Merge pull request #29 from Connie120/dev

Integrated CUDA samples as a submodule

Author: tgrogers

.gitmodules

@@ -7,3 +7,6 @@
 [submodule "src/cuda/pytorch_examples"]
 	path = src/cuda/pytorch_examples
 	url = https://github.com/pytorch/examples.git
+[submodule "src/cuda/cuda-samples"]
+	path = src/cuda/cuda-samples
+	url = https://github.com/NVIDIA/cuda-samples.git

src/Makefile

@@ -4,7 +4,7 @@ $(error You must run "source setup_environment before calling make")
 endif
 
 ifeq ($(CUDA_GT_7), 1)
-all:  GPU_Microbenchmark microbench rodinia_2.0-ft cutlass rodinia-3.1 pannotia proxy-apps ispass-2009 lonestargpu-2.0 polybench custom_apps heterosync#mlperf_inference mlperf_training
+all:  GPU_Microbenchmark microbench rodinia_2.0-ft cutlass rodinia-3.1 pannotia proxy-apps ispass-2009 lonestargpu-2.0 polybench custom_apps heterosync cuda_samples#mlperf_inference mlperf_training
 #all:   pannotia rodinia_2.0-ft proxy-apps  microbench rodinia-3.1 ispass-2009 lonestargpu-2.0 polybench parboil shoc custom_apps deeplearning cutlass GPU_Microbenchmark heterosync Deepbench_nvidia dragon-naive dragon-cdp
 else
 ifeq ($(CUDA_GT_4), 1)
@@ -19,8 +19,8 @@ accelwattch_hw_power: rodinia-3.1_hw_power parboil_hw_power cuda_samples-11.0_hw
 #Disable clean for now, It has a bug!
 # clean_dragon-naive clean_pannotia clean_proxy-apps
 
-clean: clean_rodinia_2.0-ft clean_dragon-cdp  clean_ispass-2009 clean_lonestargpu-2.0 clean_custom_apps clean_parboil clean_cutlass clean_rodinia-3.1 clean_heterosync clean_UVMSmart_test
-clean_accelwattch: clean_rodinia-3.1 clean_parboil clean_cutlass clean_cuda_samples clean_cuda_samples_hw_power clean_rodinia-3.1_hw_power clean_parboil_hw_power clean_accelwattch_ubench
+clean: clean_rodinia_2.0-ft clean_dragon-cdp  clean_ispass-2009 clean_lonestargpu-2.0 clean_custom_apps clean_parboil clean_cutlass clean_rodinia-3.1 clean_heterosync clean_UVMSmart_test clean_cuda_samples
+clean_accelwattch: clean_rodinia-3.1 clean_parboil clean_cutlass clean_cuda_samples-11.0 clean_cuda_samples_hw_power clean_rodinia-3.1_hw_power clean_parboil_hw_power clean_accelwattch_ubench
 
 clean_data:
 	./clean_data.sh
@@ -492,6 +492,10 @@ mlperf_training:
 	cd $(BINDIR)/$(BINSUBDIR) && sh ./install_mlperf_and_virtual_environments_training.sh | tee $(BINDIR)/$(BINSUBDIR)/mlperf_install_training.log
 	cd $(BINDIR)/$(BINSUBDIR) && rm install_mlperf_and_virtual_environments_training.sh
 
+cuda_samples:
+	mkdir -p $(BINDIR)/$(BINSUBDIR)/
+	make -C ./cuda/cuda-samples/
+	find $(GPUAPPS_ROOT)/src/cuda/cuda-samples/bin/x86_64/linux/release -maxdepth 1 -type f -executable -exec mv {} "$(BINDIR)/$(BINSUBDIR)/" \; ;
 
 clean_heterosync:
 	rm -rf cuda/heterosync
@@ -582,7 +586,7 @@ clean_rodinia-3.1_hw_power:
 	$(SETENV) make clean $(MAKE_ARGS) noinline=$(noinline) -C cuda/rodinia/3.1/cuda/pathfinder_k1/
 	$(SETENV) make clean $(MAKE_ARGS) noinline=$(noinline) -C cuda/rodinia/3.1/cuda/srad/srad_v1_k1
 
-clean_cuda_samples:
+clean_cuda_samples-11.0:
 	$(SETENV) make clean $(MAKE_ARGS) noinline=$(noinline) -C cuda/NVIDIA_CUDA-11.0_Samples/binomialOptions
 	$(SETENV) make clean $(MAKE_ARGS) noinline=$(noinline) -C cuda/NVIDIA_CUDA-11.0_Samples/dct8x8
 	$(SETENV) make clean $(MAKE_ARGS) noinline=$(noinline) -C cuda/NVIDIA_CUDA-11.0_Samples/fastWalshTransform
@@ -658,3 +662,6 @@ clean_mlperf_inference:
 
 clean_mlperf_training:
 	rm -rf $(BINDIR)/$(BINSUBDIR)/mlperf_training
+	
+clean_cuda_samples:
+	make clean -C ./cuda/cuda-samples