refactor: remove extern decl up to gpgpusim_entrypoint.cc

[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 7178]

14 files changed, 42 insertions, 74 deletions

diff --git a/src/cuda-sim/cuda-sim.cc b/src/cuda-sim/cuda-sim.cc
index a55fb79..9d62a59 100644
--- a/src/cuda-sim/cuda-sim.cc
+++ b/src/cuda-sim/cuda-sim.cc
@@ -80,20 +80,17 @@
 #include "ptx_loader.h"
 #include "ptx_parser.h"
 #include "../gpgpu-sim/gpu-sim.h"
-
-extern bool g_interactive_debugger_enabled;
+#include "ptx_sim.h"
+#include "../gpgpusim_entrypoint.h"
 
 int gpgpu_ptx_instruction_classification=0;
 void ** g_inst_classification_stat = NULL;
 void ** g_inst_op_classification_stat= NULL;
 int g_ptx_kernel_count = -1; // used for classification stat collection purposes 
-
 int g_debug_execution = 0;
 int g_debug_thread_uid = 0;
 addr_t g_debug_pc = 0xBEEF1518;
-
 unsigned g_ptx_sim_num_insn = 0;
-
 std::map<const struct textureReference*,const struct cudaArray*> TextureToArrayMap; // texture bindings
 std::map<const struct textureReference*, const struct textureInfo*> TextureToInfoMap;
 std::map<std::string, const struct textureReference*> NameToTextureMap;
@@ -1115,7 +1112,6 @@ void init_inst_classification_stat() {
 
 std::map<std::string,function_info*> *g_kernel_name_to_function_lookup=NULL;
 std::map<const void*,std::string> *g_host_to_kernel_entrypoint_name_lookup=NULL;
-extern unsigned g_ptx_thread_info_uid_next;
 
 void gpgpu_cuda_ptx_sim_init_grid( const char *kernel_key, struct gpgpu_ptx_sim_arg* args,
                                          struct dim3 gridDim, struct dim3 blockDim ) 
@@ -1388,13 +1384,6 @@ void read_sim_environment_variables()
    }
 }
 
-
-
-
-
-
-extern time_t g_simulation_starttime;
-
 ptx_cta_info *g_func_cta_info = NULL;
 
 #define MAX(a,b) (((a)>(b))?(a):(b))
diff --git a/src/cuda-sim/ptx_sim.h b/src/cuda-sim/ptx_sim.h
index 0379686..e9270be 100644
--- a/src/cuda-sim/ptx_sim.h
+++ b/src/cuda-sim/ptx_sim.h
@@ -90,13 +90,11 @@ struct gpgpu_ptx_sim_kernel_info {
 #include <assert.h>
 #include "opcodes.h"
 
-#ifdef __cplusplus
-
-   #include <string>
-   #include <map>
-   #include <set>
-   #include <list>
-   #include <unordered_map>
+#include <string>
+#include <map>
+#include <set>
+#include <list>
+#include <unordered_map>
 
 #include "memory.h"
 
@@ -484,6 +482,6 @@ bool isspace_shared( unsigned smid, addr_t addr );
 bool isspace_global( addr_t addr );
 memory_space_t whichspace( addr_t addr );
 
-#endif
+extern unsigned g_ptx_thread_info_uid_next;
 
 #endif
diff --git a/src/gpgpu-sim/addrdec.cc b/src/gpgpu-sim/addrdec.cc
index 333d6de..0ed44d0 100644
--- a/src/gpgpu-sim/addrdec.cc
+++ b/src/gpgpu-sim/addrdec.cc
@@ -65,11 +65,10 @@
 
 #include <string.h>
 #include "addrdec.h"
-//#include "gpu-sim.h"
+#include "gpu-sim.h"
 #include "../option_parser.h"
 
 int ADDR_CHIP_S = 10;
-extern int gpgpu_mem_address_mask;
 
 long int powli( long int x, long int y ) // compute x to the y
 {
diff --git a/src/gpgpu-sim/dram.cc b/src/gpgpu-sim/dram.cc
index a4e69f8..f86d968 100644
--- a/src/gpgpu-sim/dram.cc
+++ b/src/gpgpu-sim/dram.cc
@@ -66,16 +66,9 @@
 
 #include "gpu-sim.h"
 #include "gpu-misc.h"
-//#include "shader.h"
 #include "dram.h"
 #include "mem_latency_stat.h"
 
