4e + + /* x^125952 mod p(x)` << 1, x^126016 mod p(x)` << 1 */ + .octa 0x00000000d94869b2000000010e84da42 + + /* x^124928 mod p(x)` << 1, x^124992 mod p(x)` << 1 */ + .octa 0x00000001c8b203ae00000001b61ba3d0 + + /* x^123904 mod p(x)` << 1, x^123968 mod p(x)` << 1 */ + .octa 0x000000005704aea000000000680f2de8 + + /* x^122880 mod p(x)` << 1, x^122944 mod p(x)` << 1 */ + .octa 0x000000012e295fa2000000008772a9a8 + + /* x^121856 mod p(x)` << 1, x^121920 mod p(x)` << 1 */ + .octa 0x000000011d0908bc0000000155f295bc + + /* x^120832 mod p(x)` << 1, x^120896 mod p(x)` << 1 */ + .octa 0x0000000193ed97ea00000000595f9282 + + /* x^119808 mod p(x)` << 1, x^119872 mod p(x)` << 1 */ + .octa 0x000000013a0f1c520000000164b1c25a + + /* x^118784 mod p(x)` << 1, x^118848 mod p(x)` << 1 */ + .octa 0x000000010c2c40c000000000fbd67c50 + + /* x^117760 mod p(x)` << 1, x^117824 mod p(x)` << 1 */ + .octa 0x00000000ff6fac3e0000000096076268 + + /* x^116736 mod p(x)` << 1, x^116800 mod p(x)` << 1 */ + .octa 0x000000017b3609c000000001d288e4cc + + /* x^115712 mod p(x)` << 1, x^115776 mod p(x)` << 1 */ + .octa 0x0000000088c8c92200000001eaac1bdc + + /* x^114688 mod p(x)` << 1, x^114752 mod p(x)` << 1 */ + .octa 0x00000001751baae600000001f1ea39e2 + + /* x^113664 mod p(x)` << 1, x^113728 mod p(x)` << 1 */ + .octa 0x000000010795297200000001eb6506fc + + /* x^112640 mod p(x)` << 1, x^112704 mod p(x)` << 1 */ + .octa 0x0000000162b00abe000000010f806ffe + + /* x^111616 mod p(x)` << 1, x^111680 mod p(x)` << 1 */ + .octa 0x000000000d7b404c000000010408481e + + /* x^110592 mod p(x)` << 1, x^110656 mod p(x)` << 1 */ + .octa 0x00000000763b13d40000000188260534 + + /* x^109568 mod p(x)` << 1, x^109632 mod p(x)` << 1 */ + .octa 0x00000000f6dc22d80000000058fc73e0 + + /* x^108544 mod p(x)` << 1, x^108608 mod p(x)` << 1 */ + .octa 0x000000007daae06000000000391c59b8 + + /* x^107520 mod p(x)` << 1, x^107584 mod p(x)` << 1 */ + .octa 0x000000013359ab7c000000018b638400 + + /* x^106496 mod p(x)` << 1, x^106560 mod p(x)` << 1 */ + .octa 0x000000008add438a000000011738f5c4 + + /* x^105472 mod p(x)` << 1, x^105536 mod p(x)` << 1 */ + .octa 0x00000001edbefdea000000008cf7c6da + + /* x^104448 mod p(x)` << 1, x^104512 mod p(x)` << 1 */ + .octa 0x000000004104e0f800000001ef97fb16 + + /* x^103424 mod p(x)` << 1, x^103488 mod p(x)` << 1 */ + .octa 0x00000000b48a82220000000102130e20 + + /* x^102400 mod p(x)` << 1, x^102464 mod p(x)` << 1 */ + .octa 0x00000001bcb4684400000000db968898 + + /* x^101376 mod p(x)` << 1, x^101440 mod p(x)` << 1 */ + .octa 0x000000013293ce0a00000000b5047b5e + + /* x^100352 mod p(x)` << 1, x^100416 mod p(x)` << 1 */ + .octa 0x00000001710d0844000000010b90fdb2 + + /* x^99328 mod p(x)` << 1, x^99392 mod