chapter 11 Makefile

Author: nvixnu

chapter_10/ch10__spmv.cu

@@ -227,11 +227,11 @@ int main(){
 	printf("Running on Device with 1024 threads per block...");
 	ch10__spmv(Device, {.block_dim = {1024, 1, 1}, .kernel_version = CH10__SPMV_CSR});
 
-	printf("\n_____ spmv [ELL] _____\n\n");
+	printf("\n_____ spmv [ELL] _____\n");
 	printf("Running on Device with 1024 threads per block...");
 	ch10__spmv(Device, {.block_dim = {1024, 1, 1}, .kernel_version = CH10__SPMV_ELL});
 
-	printf("\n_____ spmv_CPU [CSR] _____\n\n");
+	printf("\n_____ spmv_CPU [CSR] _____\n");
 	ch10__spmv(Host, {});
 
 	return 0;

chapter_11/Makefile

@@ -0,0 +1,44 @@
+CC = nvcc
+OBJDIR = ./obj
+
+DEPSNAMES = \
+nvixnu__array_utils \
+nvixnu__populate_arrays_utils \
+nvixnu__error_utils \
+nvixnu__cuda_devices_props
+
+CFLAGS = -g -G --compiler-options -Wall -lm
+
+INCLUDES = $(patsubst %,-I ../../%, $(DEPSNAMES))
+
+# List with all .cu files inside $(REPOSDIR)/<repoName>
+CUFILES = $(foreach dep,$(DEPSNAMES), $(wildcard ../../$(dep)/*.cu))
+
+# List with all .o paths
+OBJS = $(patsubst %.cu,%.o,$(CUFILES))
+
+# Compiled objects path
+COMPILEDOBJS := $(patsubst %,$(OBJDIR)/%,$(notdir $(OBJS))) 
+
+# Creates the obj dir, compiles each dependency and then the main app
+all: objdir $(OBJS)
+	nvcc ch11__merge_sort.cu -o merge_sort.out $(COMPILEDOBJS) $(CFLAGS) $(INCLUDES)
+
+# Creates the ./obj dir
+objdir:
+	mkdir	-p	$(OBJDIR)
+
+# Compile a dependency
+%.o: %.cu
+	nvcc -c $< -o $(OBJDIR)/$(notdir $@) $(CFLAGS)
+
+# Run the executable
+run:
+	./main
+
+# Remove the generated artifacts
+clean:
+	rm -Rf $(OBJDIR)/*.o
+	rm -Rf merge_sort.out
+
+.PHONY: all clean app run objdir

chapter_11/ch11__merge_sort.cu

@@ -382,3 +382,30 @@ void ch11__merge_sort(env_e env, kernel_config_t config){
 
 	return;
 }
+
+int main(){
+	printf("Chapter 11\n");
+	printf("Array A length: %d\n", CH11__A_LENGTH);
+	printf("Array B length: %d\n", CH11__B_LENGTH);
+
+	printf("\n_____ merge_sort _____\n\n");
+
+	printf("Running on Device with 256 threads per block...");
+	ch11__merge_sort(Device, {.block_dim = {256, 1, 1}, .kernel_version = CH11__BASIC_MERGE_SORT});
+
+	printf("\n_____ merge_sort_tiled _____\n\n");
+	
+	printf("Running on Device with 256 threads per block...");
+	ch11__merge_sort(Device, {.block_dim = {256}, .kernel_version = CH11__TILED_MERGE_SORT});
+
+	printf("\n_____ merge_sort_circular_buffer _____\n\n");
+
+	printf("Running on Device with 256 threads per block...");
+	ch11__merge_sort(Device, {.block_dim = {256, 1, 1}, .kernel_version = CH11__CIRCULAR_BUFFER_MERGE_SORT});
+
+
+	printf("\n_____ merge_sort_CPU_____\n");
+	ch11__merge_sort(Host, {});
+
+	return 0;
+}