Merge pull request #132 from DLTcollab/develop

Pre-release v0.2.0

LICENSE

@@ -1,4 +1,4 @@
-Copyright (C) 2018 dcurl Developers.
+Copyright (C) 2018-2019 BiiLabs Co., Ltd. and Contributors.
 Copyright (C) 2017 IOTA AS, IOTA Foundation and Developers.
 Copyright (C) 2016 Shinya Yagyu.
 

Makefile

@@ -58,6 +58,10 @@ else ifeq ("$(BUILD_GENERIC)","1")
 CFLAGS += -DENABLE_GENERIC
 endif
 
+ifeq ("$(call cpu_feature,SSE4_2)","1")
+    CFLAGS += -msse4.2
+endif
+
 ifeq ("$(BUILD_GPU)","1")
 include mk/opencl.mk
 endif
@@ -78,13 +82,8 @@ TESTS = \
 	trinary \
 	curl \
 	dcurl \
-	multi-pow \
 	pow
 
-ifeq ("$(BUILD_COMPAT)", "1")
-TESTS += ccurl-multi_pow
-endif
-
 TESTS := $(addprefix $(OUT)/test-, $(TESTS))
 
 LIBS = libdcurl.so
@@ -132,14 +131,6 @@ endif
 
 OBJS := $(addprefix $(OUT)/, $(OBJS))
 
-# Add the libtuv PIC(position independent code) library into the object files
-# if the specified hardware is CPU
-CPU_PLATFORMS := $(BUILD_AVX) $(BUILD_SSE) $(BUILD_GENERIC)
-ENABLE_CPU_PLATFORMS := $(findstring 1,$(CPU_PLATFORMS))
-ifeq ("$(ENABLE_CPU_PLATFORMS)","1")
-    OBJS += $(LIBTUV_LIBRARY)
-endif
-
 $(OUT)/test-%.o: tests/test-%.c $(LIBTUV_PATH)/include
 	$(VECHO) "  CC\t$@\n"
 	$(Q)$(CC) -o $@ $(CFLAGS) -I $(SRC) $(LIBTUV_INCLUDE) -c -MMD -MF $@.d $<
@@ -148,20 +139,14 @@ $(OUT)/%.o: $(SRC)/%.c $(LIBTUV_PATH)/include
 	$(VECHO) "  CC\t$@\n"
 	$(Q)$(CC) -o $@ $(CFLAGS) $(LIBTUV_INCLUDE) -c -MMD -MF $@.d $<
 
-$(OUT)/test-%: $(OUT)/test-%.o $(OBJS)
+$(OUT)/test-%: $(OUT)/test-%.o $(OBJS) $(LIBTUV_LIBRARY)
 	$(VECHO) "  LD\t$@\n"
 	$(Q)$(CC) -o $@ $^ $(LDFLAGS)
 
-$(OUT)/libdcurl.so: $(OBJS)
+$(OUT)/libdcurl.so: $(OBJS) $(LIBTUV_LIBRARY)
 	$(VECHO) "  LD\t$@\n"
 	$(Q)$(CC) -shared -o $@ $^ $(LDFLAGS)
 
-$(OUT)/test-%: tests/test-%.py $(OUT)/libdcurl.so
-	$(Q)echo "#!$(PYTHON)" > $@
-	$(call py_prepare_cmd)
-	$(Q)chmod +x $@
-
 include mk/common.mk
-include mk/python.mk
 
 -include $(deps)

README.md

@@ -1,6 +1,6 @@
 # dcurl - Multi-threaded Curl implementation
 
-[![Build Status](https://travis-ci.org/DLTcollab/dcurl.svg?branch=dev)](https://travis-ci.org/DLTcollab/dcurl)
+[![Build Status](https://travis-ci.org/DLTcollab/dcurl.svg?branch=develop)](https://travis-ci.org/DLTcollab/dcurl)
 ![Supported IRI version](https://img.shields.io/badge/Supported%20IRI%20Version-1.6.0-brightgreen.svg)
 ![Release version](https://img.shields.io/github/release-pre/DLTcollab/dcurl.svg)
 
@@ -23,7 +23,7 @@ After integrating dcurl into IRI, performance of [attachToTangle](https://iota.r
 * Each sampling is measured with 30 transaction trytes and total 200 samples are measured.
 * mwm = 14, 26 CPU threads to find nonce
 * Settings: enable 2 pow tasks in CPU, 1 pow tasks in GPU at the same time
-![](https://github.com/raw/DLTcollab/dcurl/dev/docs/benchmark.png)
+![](https://github.com/raw/DLTcollab/dcurl/develop/docs/benchmark.png)
 