p(x)` << 1 */ + .octa 0x0000000117907f6e000000004834a32e + + /* x^98304 mod p(x)` << 1, x^98368 mod p(x)` << 1 */ + .octa 0x0000000087ddf93e0000000059c8f2b0 + + /* x^97280 mod p(x)` << 1, x^97344 mod p(x)` << 1 */ + .octa 0x000000005970e9b00000000122cec508 + + /* x^96256 mod p(x)` << 1, x^96320 mod p(x)` << 1 */ + .octa 0x0000000185b2b7d0000000000a330cda + + /* x^95232 mod p(x)` << 1, x^95296 mod p(x)` << 1 */ + .octa 0x00000001dcee0efc000000014a47148c + + /* x^94208 mod p(x)` << 1, x^94272 mod p(x)` << 1 */ + .octa 0x0000000030da27220000000042c61cb8 + + /* x^93184 mod p(x)` << 1, x^93248 mod p(x)` << 1 */ + .octa 0x000000012f925a180000000012fe6960 + + /* x^92160 mod p(x)` << 1, x^92224 mod p(x)` << 1 */ + .octa 0x00000000dd2e357c00000000dbda2c20 + + /* x^91136 mod p(x)` << 1, x^91200 mod p(x)` << 1 */ + .octa 0x00000000071c80de000000011122410c + + /* x^90112 mod p(x)` << 1, x^90176 mod p(x)` << 1 */ + .octa 0x000000011513140a00000000977b2070 + + /* x^89088 mod p(x)` << 1, x^89152 mod p(x)` << 1 */ + .octa 0x00000001df876e8e000000014050438e + + /* x^88064 mod p(x)` << 1, x^88128 mod p(x)` << 1 */ + .octa 0x000000015f81d6ce0000000147c840e8 + + /* x^87040 mod p(x)` << 1, x^87104 mod p(x)` << 1 */ + .octa 0x000000019dd94dbe00000001cc7c88ce + + /* x^86016 mod p(x)` << 1, x^86080 mod p(x)` << 1 */ + .octa 0x00000001373d206e00000001476b35a4 + + /* x^84992 mod p(x)` << 1, x^85056 mod p(x)` << 1 */ + .octa 0x00000000668ccade000000013d52d508 + + /* x^83968 mod p(x)` << 1, x^84032 mod p(x)` << 1 */ + .octa 0x00000001b192d268000000008e4be32e + + /* x^82944 mod p(x)` << 1, x^83008 mod p(x)` << 1 */ + .octa 0x00000000e30f3a7800000000024120fe + + /* x^81920 mod p(x)` << 1, x^81984 mod p(x)` << 1 */ + .octa 0x000000010ef1f7bc00000000ddecddb4 + + /* x^80896 mod p(x)` << 1, x^80960 mod p(x)` << 1 */ + .octa 0x00000001f5ac738000000000d4d403bc + + /* x^79872 mod p(x)` << 1, x^79936 mod p(x)` << 1 */ + .octa 0x000000011822ea7000000001734b89aa + + /* x^78848 mod p(x)` << 1, x^78912 mod p(x)` << 1 */ + .octa 0x00000000c3a33848000000010e7a58d6 + + /* x^77824 mod p(x)` << 1, x^77888 mod p(x)` << 1 */ + .octa 0x00000001bd151c2400000001f9f04e9c + + /* x^76800 mod p(x)` << 1, x^76864 mod p(x)` << 1 */ + .octa 0x0000000056002d7600000000b692225e + + /* x^75776 mod p(x)` << 1, x^75840 mod p(x)` << 1 */ + .octa 0x000000014657c4f4000000019b8d3f3e + + /* x^74752 mod p(x)` << 1, x^74816 mod p(x)` << 1 */ + .octa 0x0000000113742d7c00000001a874f11e + + /* x^73728 mod p(x)` << 1, x^73792 mod p(x)` << 1 */ + .octa 0x000000019c5920ba000000010d5a4254 + + /* x^72704 mod p(x)` << 1, x^72768 mod p(x)` << 1 */ + .octa 0x000000005216d2d600000000bbb2f5d6 + + /* x^71680 mod p(x)` << 1, x^71744 mod