-extern unsigned max_dq_latency;
-extern unsigned max_dq_latency;
-unsigned int gpu_n_warps;
-extern unsigned int gpu_n_mem_per_ctrlr;
-extern unsigned int recent_dram_util;
-
 #ifdef DRAM_VERIFY
 int PRINT_CYCLE = 0;
 #endif
diff --git a/src/gpgpu-sim/dram_sched.cc b/src/gpgpu-sim/dram_sched.cc
index 30c8752..e678042 100644
--- a/src/gpgpu-sim/dram_sched.cc
+++ b/src/gpgpu-sim/dram_sched.cc
@@ -68,11 +68,7 @@
 #include "gpu-misc.h"
 #include "gpu-sim.h"
 #include "../abstract_hardware_model.h"
-
-extern unsigned max_mrq_latency;
-extern unsigned mrq_lat_table[24];
-extern int gpgpu_memlatency_stat;
-extern unsigned int **concurrent_row_access; //concurrent_row_access[dram chip id][bank id]
+#include "mem_latency_stat.h"
 
 ideal_dram_scheduler::ideal_dram_scheduler( dram_t *dm )
 {
diff --git a/src/gpgpu-sim/gpu-sim.cc b/src/gpgpu-sim/gpu-sim.cc
index 0cef489..f0646d2 100644
--- a/src/gpgpu-sim/gpu-sim.cc
+++ b/src/gpgpu-sim/gpu-sim.cc
@@ -86,7 +86,7 @@
 #include "addrdec.h"
 #include "dwf.h"
 #include "warp_tracker.h"
-#include "cflogger.h"
+#include "stat-tool.h"
 #include "l2cache.h"
 
 #include "../cuda-sim/ptx-stats.h"
@@ -209,7 +209,6 @@ void visualizer_printstat();
 void time_vector_create(int ld_size,int st_size);
 void time_vector_print(void);
 void time_vector_update(unsigned int uid,int slot ,long int cycle,int type);
-void check_time_vector_update(unsigned int uid,int slot ,long int latency,int type);
 void node_req_hist_clear(void *p);
 void node_req_hist_dump(void *p);
 void node_req_hist_update(void * p,int node, long long cycle);
diff --git a/src/gpgpu-sim/gpu-sim.h b/src/gpgpu-sim/gpu-sim.h
index 58ff4de..338336e 100644
--- a/src/gpgpu-sim/gpu-sim.h
+++ b/src/gpgpu-sim/gpu-sim.h
@@ -72,10 +72,6 @@
 #define NUM_SHADERS 8
 #define STREAMS_PER_FILE 128
 
-//unsigned int run_gpu_sim(int grid_num);
-
-unsigned int get_converge_point(unsigned int pc, void *thd);
-
 #define GPU_RSTAT_SHD_INFO 0x1
 #define GPU_RSTAT_BW_STAT  0x2
 #define GPU_RSTAT_WARP_DIS 0x4
@@ -95,17 +91,16 @@ unsigned int get_converge_point(unsigned int pc, void *thd);
 #define GPU_MEMLATSTAT_MC 0x2
 #define GPU_MEMLATSTAT_QUEUELOGS 0x4
 
-//void gpu_reg_options(option_parser_t opp);
-//void init_gpu();
-void gpu_print_stat();
-
-int mem_ctrl_full( int mc_id );
+#define MhZ *1000000
 
-void dramqueue_latency_log_dump();
-void dump_pipeline_impl( int mask, int s, int m );
-unsigned int run_gpu_sim(int grid_num);
-extern void gpu_reg_options(class OptionParser * opp);
-extern void init_gpu();
+extern void         init_gpu();
+extern void         gpu_reg_options(class OptionParser * opp);
+extern unsigned int run_gpu_sim(int grid_num);
+extern unsigned int get_converge_point(unsigned int pc, void *thd);
+extern void         gpu_print_stat();
+extern int          mem_ctrl_full( int mc_id );
+extern void         dramqueue_latency_log_dump();
+extern void         dump_pipeline_impl( int mask, int s, int m );
 
 extern unsigned int L1_write_miss;
 extern unsigned int L1_read_miss;
@@ -167,11 +162,11 @@ extern int pdom_sched_type;
 extern int n_pdom_sc_orig_stat;
 extern int n_pdom_sc_single_stat;
 extern int gpgpu_cuda_sim;
-
-#ifndef MhZ
-   #define MhZ *1000000
-#endif
-
-extern void check_time_vector_update(unsigned int uid,int slot ,long int latency,int type);
+extern int gpgpu_mem_address_mask;
+extern bool g_interactive_debugger_enabled;
+extern unsigned int gpu_n_mem_per_ctrlr;
+extern unsigned int recent_dram_util;
+extern unsigned int **concurrent_row_access; //concurrent_row_access[dram chip id][bank id]
+extern unsigned long long  gpu_tot_sim_insn;
 
 #endif
diff --git a/src/gpgpu-sim/mem_latency_stat.cc b/src/gpgpu-sim/mem_latency_stat.cc
index 81589de..19cc76b 100644
--- a/src/gpgpu-sim/mem_latency_stat.cc
+++ b/src/gpgpu-sim/mem_latency_stat.cc
@@ -70,8 +70,10 @@
 #include "gpu-misc.h"
 #include "shader.h"
 #include "mem_fetch.h"
-#include "cflogger.h"
+#include "stat-tool.h"
 #include "../cuda-sim/ptx-stats.h"
+#include "visualizer.h"
+
 #include <string.h>
 #include <stdlib.h>
 #include <stdio.h>
diff --git a/src/gpgpu-sim/shader.cc b/src/gpgpu-sim/shader.cc
index ef3a8c0..6cafc48 100644
--- a/src/gpgpu-sim/shader.cc
+++ b/src/gpgpu-sim/shader.cc
@@ -71,7 +71,7 @@
 #include "dram.h"
 #include "dwf.h"
 #include "warp_tracker.h"
-#include "cflogger.h"
+#include "stat-tool.h"
 #include "gpu-misc.h"
 #include "../cuda-sim/ptx_sim.h"
 #include "../cuda-sim/ptx-stats.h"
diff --git a/src/gpgpu-sim/stat-tool.cc b/src/gpgpu-sim/stat-tool.cc
index 98a6917..d957e0c 100644
--- a/src/gpgpu-sim/stat-tool.cc
+++ b/src/gpgpu-sim/stat-tool.cc
@@ -62,6 +62,8 @@
  * Vancouver, BC V6T 1Z4
  */
 
