avx512 and sse4.1

README

@@ -0,0 +1,11 @@
+# quick compile yafu 2.0
+lanch compile.sh !
+
+warning: This yafu use avx512 and sse4.1 in default.
+
+
+# references
+https://www.mersenneforum.org/showthread.php?t=23087
+https://www.mersenneforum.org/showthread.php?t=26681
+https://github.com/bbuhrow/yafu
+

avx-ecm/LICENSE

@@ -0,0 +1,25 @@
+BSD 2-Clause License
+
+Copyright (c) 2019, bbuhrow
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions are met:
+
+1. Redistributions of source code must retain the above copyright notice, this
+   list of conditions and the following disclaimer.
+
+2. Redistributions in binary form must reproduce the above copyright notice,
+   this list of conditions and the following disclaimer in the documentation
+   and/or other materials provided with the distribution.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE
+FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
+CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

avx-ecm/Makefile

@@ -0,0 +1,193 @@
+# 
+# Copyright (c) 2019, Ben Buhrow
+# All rights reserved.
+# 
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions are met:
+# 
+# 1. Redistributions of source code must retain the above copyright notice, this
+#    list of conditions and the following disclaimer. 
+# 2. Redistributions in binary form must reproduce the above copyright notice,
+#    this list of conditions and the following disclaimer in the documentation
+#    and/or other materials provided with the distribution.
+# 
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+# ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+# WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+# ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+# (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+# LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+# ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+# 
+# The views and conclusions contained in the software and documentation are those
+# of the authors and should not be interpreted as representing official policies, 
+# either expressed or implied, of the FreeBSD Project.
+# 
+# 
+
+#--------------------------- flags -------------------------
+CC = gcc
+WARN_FLAGS = -Wall #-W -Wconversion
+OPT_FLAGS = -O3
+INC = -I. 
+LIBS =
+BINNAME = avx-ecm
+COMPILER = gcc
+
+
+#--------------------------- make options -------------------------
+
+ifeq ($(COMPILER),mingw)
+# NOTE: Using -fcall-used instead of -ffixed is much better and still works.
+# -fcall-used simply prevents the named registers from being saved/restored while
+# -ffixed prevents them from being used at all.  The code benefits a lot from being
+# able to use all 32 zmm registers.
+	CC = gcc
+    CFLAGS += -fcall-used-xmm16 -fcall-used-xmm17 -fcall-used-xmm18 -fcall-used-xmm19
+    CFLAGS += -fcall-used-xmm20 -fcall-used-xmm21 -fcall-used-xmm22 -fcall-used-xmm23
+    CFLAGS += -fcall-used-xmm24 -fcall-used-xmm25 -fcall-used-xmm26 -fcall-used-xmm27