p(x)` << 1 */ + .octa 0x0000000136f5ad8a0000000179cc0e36 + + /* x^70656 mod p(x)` << 1, x^70720 mod p(x)` << 1 */ + .octa 0x000000018b07beb600000001dca1da4a + + /* x^69632 mod p(x)` << 1, x^69696 mod p(x)` << 1 */ + .octa 0x00000000db1e93b000000000feb1a192 + + /* x^68608 mod p(x)` << 1, x^68672 mod p(x)` << 1 */ + .octa 0x000000000b96fa3a00000000d1eeedd6 + + /* x^67584 mod p(x)` << 1, x^67648 mod p(x)` << 1 */ + .octa 0x00000001d9968af0000000008fad9bb4 + + /* x^66560 mod p(x)` << 1, x^66624 mod p(x)` << 1 */ + .octa 0x000000000e4a77a200000001884938e4 + + /* x^65536 mod p(x)` << 1, x^65600 mod p(x)` << 1 */ + .octa 0x00000000508c2ac800000001bc2e9bc0 + + /* x^64512 mod p(x)` << 1, x^64576 mod p(x)` << 1 */ + .octa 0x0000000021572a8000000001f9658a68 + + /* x^63488 mod p(x)` << 1, x^63552 mod p(x)` << 1 */ + .octa 0x00000001b859daf2000000001b9224fc + + /* x^62464 mod p(x)` << 1, x^62528 mod p(x)` << 1 */ + .octa 0x000000016f7884740000000055b2fb84 + + /* x^61440 mod p(x)` << 1, x^61504 mod p(x)` << 1 */ + .octa 0x00000001b438810e000000018b090348 + + /* x^60416 mod p(x)` << 1, x^60480 mod p(x)` << 1 */ + .octa 0x0000000095ddc6f2000000011ccbd5ea + + /* x^59392 mod p(x)` << 1, x^59456 mod p(x)` << 1 */ + .octa 0x00000001d977c20c0000000007ae47f8 + + /* x^58368 mod p(x)` << 1, x^58432 mod p(x)` << 1 */ + .octa 0x00000000ebedb99a0000000172acbec0 + + /* x^57344 mod p(x)` << 1, x^57408 mod p(x)` << 1 */ + .octa 0x00000001df9e9e9200000001c6e3ff20 + + /* x^56320 mod p(x)` << 1, x^56384 mod p(x)` << 1 */ + .octa 0x00000001a4a3f95200000000e1b38744 + + /* x^55296 mod p(x)` << 1, x^55360 mod p(x)` << 1 */ + .octa 0x00000000e2f5122000000000791585b2 + + /* x^54272 mod p(x)` << 1, x^54336 mod p(x)` << 1 */ + .octa 0x000000004aa01f3e00000000ac53b894 + + /* x^53248 mod p(x)` << 1, x^53312 mod p(x)` << 1 */ + .octa 0x00000000b3e90a5800000001ed5f2cf4 + + /* x^52224 mod p(x)` << 1, x^52288 mod p(x)` << 1 */ + .octa 0x000000000c9ca2aa00000001df48b2e0 + + /* x^51200 mod p(x)` << 1, x^51264 mod p(x)` << 1 */ + .octa 0x000000015168231600000000049c1c62 + + /* x^50176 mod p(x)` << 1, x^50240 mod p(x)` << 1 */ + .octa 0x0000000036fce78c000000017c460c12 + + /* x^49152 mod p(x)` << 1, x^49216 mod p(x)` << 1 */ + .octa 0x000000009037dc10000000015be4da7e + + /* x^48128 mod p(x)` << 1, x^48192 mod p(x)` << 1 */ + .octa 0x00000000d3298582000000010f38f668 + + /* x^47104 mod p(x)` << 1, x^47168 mod p(x)` << 1 */ + .octa 0x00000001b42e8ad60000000039f40a00 + + /* x^46080 mod p(x)` << 1, x^46144 mod p(x)` << 1 */ + .octa 0x00000000142a983800000000bd4c10c4 + + /* x^45056 mod p(x)` << 1, x^45120 mod p(x)` << 1 */ + .octa 0x0000000109c7f1900000000042db1d98 + + /* x^44032 mod p(x)` << 1, x^44096 