 
 ## IRI Adaptation

docs/build-n-test.md

@@ -23,6 +23,8 @@
     - ``BUILD_STAT``: show the statistics of the PoW information.
     - ``BUILD_DEBUG``: dump verbose messages internally.
                        Build with the corresponding Sanitizer to detect software potential bugs if the value is `address`, `undefined` or `thread`.
+    - ``BOARD``: specify the board which builds the source code.
+                 The supported boards are `de10nano` (DE10-nano board), `arrowsockit` (Arrow SoCKit board), `rpi2` and `rpi3` (Raspberry Pi 2/3 board).
 * Alternatively, you can specify conditional build as following:
 ```shell
 $ make BUILD_GPU=0 BUILD_JNI=1 BUILD_AVX=1

mk/submodule.mk

@@ -8,16 +8,21 @@ ifeq ($(UNAME_S),Darwin)
     # macOS
     UNAME_S := darwin
 endif
+# For de10-nano, arrow sockit and raspberry pi 2/3 board
+ifeq ($(UNAME_M),armv7l)
+    UNAME_M := arm
+endif
 
 # libtuv related variables
 LIBTUV_PATH = deps/libtuv
 LIBTUV_INCLUDE := -I $(LIBTUV_PATH)/include
 LIBTUV_PLATFORM := $(UNAME_M)-$(UNAME_S)
+LIBTUV_BOARD := $(BUILD_BOARD)
 # PIC (Position-Independent-Code) library
 LIBTUV_LIBRARY := $(LIBTUV_PATH)/build/$(LIBTUV_PLATFORM)/release/lib/libtuv.o
 
 $(LIBTUV_PATH)/include:
 	git submodule update --init $(LIBTUV_PATH)
 
 $(LIBTUV_LIBRARY):
-	$(MAKE) -C $(LIBTUV_PATH) TUV_BUILD_TYPE=release TUV_CREATE_PIC_LIB=yes
+	$(MAKE) -C $(LIBTUV_PATH) TUV_BUILD_TYPE=release TUV_CREATE_PIC_LIB=yes TUV_PLATFORM=$(LIBTUV_PLATFORM) TUV_BOARD=$(LIBTUV_BOARD)

src/curl.c

@@ -20,7 +20,7 @@ static void _transform(int8_t state[])
         for (i = 0; i < STATE_TRITS_LENGTH; i++) {
             int aa = indices[i];
             int bb = indices[i + 1];
-            to[i] = truthTable[from[aa] + (from[bb] << 2) + 5];
+            to[i] = truthTable[from[aa] + (from[bb] * 4) + 5];
         }
         int8_t *tmp = from;
         from = to;

src/dcurl.c

@@ -17,6 +17,7 @@
 #endif
 #include "implcontext.h"
 #include "trinary.h"
+#include "uv.h"
 #if defined(ENABLE_AVX)
 #include "pow_avx.h"
 #elif defined(ENABLE_SSE)
@@ -25,20 +26,9 @@
 #include "pow_c.h"
 #endif
 
-#ifdef __APPLE__
-#include <dispatch/dispatch.h>
-#else
-#include <semaphore.h>
-#endif
-
 /* check whether dcurl is initialized */
 static bool isInitialized = false;
-
-#ifdef __APPLE__
-static dispatch_semaphore_t notify;
-#else
-static sem_t notify;
-#endif
+static uv_sem_t notify;
 
 LIST_HEAD(IMPL_LIST);
 
@@ -78,11 +68,7 @@ bool dcurl_init()
     ret &= registerImplContext(&PoWFPGAAccel_Context);
 #endif
 
-#ifdef __APPLE__
-    notify = dispatch_semaphore_create(0);
-#else
-    sem_init(&notify, 0, 0);
-#endif
+    uv_sem_init(&notify, 0);
     return isInitialized = ret;
 }
 
@@ -118,11 +104,7 @@ int8_t *dcurl_entry(int8_t *trytes, int mwm, int threads)
                 goto do_pow;
             }
         }
-#ifdef __APPLE__
-        dispatch_semaphore_wait(notify, DISPATCH_TIME_FOREVER);
-#else
-        sem_wait(&notify);
-#endif
+        uv_sem_wait(&notify);
     } while (1);
 
 do_pow:
@@ -133,10 +115,6 @@ int8_t *dcurl_entry(int8_t *trytes, int mwm, int threads)
     }
     freePoWContext(impl, pow_ctx);
     exitImplContext(impl);
-#ifdef __APPLE__
-    dispatch_semaphore_signal(notify);
-#else
-    sem_post(&notify);
-#endif
+    uv_sem_post(&notify);
     return res;
 }

src/implcontext.c

@@ -30,21 +30,21 @@ void destroyImplContext(ImplContext *impl_ctx)
 
 bool enterImplContext(ImplContext *impl_ctx)
 {
-    pthread_mutex_lock(&impl_ctx->lock);
+    uv_mutex_lock(&impl_ctx->lock);
     if (impl_ctx->num_working_thread >= impl_ctx->num_max_thread) {
-        pthread_mutex_unlock(&impl_ctx->lock);
+        uv_mutex_unlock(&impl_ctx->lock);
         return false; /* Access Failed */
     }
     impl_ctx->num_working_thread++;
-    pthread_mutex_unlock(&impl_ctx->lock);
+    uv_mutex_unlock(&impl_ctx->lock);
     return true; /* Access Success */
 }
 
 void exitImplContext(ImplContext *impl_ctx)
 {
-    pthread_mutex_lock(&impl_ctx->lock);
+    uv_mutex_lock(&impl_ctx->lock);
     impl_ctx->num_working_thread--;
-    pthread_mutex_unlock(&impl_ctx->lock);
+    uv_mutex_unlock(&impl_ctx->lock);
 }
 
 void *getPoWContext(ImplContext *impl_ctx, int8_t *trytes, int mwm, int threads)

src/implcontext.h

@@ -1,11 +1,11 @@
 #ifndef IMPL_CTX_H_
 #define IMPL_CTX_H_
 
-#include <pthread.h>
 #include <stdbool.h>
 #include <stdint.h>
 #include "common.h"
 #include "list.h"
+#include "uv.h"
 
 typedef struct _impl_context ImplContext;
 
@@ -14,7 +14,7 @@ struct _impl_context {
     char *description;
 
     /* Multi-thread Management */
-    pthread_mutex_t lock;
+    uv_mutex_t lock;
     int bitmap; /* Used to tell which slot is available */
     int num_max_thread;
     int num_working_thread;