BINDIR = $(PREFIX)/bin
### Built-in benchmark for pgo-builds
-ifeq ($(SDE_PATH),)
- PGOBENCH = $(WINE_PATH) ./$(EXE) bench
-else
- PGOBENCH = $(SDE_PATH) -- $(WINE_PATH) ./$(EXE) bench
-endif
+PGOBENCH = $(WINE_PATH) ./$(EXE) bench
### Source and object files
-SRCS = benchmark.cpp bitbase.cpp bitboard.cpp endgame.cpp evaluate.cpp main.cpp \
- material.cpp misc.cpp movegen.cpp movepick.cpp pawns.cpp position.cpp psqt.cpp \
+SRCS = benchmark.cpp bitboard.cpp evaluate.cpp main.cpp \
+ misc.cpp movegen.cpp movepick.cpp position.cpp \
search.cpp thread.cpp timeman.cpp tt.cpp uci.cpp ucioption.cpp tune.cpp syzygy/tbprobe.cpp \
nnue/evaluate_nnue.cpp nnue/features/half_ka_v2_hm.cpp
### Section 2. High-level Configuration
### ==========================================================================
#
-# flag --- Comp switch --- Description
+# flag --- Comp switch --- Description
# ----------------------------------------------------------------------------
#
-# debug = yes/no --- -DNDEBUG --- Enable/Disable debug mode
+# debug = yes/no --- -DNDEBUG --- Enable/Disable debug mode
# sanitize = none/<sanitizer> ... (-fsanitize )
-# --- ( undefined ) --- enable undefined behavior checks
-# --- ( thread ) --- enable threading error checks
-# --- ( address ) --- enable memory access checks
-# --- ...etc... --- see compiler documentation for supported sanitizers
-# optimize = yes/no --- (-O3/-fast etc.) --- Enable/Disable optimizations
-# arch = (name) --- (-arch) --- Target architecture
-# bits = 64/32 --- -DIS_64BIT --- 64-/32-bit operating system
-# prefetch = yes/no --- -DUSE_PREFETCH --- Use prefetch asm-instruction
-# popcnt = yes/no --- -DUSE_POPCNT --- Use popcnt asm-instruction
-# pext = yes/no --- -DUSE_PEXT --- Use pext x86_64 asm-instruction
-# sse = yes/no --- -msse --- Use Intel Streaming SIMD Extensions
-# mmx = yes/no --- -mmmx --- Use Intel MMX instructions
-# sse2 = yes/no --- -msse2 --- Use Intel Streaming SIMD Extensions 2
-# ssse3 = yes/no --- -mssse3 --- Use Intel Supplemental Streaming SIMD Extensions 3
-# sse41 = yes/no --- -msse4.1 --- Use Intel Streaming SIMD Extensions 4.1
-# avx2 = yes/no --- -mavx2 --- Use Intel Advanced Vector Extensions 2
-# avxvnni = yes/no --- -mavxvnni --- Use Intel Vector Neural Network Instructions AVX
-# avx512 = yes/no --- -mavx512bw --- Use Intel Advanced Vector Extensions 512
-# vnni256 = yes/no --- -mavx512vnni --- Use Intel Vector Neural Network Instructions 256
-# vnni512 = yes/no --- -mavx512vnni --- Use Intel Vector Neural Network Instructions 512
-# neon = yes/no --- -DUSE_NEON --- Use ARM SIMD architecture
+# --- ( undefined ) --- enable undefined behavior checks
+# --- ( thread ) --- enable threading error checks
+# --- ( address ) --- enable memory access checks
+# --- ...etc... --- see compiler documentation for supported sanitizers
+# optimize = yes/no --- (-O3/-fast etc.) --- Enable/Disable optimizations
+# arch = (name) --- (-arch) --- Target architecture
+# bits = 64/32 --- -DIS_64BIT --- 64-/32-bit operating system
+# prefetch = yes/no --- -DUSE_PREFETCH --- Use prefetch asm-instruction
+# popcnt = yes/no --- -DUSE_POPCNT --- Use popcnt asm-instruction
+# pext = yes/no --- -DUSE_PEXT --- Use pext x86_64 asm-instruction
+# sse = yes/no --- -msse --- Use Intel Streaming SIMD Extensions