mod p(x)` << 1 */ + .octa 0x0000000056ff931000000001c905bae6 + + /* x^43008 mod p(x)` << 1, x^43072 mod p(x)` << 1 */ + .octa 0x00000001594513aa00000000069d40ea + + /* x^41984 mod p(x)` << 1, x^42048 mod p(x)` << 1 */ + .octa 0x00000001e3b5b1e8000000008e4fbad0 + + /* x^40960 mod p(x)` << 1, x^41024 mod p(x)` << 1 */ + .octa 0x000000011dd5fc080000000047bedd46 + + /* x^39936 mod p(x)` << 1, x^40000 mod p(x)` << 1 */ + .octa 0x00000001675f0cc20000000026396bf8 + + /* x^38912 mod p(x)` << 1, x^38976 mod p(x)` << 1 */ + .octa 0x00000000d1c8dd4400000000379beb92 + + /* x^37888 mod p(x)` << 1, x^37952 mod p(x)` << 1 */ + .octa 0x0000000115ebd3d8000000000abae54a + + /* x^36864 mod p(x)` << 1, x^36928 mod p(x)` << 1 */ + .octa 0x00000001ecbd0dac0000000007e6a128 + + /* x^35840 mod p(x)` << 1, x^35904 mod p(x)` << 1 */ + .octa 0x00000000cdf67af2000000000ade29d2 + + /* x^34816 mod p(x)` << 1, x^34880 mod p(x)` << 1 */ + .octa 0x000000004c01ff4c00000000f974c45c + + /* x^33792 mod p(x)` << 1, x^33856 mod p(x)` << 1 */ + .octa 0x00000000f2d8657e00000000e77ac60a + + /* x^32768 mod p(x)` << 1, x^32832 mod p(x)` << 1 */ + .octa 0x000000006bae74c40000000145895816 + + /* x^31744 mod p(x)` << 1, x^31808 mod p(x)` << 1 */ + .octa 0x0000000152af8aa00000000038e362be + + /* x^30720 mod p(x)` << 1, x^30784 mod p(x)` << 1 */ + .octa 0x0000000004663802000000007f991a64 + + /* x^29696 mod p(x)` << 1, x^29760 mod p(x)` << 1 */ + .octa 0x00000001ab2f5afc00000000fa366d3a + + /* x^28672 mod p(x)` << 1, x^28736 mod p(x)` << 1 */ + .octa 0x0000000074a4ebd400000001a2bb34f0 + + /* x^27648 mod p(x)` << 1, x^27712 mod p(x)` << 1 */ + .octa 0x00000001d7ab3a4c0000000028a9981e + + /* x^26624 mod p(x)` << 1, x^26688 mod p(x)` << 1 */ + .octa 0x00000001a8da60c600000001dbc672be + + /* x^25600 mod p(x)` << 1, x^25664 mod p(x)` << 1 */ + .octa 0x000000013cf6382000000000b04d77f6 + + /* x^24576 mod p(x)` << 1, x^24640 mod p(x)` << 1 */ + .octa 0x00000000bec12e1e0000000124400d96 + + /* x^23552 mod p(x)` << 1, x^23616 mod p(x)` << 1 */ + .octa 0x00000001c6368010000000014ca4b414 + + /* x^22528 mod p(x)` << 1, x^22592 mod p(x)` << 1 */ + .octa 0x00000001e6e78758000000012fe2c938 + + /* x^21504 mod p(x)` << 1, x^21568 mod p(x)` << 1 */ + .octa 0x000000008d7f2b3c00000001faed01e6 + + /* x^20480 mod p(x)` << 1, x^20544 mod p(x)` << 1 */ + .octa 0x000000016b4a156e000000007e80ecfe + + /* x^19456 mod p(x)` << 1, x^19520 mod p(x)` << 1 */ + .octa 0x00000001c63cfeb60000000098daee94 + + /* x^18432 mod p(x)` << 1, x^18496 mod p(x)` << 1 */ + .octa 0x000000015f902670000000010a04edea + + /* x^17408 mod p(x)` << 1, x^17472 mod p(x)` << 1 */ + .octa 0x00000001cd5de11e00000001c00b4524 + + /* x^16384 mod p(x)` << 1, x^16448 