fixing a segfault problem for Quadro config with interconnect stats

[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 14737]

4 files changed, 12 insertions, 6 deletions

diff --git a/src/gpgpu-sim/gpu-sim.cc b/src/gpgpu-sim/gpu-sim.cc
index b2b00a4..9d3145b 100644
--- a/src/gpgpu-sim/gpu-sim.cc
+++ b/src/gpgpu-sim/gpu-sim.cc
@@ -759,7 +759,7 @@ void gpgpu_sim::gpu_print_stat()
       m_memory_partition_unit[i]->set_icnt_power_stats(temp);
       total_mem_to_simt += temp;
    }
-   for(unsigned i=0; i<m_config.num_shader(); i++){
+   for(unsigned i=0; i<m_config.num_cluster(); i++){
 	   unsigned temp=0;
 	   m_cluster[i]->set_icnt_stats(temp);
 	   total_simt_to_mem += temp;
diff --git a/src/gpgpu-sim/gpu-sim.h b/src/gpgpu-sim/gpu-sim.h
index 7866688..98f719f 100644
--- a/src/gpgpu-sim/gpu-sim.h
+++ b/src/gpgpu-sim/gpu-sim.h
@@ -302,6 +302,7 @@ public:
     }
 
     unsigned num_shader() const { return m_shader_config.num_shader(); }
+    unsigned num_cluster() const { return m_shader_config.n_simt_clusters; }
     unsigned get_max_concurrent_kernel() const { return max_concurrent_kernel; }
 
 private:
diff --git a/src/gpgpu-sim/shader.cc b/src/gpgpu-sim/shader.cc
index 4095cbc..ad097e8 100644
--- a/src/gpgpu-sim/shader.cc
+++ b/src/gpgpu-sim/shader.cc
@@ -971,10 +971,15 @@ void ldst_unit::set_icnt_power_stats(unsigned &simt_to_mem) const{
 	unsigned tex=0;
 	unsigned l1c=0;
 
-	m_L1D->set_icnt_power_stats(l1d);
-	m_L1T->set_icnt_power_stats(tex);
-	m_L1C->set_icnt_power_stats(l1c);
-
+	if( m_L1D ) {
+		m_L1D->set_icnt_power_stats(l1d);
+	}
+	if( m_L1T ){
+		m_L1T->set_icnt_power_stats(tex);
+	}
+	if( m_L1C ){
+		m_L1C->set_icnt_power_stats(l1c);
+	}
 	simt_to_mem = n_simt_to_mem+l1d+tex+l1c; // All components that push packets into the interconnect
 }
 
diff --git a/src/mcpat/gpgpu_sim_wrapper.cc b/src/mcpat/gpgpu_sim_wrapper.cc
index 0ed7f57..1b06e26 100644
--- a/src/mcpat/gpgpu_sim_wrapper.cc
+++ b/src/mcpat/gpgpu_sim_wrapper.cc
@@ -241,7 +241,7 @@ void gpgpu_sim_wrapper::set_inst_power(bool clk_gated_lanes, double tot_cycles,
 	p->sys.core[0].gpgpu_clock_gated_lanes = clk_gated_lanes;
 	p->sys.core[0].total_cycles = tot_cycles;
 	p->sys.core[0].busy_cycles = busy_cycles;
-	p->sys.core[0].total_instructions  = tot_inst * p->sys.scaling_coefficients[TOT_INST];;
+	p->sys.core[0].total_instructions  = tot_inst * p->sys.scaling_coefficients[TOT_INST];
 	p->sys.core[0].int_instructions    = int_inst * p->sys.scaling_coefficients[FP_INT];
 	p->sys.core[0].fp_instructions     = fp_inst  * p->sys.scaling_coefficients[FP_INT];
 	p->sys.core[0].load_instructions  = load_inst;