+# mmx = yes/no --- -mmmx --- Use Intel MMX instructions
+# sse2 = yes/no --- -msse2 --- Use Intel Streaming SIMD Extensions 2
+# ssse3 = yes/no --- -mssse3 --- Use Intel Supplemental Streaming SIMD Extensions 3
+# sse41 = yes/no --- -msse4.1 --- Use Intel Streaming SIMD Extensions 4.1
+# avx2 = yes/no --- -mavx2 --- Use Intel Advanced Vector Extensions 2
+# avxvnni = yes/no --- -mavxvnni --- Use Intel Vector Neural Network Instructions AVX
+# avx512 = yes/no --- -mavx512bw --- Use Intel Advanced Vector Extensions 512
+# vnni256 = yes/no --- -mavx256vnni --- Use Intel Vector Neural Network Instructions 512 with 256bit operands
+# vnni512 = yes/no --- -mavx512vnni --- Use Intel Vector Neural Network Instructions 512
+# neon = yes/no --- -DUSE_NEON --- Use ARM SIMD architecture
+# dotprod = yes/no --- -DUSE_NEON_DOTPROD --- Use ARM advanced SIMD Int8 dot product instructions
#
# Note that Makefile is space sensitive, so when adding new architectures
# or modifying existing flags, you have to make sure there are no extra spaces
### 2.1. General and architecture defaults
ifeq ($(ARCH),)
- ARCH = x86-64-modern
+ ARCH = x86-64-avx2
help_skip_sanity = yes
endif
# explicitly check for the list of supported architectures (as listed with make help),
x86-64-vnni512 x86-64-vnni256 x86-64-avx512 x86-64-avxvnni x86-64-bmi2 \
x86-64-avx2 x86-64-sse41-popcnt x86-64-modern x86-64-ssse3 x86-64-sse3-popcnt \
x86-64 x86-32-sse41-popcnt x86-32-sse2 x86-32 ppc-64 ppc-32 e2k \
- armv7 armv7-neon armv8 apple-silicon general-64 general-32 riscv64))
+ armv7 armv7-neon armv8 armv8-dotprod apple-silicon general-64 general-32 riscv64))
SUPPORTED_ARCH=true
else
SUPPORTED_ARCH=false
vnni256 = no
vnni512 = no
neon = no
+dotprod = no
arm_version = 0
STRIP = strip
endif
ifeq ($(findstring -modern,$(ARCH)),-modern)
+ $(warning *** ARCH=$(ARCH) is deprecated, defaulting to ARCH=x86-64-sse41-popcnt. Execute `make help` for a list of available architectures. ***)
+ $(shell sleep 5)
popcnt = yes
sse = yes
sse2 = yes
arm_version = 8
endif
+ifeq ($(ARCH),armv8-dotprod)
+ arch = armv8
+ prefetch = yes
+ popcnt = yes
+ neon = yes
+ dotprod = yes
+ arm_version = 8
+endif
+
ifeq ($(ARCH),apple-silicon)
arch = arm64
prefetch = yes
popcnt = yes
neon = yes
+ dotprod = yes
arm_version = 8
endif
CXXFLAGS += -pedantic -Wextra -Wshadow -Wmissing-declarations
endif
-ifeq ($(COMP),icc)
- comp=icc
- CXX=icpc
- CXXFLAGS += -diag-disable 1476,10120 -Wcheck -Wabi -Wdeprecated -strict-ansi
+ifeq ($(COMP),icx)
+ comp=icx
+ CXX=icpx
+ CXXFLAGS += --intel -pedantic -Wextra -Wshadow -Wmissing-prototypes \
+ -Wconditional-uninitialized -Wabi -Wdeprecated
endif
ifeq ($(COMP),clang)
LDFLAGS += -static-libstdc++ -pie -lm -latomic
endif
-ifeq ($(comp),icc)
- profile_make = icc-profile-make
- profile_use = icc-profile-use
+ifeq ($(comp),icx)
+ profile_make = icx-profile-make
+ profile_use = icx-profile-use
else ifeq ($(comp),clang)
profile_make = clang-profile-make
profile_use = clang-profile-use
endif
ifeq ($(KERNEL),Darwin)
- ifeq ($(comp),$(filter $(comp),clang icc))
+ ifeq ($(comp),$(filter $(comp),clang icx))
CXXFLAGS += -mdynamic-no-pic
endif
endif
ifeq ($(comp),clang)
- CXXFLAGS += -fexperimental-new-pass-manager
+ clangmajorversion = $(shell $(CXX) -dumpversion 2>/dev/null | cut -f1 -d.)
