Merge branch 'dev' of https://github.com/gpgpu-sim/gpgpu-sim_distribution into dev-private

5 files changed, 37 insertions, 7 deletions

diff --git a/.travis.yml b/.travis.yml
index b38f468..82f0604 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -18,5 +18,10 @@ matrix:
         - CONFIG=TITANV
         - CUDA_INSTALL_PATH=/usr/local/cuda-9.1/
         - PTXAS_CUDA_INSTALL_PATH=/usr/local/cuda-9.1/
+    - services: docker
+      env:
+        - CONFIG=TITANV-LOCALXBAR
+        - CUDA_INSTALL_PATH=/usr/local/cuda-9.1/
+        - PTXAS_CUDA_INSTALL_PATH=/usr/local/cuda-9.1/
 
 script: docker run -v `pwd`:/home/runner/gpgpu-sim_distribution:rw tgrogers/gpgpu-sim_regress:volta_update /bin/bash -c "./start_torque.sh; chown -R runner /home/runner/gpgpu-sim_distribution; su - runner -c 'export CUDA_INSTALL_PATH=$CUDA_INSTALL_PATH && export PTXAS_CUDA_INSTALL_PATH=$PTXAS_CUDA_INSTALL_PATH && source /home/runner/gpgpu-sim_distribution/setup_environment && make -j -C /home/runner/gpgpu-sim_distribution && cd /home/runner/gpgpu-sim_simulations/ && git pull && /home/runner/gpgpu-sim_simulations/util/job_launching/run_simulations.py -C $CONFIG -B rodinia_2.0-ft -N regress && /home/runner/gpgpu-sim_simulations/util/job_launching/monitor_func_test.py -v -N regress'"
diff --git a/Makefile b/Makefile
index 7a3bb47..f4e8e4b 100644
--- a/Makefile
+++ b/Makefile
@@ -168,6 +168,7 @@ $(SIM_LIB_DIR)/libcudart.so: makedirs $(LIBS) cudalib
 	if [ ! -f $(SIM_LIB_DIR)/libcudart.so.10.0 ]; then ln -s libcudart.so $(SIM_LIB_DIR)/libcudart.so.10.0; fi
 	if [ ! -f $(SIM_LIB_DIR)/libcudart.so.10.1 ]; then ln -s libcudart.so $(SIM_LIB_DIR)/libcudart.so.10.1; fi
 
+
 $(SIM_LIB_DIR)/libcudart.dylib: makedirs $(LIBS) cudalib
 	g++ -dynamiclib -Wl,-headerpad_max_install_names,-undefined,dynamic_lookup,-compatibility_version,1.1,-current_version,1.1\
 			$(SIM_OBJ_FILES_DIR)/libcuda/*.o \
diff --git a/README b/README
index 998cb51..ff6664f 100644
--- a/README
+++ b/README
@@ -177,7 +177,7 @@ GPGPU-Sim dependencies:
 * flex
 * zlib
 * CUDA Toolkit
-	
+
 GPGPU-Sim documentation dependencies:
 * doxygen
 * graphvi
@@ -208,6 +208,9 @@ python-matplotlib"
 CUDA SDK dependencies:
 "sudo apt-get install libxi-dev libxmu-dev libglut3-dev"
 
+If you are running applications which use NVIDIA libraries such as cuDNN and 
+cuBLAS, install them too.
+
 Finally, ensure CUDA_INSTALL_PATH is set to the location where you installed
 the CUDA Toolkit (e.g., /usr/local/cuda) and that $CUDA_INSTALL_PATH/bin is in
 your PATH.  You probably want to modify your .bashrc file to incude the
@@ -216,6 +219,10 @@ following (this assumes the CUDA Toolkit was installed in /usr/local/cuda):
 	export CUDA_INSTALL_PATH=/usr/local/cuda
 	export PATH=$CUDA_INSTALL_PATH/bin
 
+If running applications which use cuDNN or cuBLAS:
+	export CUDNN_PATH=<Path To cuDNN Directory>
+	export LD_LIBRARY_PATH=$CUDA_INSTALL_PATH/lib64:$CUDA_INSTALL_PATH/lib:$CUDNN_PATH/lib64
+	
 
 Step 2: Build
 =============
@@ -261,6 +268,23 @@ ldd <your_application_name>
 
 You should see that your application is using libcudart.so file in GPGPUSim directory.
 
+If running applications which use cuDNN or cuBLAS:
+
+ * Modify the Makefile or the compilation command of the application to change 
+   all the dynamic links to static ones, for example:
+	* -L$(CUDA_PATH)/lib64 -lcublas to
+	  -L$(CUDA_PATH)/lib64 -lcublas_static
+
+	* -L$(CUDNN_PATH)/lib64 -lcudnn to
+	  -L$(CUDNN_PATH)/lib64 -lcudnn_static
+
+ * Modify the Makefile or the compilation command such that the following 
+   flags are used by the nvcc compiler:
+	-gencode arch=compute_61,code=compute_61
+
+   (the number 61 refers to the SM version. You would need to set it based 
+   on the GPGPU-Sim config "-gpgpu ptx force max capability" you use)
+
 Copy the contents of configs/QuadroFX5800/ or configs/GTX480/ to your
 application's working directory.  These files configure the microarchitecture
 models to resemble the respective GPGPU architectures.
diff --git a/configs/tested-cfgs/SM7_QV100/gpgpusim.config b/configs/tested-cfgs/SM7_QV100/gpgpusim.config
index 16c131b..1a34d0f 100644
--- a/configs/tested-cfgs/SM7_QV100/gpgpusim.config
+++ b/configs/tested-cfgs/SM7_QV100/gpgpusim.config
@@ -11,7 +11,7 @@
 # functional simulator specification
 -gpgpu_ptx_instruction_classification 0
 -gpgpu_ptx_sim_mode 0
--gpgpu_ptx_force_max_capability 70 
+-gpgpu_ptx_force_max_capability 60 
 
 
 # Device Limits
@@ -21,7 +21,7 @@
 -gpgpu_runtime_pending_launch_count_limit 2048
 
 # Compute Capability
--gpgpu_compute_capability_major 7
+-gpgpu_compute_capability_major 6
 -gpgpu_compute_capability_minor 0
 
 # SASS execution (only supported with CUDA >= 4.0)
@@ -44,7 +44,7 @@
 
 # shader core pipeline config
 -gpgpu_shader_registers 65536
--gpgpu_occupancy_sm_number 70
+-gpgpu_occupancy_sm_number 60
 
 # This implies a maximum of 64 warps/SM
 -gpgpu_shader_core_pipeline 2048:32 
diff --git a/src/gpgpu-sim/gpu-cache.h b/src/gpgpu-sim/gpu-cache.h
index 673e128..9725270 100644
--- a/src/gpgpu-sim/gpu-cache.h
+++ b/src/gpgpu-sim/gpu-cache.h
@@ -681,15 +681,15 @@ public:
         // tag + index is required to check for hit/miss. Tag is now identical to the block address.
 
         //return addr >> (m_line_sz_log2+m_nset_log2);
-        return addr & ~(m_line_sz-1);
+        return addr & ~(new_addr_type)(m_line_sz-1);
     }
     new_addr_type block_addr( new_addr_type addr ) const
     {
-        return addr & ~(m_line_sz-1);
+        return addr & ~(new_addr_type)(m_line_sz-1);
     }
     new_addr_type mshr_addr( new_addr_type addr ) const
 	{
-    	return addr & ~(m_atom_sz-1);
+    	return addr & ~(new_addr_type)(m_atom_sz-1);
 	}
     enum mshr_config_t get_mshr_type() const
 	{