Merge pull request #117 from PSuchita/dev

Texture + Stream Bug Fix 

2 files changed, 39 insertions, 1 deletions

diff --git a/src/gpgpu-sim/gpu-sim.cc b/src/gpgpu-sim/gpu-sim.cc
index ec570bf..3687448 100644
--- a/src/gpgpu-sim/gpu-sim.cc
+++ b/src/gpgpu-sim/gpu-sim.cc
@@ -499,6 +499,12 @@ void gpgpu_sim_config::reg_options(option_parser_t opp)
    option_parser_register(opp, "-liveness_message_freq", OPT_INT64, &liveness_message_freq, 
                "Minimum number of seconds between simulation liveness messages (0 = always print)",
                "1");
+   option_parser_register(opp, "-gpgpu_compute_capability_major", OPT_UINT32, &gpgpu_compute_capability_major,
+                 "Major compute capability version number",
+                 "7");
+   option_parser_register(opp, "-gpgpu_compute_capability_minor", OPT_UINT32, &gpgpu_compute_capability_minor,
+                 "Minor compute capability version number",
+                 "0");
    option_parser_register(opp, "-gpgpu_flush_l1_cache", OPT_BOOL, &gpgpu_flush_l1_cache,
                 "Flush L1 cache at the end of each kernel call",
                 "0");
@@ -532,6 +538,14 @@ void gpgpu_sim_config::reg_options(option_parser_t opp)
    option_parser_register(opp, "-visualizer_zlevel", OPT_INT32,
                           &g_visualizer_zlevel, "Compression level of the visualizer output log (0=no comp, 9=highest)",
                           "6");
+   option_parser_register(opp, "-gpgpu_stack_size_limit", OPT_INT32, &stack_size_limit,
+                          "GPU thread stack size", "1024" );
+   option_parser_register(opp, "-gpgpu_heap_size_limit", OPT_INT32, &heap_size_limit,
+                          "GPU malloc heap size ", "8388608" );
+   option_parser_register(opp, "-gpgpu_runtime_sync_depth_limit", OPT_INT32, &runtime_sync_depth_limit,
+                          "GPU device runtime synchronize depth", "2" );
+   option_parser_register(opp, "-gpgpu_runtime_pending_launch_count_limit", OPT_INT32, &runtime_pending_launch_count_limit,
+                          "GPU device runtime pending launch count", "2048" );
     option_parser_register(opp, "-trace_enabled", OPT_BOOL, 
                           &Trace::enabled, "Turn on traces",
                           "0");
@@ -794,6 +808,16 @@ void gpgpu_sim::set_prop( cudaDeviceProp *prop )
    m_cuda_properties = prop;
 }
 
+int gpgpu_sim::compute_capability_major() const
+{
+   return m_config.gpgpu_compute_capability_major;
+}
+
+int gpgpu_sim::compute_capability_minor() const
+{
+   return m_config.gpgpu_compute_capability_minor;
+}
+
 const struct cudaDeviceProp *gpgpu_sim::get_prop() const
 {
    return m_cuda_properties;
diff --git a/src/gpgpu-sim/gpu-sim.h b/src/gpgpu-sim/gpu-sim.h
index 6ce5524..d4cc18f 100644
--- a/src/gpgpu-sim/gpu-sim.h
+++ b/src/gpgpu-sim/gpu-sim.h
@@ -337,6 +337,11 @@ public:
     unsigned get_max_concurrent_kernel() const { return max_concurrent_kernel; }
     unsigned checkpoint_option;
 
+    size_t stack_limit() const {return stack_size_limit; }
+    size_t heap_limit() const {return heap_size_limit; }
+    size_t sync_depth_limit() const {return runtime_sync_depth_limit; }
+    size_t pending_launch_count_limit() const {return runtime_pending_launch_count_limit;}
+
 private:
     void init_clock_domains(void ); 
 
@@ -377,8 +382,15 @@ private:
     int gpu_stat_sample_freq;
     int gpu_runtime_stat_flag;
 
+    // Device Limits
+    size_t stack_size_limit;
+    size_t heap_size_limit;
+    size_t runtime_sync_depth_limit;
+    size_t runtime_pending_launch_count_limit;	
 
-
+ //gpu compute capability options
+    unsigned int gpgpu_compute_capability_major;
+    unsigned int gpgpu_compute_capability_minor;
     unsigned long long liveness_message_freq; 
 
     friend class gpgpu_sim;
@@ -438,6 +450,8 @@ public:
 
    int shared_mem_size() const;
    int shared_mem_per_block() const;
+   int compute_capability_major() const;
+   int compute_capability_minor() const;
    int num_registers_per_core() const;
    int num_registers_per_block() const;
    int wrp_size() const;