+ ifeq ($(shell expr $(clangmajorversion) \< 16),1)
+ CXXFLAGS += -fexperimental-new-pass-manager
+ endif
endif
endif
ifeq ($(popcnt),yes)
ifeq ($(arch),$(filter $(arch),ppc64 armv7 armv8 arm64))
CXXFLAGS += -DUSE_POPCNT
- else ifeq ($(comp),icc)
- CXXFLAGS += -msse3 -DUSE_POPCNT
else
CXXFLAGS += -msse3 -mpopcnt -DUSE_POPCNT
endif
### 3.6 SIMD architectures
ifeq ($(avx2),yes)
CXXFLAGS += -DUSE_AVX2
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -mavx2 -mbmi
endif
endif
ifeq ($(avxvnni),yes)
CXXFLAGS += -DUSE_VNNI -DUSE_AVXVNNI
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -mavxvnni
endif
endif
ifeq ($(avx512),yes)
CXXFLAGS += -DUSE_AVX512
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -mavx512f -mavx512bw
endif
endif
ifeq ($(vnni256),yes)
CXXFLAGS += -DUSE_VNNI
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -mprefer-vector-width=256
endif
endif
ifeq ($(vnni512),yes)
CXXFLAGS += -DUSE_VNNI
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
- CXXFLAGS += -mavx512vnni -mavx512dq -mavx512vl
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
+ CXXFLAGS += -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -mprefer-vector-width=512
endif
endif
ifeq ($(sse41),yes)
CXXFLAGS += -DUSE_SSE41
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -msse4.1
endif
endif
ifeq ($(ssse3),yes)
CXXFLAGS += -DUSE_SSSE3
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -mssse3
endif
endif
ifeq ($(sse2),yes)
CXXFLAGS += -DUSE_SSE2
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -msse2
endif
endif
ifeq ($(mmx),yes)
CXXFLAGS += -DUSE_MMX
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -mmmx
endif
endif
endif
endif
+ifeq ($(dotprod),yes)
+ CXXFLAGS += -march=armv8.2-a+dotprod -DUSE_NEON_DOTPROD
+endif
+
### 3.7 pext
ifeq ($(pext),yes)
CXXFLAGS += -DUSE_PEXT
- ifeq ($(comp),$(filter $(comp),gcc clang mingw))
+ ifeq ($(comp),$(filter $(comp),gcc clang mingw icx))
CXXFLAGS += -mbmi2
endif
endif
### 3.7.1 Try to include git commit sha for versioning
-GIT_SHA = $(shell git rev-parse --short HEAD 2>/dev/null)
+GIT_SHA = $(shell git rev-parse HEAD 2>/dev/null | cut -c 1-8)
ifneq ($(GIT_SHA), )
- CXXFLAGS += -DGIT_SHA=\"$(GIT_SHA)\"
+ CXXFLAGS += -DGIT_SHA=$(GIT_SHA)
endif
### 3.7.2 Try to include git commit date for versioning
GIT_DATE = $(shell git show -s --date=format:'%Y%m%d' --format=%cd HEAD 2>/dev/null)
ifneq ($(GIT_DATE), )
- CXXFLAGS += -DGIT_DATE=\"$(GIT_DATE)\"
+ CXXFLAGS += -DGIT_DATE=$(GIT_DATE)
endif
### 3.8 Link Time Optimization
### needs access to the optimization flags.
ifeq ($(optimize),yes)
ifeq ($(debug), no)
- ifeq ($(comp),clang)
+ ifeq ($(comp),$(filter $(comp),clang icx))
CXXFLAGS += -flto=full
+ ifeq ($(comp),icx)
+ CXXFLAGS += -fwhole-program-vtables
+ endif
ifeq ($(target_windows),yes)
CXXFLAGS += -fuse-ld=lld
endif
@echo ""
@echo "Supported archs:"
@echo ""
- @echo "x86-64-vnni512 > x86 64-bit with vnni support 512bit wide"
- @echo "x86-64-vnni256 > x86 64-bit with vnni support 256bit wide"
+ @echo "x86-64-vnni512 > x86 64-bit with vnni 512bit support"
+ @echo "x86-64-vnni256 > x86 64-bit with vnni 512bit support, limit operands to 256bit wide"
@echo "x86-64-avx512 > x86 64-bit with avx512 support"
- @echo "x86-64-avxvnni > x86 64-bit with avxvnni support"
+ @echo "x86-64-avxvnni > x86 64-bit with vnni 256bit support"
@echo "x86-64-bmi2 > x86 64-bit with bmi2 support"
@echo "x86-64-avx2 > x86 64-bit with avx2 support"
@echo "x86-64-sse41-popcnt > x86 64-bit with sse41 and popcnt support"
- @echo "x86-64-modern > common modern CPU, currently x86-64-sse41-popcnt"
+ @echo "x86-64-modern > deprecated, currently x86-64-sse41-popcnt"
@echo "x86-64-ssse3 > x86 64-bit with ssse3 support"
@echo "x86-64-sse3-popcnt > x86 64-bit with sse3 and popcnt support"
@echo "x86-64 > x86 64-bit generic (with sse2 support)"
@echo "armv7 > ARMv7 32-bit"
@echo "armv7-neon > ARMv7 32-bit with popcnt and neon"