mod p(x)` << 1 */ + .octa 0x000000001acaec540000000170296550 + + /* x^15360 mod p(x)` << 1, x^15424 mod p(x)` << 1 */ + .octa 0x000000002bd0ca780000000181afaa48 + + /* x^14336 mod p(x)` << 1, x^14400 mod p(x)` << 1 */ + .octa 0x0000000032d63d5c0000000185a31ffa + + /* x^13312 mod p(x)` << 1, x^13376 mod p(x)` << 1 */ + .octa 0x000000001c6d4e4c000000002469f608 + + /* x^12288 mod p(x)` << 1, x^12352 mod p(x)` << 1 */ + .octa 0x0000000106a60b92000000006980102a + + /* x^11264 mod p(x)` << 1, x^11328 mod p(x)` << 1 */ + .octa 0x00000000d3855e120000000111ea9ca8 + + /* x^10240 mod p(x)` << 1, x^10304 mod p(x)` << 1 */ + .octa 0x00000000e312563600000001bd1d29ce + + /* x^9216 mod p(x)` << 1, x^9280 mod p(x)` << 1 */ + .octa 0x000000009e8f7ea400000001b34b9580 + + /* x^8192 mod p(x)` << 1, x^8256 mod p(x)` << 1 */ + .octa 0x00000001c82e562c000000003076054e + + /* x^7168 mod p(x)` << 1, x^7232 mod p(x)` << 1 */ + .octa 0x00000000ca9f09ce000000012a608ea4 + + /* x^6144 mod p(x)` << 1, x^6208 mod p(x)` << 1 */ + .octa 0x00000000c63764e600000000784d05fe + + /* x^5120 mod p(x)` << 1, x^5184 mod p(x)` << 1 */ + .octa 0x0000000168d2e49e000000016ef0d82a + + /* x^4096 mod p(x)` << 1, x^4160 mod p(x)` << 1 */ + .octa 0x00000000e986c1480000000075bda454 + + /* x^3072 mod p(x)` << 1, x^3136 mod p(x)` << 1 */ + .octa 0x00000000cfb65894000000003dc0a1c4 + + /* x^2048 mod p(x)` << 1, x^2112 mod p(x)` << 1 */ + .octa 0x0000000111cadee400000000e9a5d8be + + /* x^1024 mod p(x)` << 1, x^1088 mod p(x)` << 1 */ + .octa 0x0000000171fb63ce00000001609bc4b4 + + .short_constants : + + /* Reduce final 1024-2048 bits to 64 bits, shifting 32 bits to include the + trailing 32 bits of zeros */ + /* x^1952 mod p(x)`, x^1984 mod p(x)`, x^2016 mod p(x)`, x^2048 mod p(x)` */ + .octa 0x7fec2963e5bf80485cf015c388e56f72 + + /* x^1824 mod p(x)`, x^1856 mod p(x)`, x^1888 mod p(x)`, x^1920 mod p(x)` */ + .octa 0x38e888d4844752a9963a18920246e2e6 + + /* x^1696 mod p(x)`, x^1728 mod p(x)`, x^1760 mod p(x)`, x^1792 mod p(x)` */ + .octa 0x42316c00730206ad419a441956993a31 + + /* x^1568 mod p(x)`, x^1600 mod p(x)`, x^1632 mod p(x)`, x^1664 mod p(x)` */ + .octa 0x543d5c543e65ddf9924752ba2b830011 + + /* x^1440 mod p(x)`, x^1472 mod p(x)`, x^1504 mod p(x)`, x^1536 mod p(x)` */ + .octa 0x78e87aaf56767c9255bd7f9518e4a304 + + /* x^1312 mod p(x)`, x^1344 mod p(x)`, x^1376 mod p(x)`, x^1408 mod p(x)` */ + .octa 0x8f68fcec1903da7f6d76739fe0553f1e + + /* x^1184 mod p(x)`, x^1216 mod p(x)`, x^1248 mod p(x)`, x^1280 mod p(x)` */ + .octa 0x3f4840246791d588c133722b1fe0b5c3 + + /* x^1056 mod p(x)`, x^1088 mod p(x)`, x^1120 mod p(x)`, x^1152 mod p(x)` */ + .octa 0x34c96751b04de25a64b67ee0e55ef1f3 + + /* x^928 mod p(x)`, x^960 mod