+    CFLAGS += -fcall-used-xmm28 -fcall-used-xmm29 -fcall-used-xmm30 -fcall-used-xmm31
+    INC = -I. -I/y/projects/factoring/gmp/include/mingw
+    LIBS = -L/y/projects/factoring/gmp/lib/mingw/x86_64
+else ifeq ($(COMPILER),gcc)
+    CC = gcc
+    INC = -I. -I../gmp-6.2.1
+    LIBS = -L../gmp-6.2.1/.libs
+else ifeq ($(COMPILER),gcc730)
+    CC = gcc
+    INC = -I. -I/sppdg/scratch/buhrow/projects/gmp_install/include
+    LIBS = -L/sppdg/scratch/buhrow/projects/gmp_install/install/lib
+else
+    CC = icc
+    INC = -I. -I/sppdg/scratch/buhrow/projects/gmp_install/include
+    LIBS = -L/sppdg/scratch/buhrow/projects/gmp_install/lib
+endif
+
+ifdef MAXBITS
+	CFLAGS += -DMAXBITS=$(MAXBITS)
+endif
+
+ifdef DIGITBITS
+	CFLAGS += -DDIGITBITS=$(DIGITBITS)
+endif
+
+OBJ_EXT = .o
+OPT_FLAGS += -mavx
+
+ifeq ($(KNL),1)
+    ifeq ($(COMPILER),icc)
+        CFLAGS += -xMIC-AVX512 -DTARGET_KNL
+    else
+        CFLAGS += -march=knl -DTARGET_KNL
+    endif
+	BINNAME := ${BINNAME:%=%_knl}
+    OBJ_EXT = .ko
+endif
+
+ifeq ($(SKYLAKEX),1)
+	OPT_FLAGS += -march=skylake-avx512 -DSKYLAKEX
+	OBJ_EXT = .o
+endif
+
+
+ifeq ($(ICELAKE),1)
+	OPT_FLAGS += -march=icelake-client -DIFMA -DICELAKE
+	OBJ_EXT = .o
+endif
+
+ifeq ($(TIGERLAKE),1)
+	OPT_FLAGS += -march=tigerlake -DIFMA
+	OBJ_EXT = .o
+endif
+
+
+ifeq ($(NO_THREADS),1)
+    CFLAGS += -DNO_THREADS
+else
+    NO_THREADS = 0
+endif
+
+ifeq ($(CC),icc)
+	ifeq ($(KNL),1)
+		CFLAGS += -mkl 
+	else
+		CFLAGS += -L/usr/lib/gcc/x86_64-redhat-linux/4.4.4 -L/lib
+        ifeq ($(NO_THREADS),0)
+            CFLAGS += -mkl
+        endif
+	endif
+endif
+
+ifeq ($(PROFILE),1)
+	CFLAGS += -pg
+	BINNAME := ${BINNAME:%=%_prof}
+endif
+
+
+CFLAGS += -g -msse4.1 $(OPT_FLAGS) $(WARN_FLAGS) $(INC)
+
+ifeq ($(STATIC),1)
+	CFLAGS += -static-intel
+	LIBS += -L/usr/lib/x86_64-redhat-linux6E/lib64/ /sppdg/scratch/buhrow/projects/gmp_install/lib/libgmp.a
+else
+	LIBS += -lm -lgmp -lpthread
+endif
+
+#--------------------------- file lists -------------------------
+SRCS = \
+	eratosthenes/presieve.c \
+	eratosthenes/count.c \
+	eratosthenes/offsets.c \
+	eratosthenes/primes.c \
+	eratosthenes/roots.c \
+	eratosthenes/linesieve.c \
+	eratosthenes/soe.c \
+	eratosthenes/tiny.c \
+	eratosthenes/worker.c \
+	eratosthenes/soe_util.c \
+	eratosthenes/wrapper.c \
+	threadpool.c \
+	main.c \
+	ecm.c \
+	util.c \
+	vecarith.c \
+	vecarith52.c \
+	vec_common.c \
+	calc.c \
+    queue.c
+
+
+OBJS = $(SRCS:.c=$(OBJ_EXT))
+
+
+
+#---------------------------Header file lists -------------------------
+HEAD = \
+	avx_ecm.h \
+	eratosthenes/soe.h \
+	threadpool.h \
+	util.h \
+	calc.h \
+	queue.h
+
+#---------------------------Make Targets -------------------------
+
+all: $(OBJS)
+	rm -f libavxecm.a
+	ar r libavxecm.a $(OBJS)
+	ranlib libavxecm.a
+	$(CC) $(CFLAGS) $(OBJS) -o $(BINNAME) libavxecm.a $(LIBS)
+
+
+clean:
+	rm -f $(OBJS)
+
+#---------------------------Build Rules -------------------------
+
+
+%$(OBJ_EXT): %.c $(HEAD)
+	$(CC) $(CFLAGS) -c -o $@ $<
+