+#include "stat-tool.h"
+
 #include <stdio.h>
 #include <stdlib.h>
 #include <assert.h>
@@ -190,8 +192,6 @@ void linear_histogram::add2bin (int sample) {
 #include <algorithm>
 #include <string>
 
-#include "cflogger.h"
-
 /////////////////////////////////////////////////////////////////////////////////////
 // logger snapshot trigger: 
 // - automate the snap_shot part of loggers to avoid modifying simulation loop everytime 
diff --git a/src/gpgpu-sim/cflogger.h b/src/gpgpu-sim/stat-tool.h
index 2afe732..79de754 100644
--- a/src/gpgpu-sim/cflogger.h
+++ b/src/gpgpu-sim/stat-tool.h
@@ -1,5 +1,5 @@
 /* 
- * cflogger.h 
+ * stat-tool.h 
  *
  * Copyright (c) 2009 by Tor M. Aamodt, Wilson W. L. Fung, and the 
  * University of British Columbia
@@ -63,10 +63,11 @@
  * Vancouver, BC V6T 1Z4
  */
 
-#ifndef CFLOGGER_H
-#define CFLOGGER_H
+#ifndef STAT_TOOL_H
+#define STAT_TOOL_H
 
 #include "../abstract_hardware_model.h"
+#include <stdio.h>
 
 void try_snap_shot (unsigned long long  current_cycle);
 void set_spill_interval (unsigned long long  interval);
diff --git a/src/gpgpu-sim/visualizer.cc b/src/gpgpu-sim/visualizer.cc
index d74c5fb..6eec3be 100644
--- a/src/gpgpu-sim/visualizer.cc
+++ b/src/gpgpu-sim/visualizer.cc
@@ -65,6 +65,9 @@
 #include "l2cache.h"
 #include "shader.h"
 #include "../option_parser.h"
+#include "mem_latency_stat.h"
+#include "stat-tool.h"
+
 #include <time.h>
 #include <string.h>
 #include <zlib.h>
@@ -78,18 +81,14 @@ extern unsigned int gpgpu_n_processed_writes;
 extern unsigned int gpgpu_n_cache_bkconflict;
 extern unsigned int gpgpu_n_shmem_bkconflict;
 extern unsigned int *max_return_queue_length;
-extern unsigned max_mrq_latency;
-extern unsigned max_dq_latency;
-extern unsigned max_mf_latency;
-extern unsigned max_icnt2mem_latency;
-extern unsigned max_icnt2sh_latency;
 extern int gpgpu_warpdistro_shader;
 extern unsigned ***mem_access_type_stats;
 
 extern unsigned int warp_size; 
 extern unsigned int *shader_cycle_distro;
-void time_vector_print_interval2file(FILE *outfile);
-void time_vector_print_interval2gzfile(gzFile outfile);
+
+static void time_vector_print_interval2file(FILE *outfile);
+static void time_vector_print_interval2gzfile(gzFile outfile);
 void cflog_visualizer_gzprint(gzFile fout);
 void shader_CTA_count_visualizer_gzprint(gzFile fout);
 float shd_cache_windowed_cache_miss_rate(shd_cache_t*, int);
diff --git a/src/gpgpu-sim/visualizer.h b/src/gpgpu-sim/visualizer.h
index bbc94bf..19e2d2a 100644
--- a/src/gpgpu-sim/visualizer.h
+++ b/src/gpgpu-sim/visualizer.h
@@ -69,8 +69,6 @@ void visualizer_options(class OptionParser *opp);
 void visualizer_printstat();
 void time_vector_create(int ld_size,int st_size);
 void time_vector_print(void);
-void time_vector_print_interval2file(FILE *outfile);
-void time_vector_print_interval2gzfile(gzFile outfile);
 void time_vector_update(unsigned int uid,int slot ,long int cycle,int type);
 void check_time_vector_update(unsigned int uid,int slot ,long int latency,int type); 
 
diff --git a/src/gpgpusim_entrypoint.cc b/src/gpgpusim_entrypoint.cc
index 7707b35..2f304b7 100644
--- a/src/gpgpusim_entrypoint.cc
+++ b/src/gpgpusim_entrypoint.cc
@@ -102,7 +102,6 @@ void gpgpu_ptx_sim_init_perf()
    g_simulation_starttime = time((time_t *)NULL);
 }
 
-extern unsigned long long  gpu_tot_sim_insn;
 
 static void print_simulation_time()
 {