p(x)`, x^992 mod p(x)`, x^1024 mod p(x)` */ + .octa 0x156c8e180b4a395b069db049b8fdb1e7 + + /* x^800 mod p(x)`, x^832 mod p(x)`, x^864 mod p(x)`, x^896 mod p(x)` */ + .octa 0xe0b99ccbe661f7bea11bfaf3c9e90b9e + + /* x^672 mod p(x)`, x^704 mod p(x)`, x^736 mod p(x)`, x^768 mod p(x)` */ + .octa 0x041d37768cd75659817cdc5119b29a35 + + /* x^544 mod p(x)`, x^576 mod p(x)`, x^608 mod p(x)`, x^640 mod p(x)` */ + .octa 0x3a0777818cfaa9651ce9d94b36c41f1c + + /* x^416 mod p(x)`, x^448 mod p(x)`, x^480 mod p(x)`, x^512 mod p(x)` */ + .octa 0x0e148e8252377a554f256efcb82be955 + + /* x^288 mod p(x)`, x^320 mod p(x)`, x^352 mod p(x)`, x^384 mod p(x)` */ + .octa 0x9c25531d19e65ddeec1631edb2dea967 + + /* x^160 mod p(x)`, x^192 mod p(x)`, x^224 mod p(x)`, x^256 mod p(x)` */ + .octa 0x790606ff9957c0a65d27e147510ac59a + + /* x^32 mod p(x)`, x^64 mod p(x)`, x^96 mod p(x)`, x^128 mod p(x)` */ + .octa 0x82f63b786ea2d55ca66805eb18b8ea18 + + .barrett_constants : + /* 33 bit reflected Barrett constant m - (4^32)/n */ + .octa 0x000000000000000000000000dea713f1 /* x^64 div p(x)` */ + /* 33 bit reflected Barrett constant n */ + .octa 0x00000000000000000000000105ec76f1 +#endif + +#endif diff --git a/util/ppc-opcode.h b/util/ppc-opcode.h new file mode 100644 index 000000000..eeb0ae08f --- /dev/null +++ b/util/ppc-opcode.h @@ -0,0 +1,31 @@ +// Copyright (c) 2017 International Business Machines Corp. +// All rights reserved. +// This source code is licensed under the BSD-style license found in the +// LICENSE file in the root directory of this source tree. An additional grant +// of patent rights can be found in the PATENTS file in the same directory. +// This source code is also licensed under the GPLv2 license found in the +// COPYING file in the root directory of this source tree. + +#ifndef __OPCODES_H +#define __OPCODES_H + +#define __PPC_RA(a) (((a)&0x1f) << 16) +#define __PPC_RB(b) (((b)&0x1f) << 11) +#define __PPC_XA(a) ((((a)&0x1f) << 16) | (((a)&0x20) >> 3)) +#define __PPC_XB(b) ((((b)&0x1f) << 11) | (((b)&0x20) >> 4)) +#define __PPC_XS(s) ((((s)&0x1f) << 21) | (((s)&0x20) >> 5)) +#define __PPC_XT(s) __PPC_XS(s) +#define VSX_XX3(t, a, b) (__PPC_XT(t) | __PPC_XA(a) | __PPC_XB(b)) +#define VSX_XX1(s, a, b) (__PPC_XS(s) | __PPC_RA(a) | __PPC_RB(b)) + +#define PPC_INST_VPMSUMW 0x10000488 +#define PPC_INST_VPMSUMD 0x100004c8 +#define PPC_INST_MFVSRD 0x7c000066 +#define PPC_INST_MTVSRD 0x7c000166 + +#define VPMSUMW(t, a, b) .long PPC_INST_VPMSUMW | VSX_XX3((t), a, b) +#define VPMSUMD(t, a, b) .long PPC_INST_VPMSUMD | VSX_XX3((t), a, b) +#define MFVRD(a, t) .long PPC_INST_MFVSRD | VSX_XX1((t) + 32, a, 0) +#define MTVRD(t, a) .long PPC_INST_MTVSRD | VSX_XX1((t) + 32, a, 0) + +#endif