@echo "armv8 > ARMv8 64-bit with popcnt and neon"
+ @echo "armv8-dotprod > ARMv8 64-bit with popcnt, neon and dot product support"
@echo "e2k > Elbrus 2000"
@echo "apple-silicon > Apple silicon ARM64"
@echo "general-64 > unspecified 64-bit"
@echo "gcc > Gnu compiler (default)"
@echo "mingw > Gnu compiler with MinGW under Windows"
@echo "clang > LLVM Clang compiler"
- @echo "icc > Intel compiler"
+ @echo "icx > Intel oneAPI DPC++/C++ Compiler"
@echo "ndk > Google NDK to cross-compile for Android"
@echo ""
@echo "Simple examples. If you don't know what to do, you likely want to run one of: "
@echo ""
@echo "make -j profile-build ARCH=x86-64-avx2 # typically a fast compile for common systems "
- @echo "make -j profile-build ARCH=x86-64-modern # A more portable compile for 64-bit systems "
+ @echo "make -j profile-build ARCH=x86-64-sse41-popcnt # A more portable compile for 64-bit systems "
@echo "make -j profile-build ARCH=x86-64 # A portable compile for 64-bit systems "
@echo ""
@echo "Advanced examples, for experienced users: "
@echo ""
- @echo "make -j profile-build ARCH=x86-64-bmi2"
- @echo "make -j profile-build ARCH=x86-64-bmi2 COMP=gcc COMPCXX=g++-9.0"
+ @echo "make -j profile-build ARCH=x86-64-avxvnni"
+ @echo "make -j profile-build ARCH=x86-64-avxvnni COMP=gcc COMPCXX=g++-12.0"
@echo "make -j build ARCH=x86-64-ssse3 COMP=clang"
@echo ""
@echo "-------------------------------"
.PHONY: help build profile-build strip install clean net objclean profileclean \
- config-sanity icc-profile-use icc-profile-make gcc-profile-use gcc-profile-make \
- clang-profile-use clang-profile-make FORCE
+ config-sanity \
+ icx-profile-use icx-profile-make \
+ gcc-profile-use gcc-profile-make \
+ clang-profile-use clang-profile-make FORCE
build: net config-sanity
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) all
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) $(profile_make)
@echo ""
@echo "Step 2/4. Running benchmark for pgo-build ..."
- $(PGOBENCH) 2>&1 | tail -n 4
+ $(PGOBENCH) > PGOBENCH.out 2>&1
+ tail -n 4 PGOBENCH.out
@echo ""
@echo "Step 3/4. Building optimized executable ..."
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) objclean
# clean auxiliary profiling files
profileclean:
@rm -rf profdir
- @rm -f bench.txt *.gcda *.gcno ./syzygy/*.gcda ./nnue/*.gcda ./nnue/features/*.gcda *.s
+ @rm -f bench.txt *.gcda *.gcno ./syzygy/*.gcda ./nnue/*.gcda ./nnue/features/*.gcda *.s PGOBENCH.out
@rm -f stockfish.profdata *.profraw
@rm -f stockfish.*args*
@rm -f stockfish.*lt*
@echo "vnni256: '$(vnni256)'"
@echo "vnni512: '$(vnni512)'"
@echo "neon: '$(neon)'"
+ @echo "dotprod: '$(dotprod)'"
@echo "arm_version: '$(arm_version)'"
+ @echo "target_windows: '$(target_windows)'"
@echo ""
@echo "Flags:"
@echo "CXX: $(CXX)"
@test "$(vnni256)" = "yes" || test "$(vnni256)" = "no"
@test "$(vnni512)" = "yes" || test "$(vnni512)" = "no"
@test "$(neon)" = "yes" || test "$(neon)" = "no"
- @test "$(comp)" = "gcc" || test "$(comp)" = "icc" || test "$(comp)" = "mingw" || test "$(comp)" = "clang" \
+ @test "$(comp)" = "gcc" || test "$(comp)" = "icx" || test "$(comp)" = "mingw" || test "$(comp)" = "clang" \
|| test "$(comp)" = "armv7a-linux-androideabi16-clang" || test "$(comp)" = "aarch64-linux-android21-clang"
$(EXE): $(OBJS)
EXTRALDFLAGS='-lgcov' \
all
-icc-profile-make:
- @mkdir -p profdir
+icx-profile-make:
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
- EXTRACXXFLAGS='-prof-gen=srcpos -prof_dir ./profdir' \
+ EXTRACXXFLAGS='-fprofile-instr-generate ' \
+ EXTRALDFLAGS=' -fprofile-instr-generate' \
all
-icc-profile-use:
+icx-profile-use:
+ $(XCRUN) llvm-profdata merge -output=stockfish.profdata *.profraw
$(MAKE) ARCH=$(ARCH) COMP=$(COMP) \
- EXTRACXXFLAGS='-prof_use -prof_dir ./profdir' \
+ EXTRACXXFLAGS='-fprofile-instr-use=stockfish.profdata' \
+ EXTRALDFLAGS='-fprofile-use ' \
all
.depend: $(SRCS)
projects / stockfish / blobdiff