avx-ecm/README.md

@@ -0,0 +1,37 @@
+# avx-ecm
+Computes parallel curves of the ECM (Elliptic Curve Method) factoring algorithm using AVX-512 vector arithmetic.  On CPUs where AVX512 is supported, AVX-ECM has been observed to have about 1.5 to 2.5 times the throughput (curves/sec) of GMP-ECM, for numbers up to a thousand bits or so.  Large numbers will be more efficient with GMP-ECM because AVX-ECM so far does not use sub-quadratic multiplier algorithms.
+
+Both stage 1 and stage 2 are computed in parallel, but stage 2 is just the standard continuation with pairing (by default, B2=100xB1).
+
+The program is also multi-threaded using the pthreads library.
+
+It will output a savefile after stage 1 that will work with GMP-ECM stage 2, if desired.
+e.g.:
+ecm -resume save_b1.txt 1000000 < input.txt
+
+Compile on linux using either icc or gcc-7.3.0 or above (I've only tested icc and gcc-7.3.0).
+
+e.g.:
+make COMPILER=gcc730 SKYLAKEX=1
+
+COMPILER options currently are gcc730 (which invokes gcc-7.3.0), gcc (which invokes gcc), mingw (using MSYS2 and mingw64 on windows) and icc (icc).  Edit the makefile if you have some other compiler name (ymmv).
+
+Use SKYLAKEX=1 to build for skylakex CPUs.  Alternatively use KNL=1 to build for Knight's Landing Xeon Phi systems.
+
+The program will run constant-time curves in steps of 208 bits (meaning, e.g., that 417 bit curves take the same amount of time as 623-bit curves) and 8 curves are performed in parallel per thread.
+
+You can optionally specify DIGITBITS=32 during make.  Doing so results in constant-time curves in steps of 128 bits and 16 curves performed in parallel per thread.
+
+The 52-bit version is generally faster (and the default) but for some sizes a 32-bit version can have higher throughput (because of the 208-bit jumps between sizes).
+
+Command line:
+avx-ecm input curves B1 threads B2
+
+The input number can be specified using these operators if desired: +,-,*,/,^,%,# (primorial), ! (factorial), fib(), and luc()
+
+Example:
+./avx-ecm "fib(791)/13/677/216416017" 8 1000000 1
+
+See https://www.mersenneforum.org/showthread.php?t=25056 for more info.
+
+Happy factoring!

avx-ecm/avx-ecm-vc/avx-ecm-vc.sln

@@ -0,0 +1,31 @@
+
+Microsoft Visual Studio Solution File, Format Version 12.00
+# Visual Studio 15
+VisualStudioVersion = 15.0.28307.271
+MinimumVisualStudioVersion = 10.0.40219.1
+Project("{8BC9CEB8-8B4A-11D0-8D11-00A0C91BC942}") = "avx-ecm-vc", "avx-ecm-vc.vcxproj", "{47524AD9-2D20-4AD2-B7FE-521EBE485A42}"
+EndProject
+Global
+	GlobalSection(SolutionConfigurationPlatforms) = preSolution
+		Debug|x64 = Debug|x64
+		Debug|x86 = Debug|x86
+		Release|x64 = Release|x64
+		Release|x86 = Release|x86
+	EndGlobalSection
+	GlobalSection(ProjectConfigurationPlatforms) = postSolution
+		{47524AD9-2D20-4AD2-B7FE-521EBE485A42}.Debug|x64.ActiveCfg = Debug|x64
+		{47524AD9-2D20-4AD2-B7FE-521EBE485A42}.Debug|x64.Build.0 = Debug|x64
+		{47524AD9-2D20-4AD2-B7FE-521EBE485A42}.Debug|x86.ActiveCfg = Debug|Win32
+		{47524AD9-2D20-4AD2-B7FE-521EBE485A42}.Debug|x86.Build.0 = Debug|Win32
+		{47524AD9-2D20-4AD2-B7FE-521EBE485A42}.Release|x64.ActiveCfg = Release|x64
+		{47524AD9-2D20-4AD2-B7FE-521EBE485A42}.Release|x64.Build.0 = Release|x64
+		{47524AD9-2D20-4AD2-B7FE-521EBE485A42}.Release|x86.ActiveCfg = Release|Win32
+		{47524AD9-2D20-4AD2-B7FE-521EBE485A42}.Release|x86.Build.0 = Release|Win32
+	EndGlobalSection
+	GlobalSection(SolutionProperties) = preSolution
+		HideSolutionNode = FALSE
+	EndGlobalSection
+	GlobalSection(ExtensibilityGlobals) = postSolution
+		SolutionGuid = {D937E06E-DE8E-4AFA-B0A3-6D2B3D4DF0D3}
+	EndGlobalSection
+EndGlobal