- add support for cvta and isspacep instructions (currently assuming

a fixed address mapping between shared,local to generic that depends
upon hardware thread context used...  might be interesting to explore
tradeoffs at some point)
- remove util.h... we don't need TRUE, FALSE anymore now that 
everything is C++
- remove some dead code from shader_decode

[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 6867]

14 files changed, 14 insertions, 24 deletions

diff --git a/src/gpgpu-sim/cflogger.h b/src/gpgpu-sim/cflogger.h
index bdf57e5..2afe732 100644
--- a/src/gpgpu-sim/cflogger.h
+++ b/src/gpgpu-sim/cflogger.h
@@ -66,6 +66,8 @@
 #ifndef CFLOGGER_H
 #define CFLOGGER_H
 
+#include "../abstract_hardware_model.h"
+
 void try_snap_shot (unsigned long long  current_cycle);
 void set_spill_interval (unsigned long long  interval);
 void spill_log_to_file (FILE *fout, int final, unsigned long long  current_cycle);
diff --git a/src/gpgpu-sim/delayqueue.h b/src/gpgpu-sim/delayqueue.h
index baa3892..a8d6fc1 100644
--- a/src/gpgpu-sim/delayqueue.h
+++ b/src/gpgpu-sim/delayqueue.h
@@ -70,8 +70,6 @@
 #ifndef DELAYQUEUE_H
 #define DELAYQUEUE_H
 
-#include "../util.h"
-
 typedef struct delay_data_t delay_data;
 struct delay_data_t {
    void *data;
diff --git a/src/gpgpu-sim/dram_sched.cc b/src/gpgpu-sim/dram_sched.cc
index 24b518f..f6fdeca 100644
--- a/src/gpgpu-sim/dram_sched.cc
+++ b/src/gpgpu-sim/dram_sched.cc
@@ -67,7 +67,7 @@
 #include "dram_sched.h"
 #include "gpu-misc.h"
 #include "gpu-sim.h"
-#include "../util.h"
+#include "../abstract_hardware_model.h"
 
 extern unsigned long long  gpu_sim_cycle;
 extern signed long long gpu_tot_sim_cycle;
diff --git a/src/gpgpu-sim/dwf.h b/src/gpgpu-sim/dwf.h
index 6328f1a..9771af5 100644
--- a/src/gpgpu-sim/dwf.h
+++ b/src/gpgpu-sim/dwf.h
@@ -78,7 +78,7 @@
 
 #endif
 
-#include "../util.h"
+#include "../abstract_hardware_model.h"
 
 extern unsigned *acc_dyn_pcs;
 
diff --git a/src/gpgpu-sim/gpu-cache.h b/src/gpgpu-sim/gpu-cache.h
index e82ad05..cdb3414 100644
--- a/src/gpgpu-sim/gpu-cache.h
+++ b/src/gpgpu-sim/gpu-cache.h
@@ -66,7 +66,7 @@
 
 #include <stdio.h>
 #include <stdlib.h>
-#include "../util.h"
+#include "../abstract_hardware_model.h"
 
 #ifndef GPU_CACHE_H
 #define GPU_CACHE_H
diff --git a/src/gpgpu-sim/gpu-sim.cc b/src/gpgpu-sim/gpu-sim.cc
index 6ae7430..d77d1e0 100644
--- a/src/gpgpu-sim/gpu-sim.cc
+++ b/src/gpgpu-sim/gpu-sim.cc
@@ -168,7 +168,7 @@ char *gpgpu_cache_dl2_opt;
 extern int gpgpu_l2_readoverwrite;
 int gpgpu_partial_write_mask = 0;
 
-int gpgpu_perfect_mem = FALSE;
+bool gpgpu_perfect_mem = false;
 char *gpgpu_shader_core_pipeline_opt;
 extern unsigned int *requests_by_warp;
 unsigned int gpgpu_dram_buswidth = 4;
@@ -258,7 +258,7 @@ extern unsigned int warp_size;
 extern int pipe_simd_width;
 extern unsigned int gpgpu_dwf_heuristic;
 extern unsigned int gpgpu_dwf_regbk;
-int gpgpu_reg_bankconflict = FALSE;
+bool gpgpu_reg_bankconflict = false;
 extern int gpgpu_shmem_port_per_bank;
 extern int gpgpu_cache_port_per_bank;
 extern int gpgpu_const_port_per_bank;
diff --git a/src/gpgpu-sim/l2cache.cc b/src/gpgpu-sim/l2cache.cc
index aef837b..03a6a25 100644
--- a/src/gpgpu-sim/l2cache.cc
+++ b/src/gpgpu-sim/l2cache.cc
@@ -13,6 +13,7 @@
 #include "histogram.h"
 #include "l2cache.h"
 #include "../intersim/statwraper.h"
+#include "../abstract_hardware_model.h"
 
 class L2c_mshr;
 class L2c_miss_tracker;
diff --git a/src/gpgpu-sim/mem_fetch.h b/src/gpgpu-sim/mem_fetch.h
index 0e12a2d..dbd3a2a 100644
--- a/src/gpgpu-sim/mem_fetch.h
+++ b/src/gpgpu-sim/mem_fetch.h
@@ -69,6 +69,7 @@
 
 #include "shader.h"
 #include "addrdec.h"
+#include "../abstract_hardware_model.h"
 
 enum mf_type {
    RD_REQ = 0,
diff --git a/src/gpgpu-sim/mem_latency_stat.h b/src/gpgpu-sim/mem_latency_stat.h
index 2fa6b1c..150a1be 100644
--- a/src/gpgpu-sim/mem_latency_stat.h
+++ b/src/gpgpu-sim/mem_latency_stat.h
@@ -80,7 +80,7 @@ extern unsigned int gpu_mem_n_bk;
    #define EXTERN_DEF extern
 #endif
 
-EXTERN_DEF int gpgpu_memlatency_stat = FALSE;
+EXTERN_DEF bool gpgpu_memlatency_stat = false;
 
 EXTERN_DEF unsigned max_mrq_latency;
 EXTERN_DEF unsigned max_dq_latency;
diff --git a/src/gpgpu-sim/shader.cc b/src/gpgpu-sim/shader.cc
index 8f3f700..c173206 100644
--- a/src/gpgpu-sim/shader.cc
+++ b/src/gpgpu-sim/shader.cc
@@ -223,7 +223,6 @@ extern unsigned g_max_regs_per_thread;
 void ptx_decode_inst( void *thd, unsigned *op, int *i1, int *i2, int *i3, int *i4, int *o1, int *o2, int *o3, int *o4, int *vectorin, int *vectorout, int *arch_reg  );
 unsigned ptx_get_inst_op( void *thd);
 void ptx_exec_inst( void *thd, address_type *addr, unsigned *space, unsigned *data_size, dram_callback_t* callback, unsigned warp_active_mask);
-int  ptx_branch_taken( void *thd );
 void ptx_sim_free_sm( void** thread_info );
 unsigned ptx_sim_init_thread( void** thread_info, int sid, unsigned tid,unsigned threads_left,unsigned num_threads, core_t *core, unsigned hw_cta_id, unsigned hw_warp_id);
 unsigned ptx_sim_cta_size();
@@ -1750,7 +1749,6 @@ void shader_decode( shader_core_ctx_t *shader,
    address_type addr;
    dram_callback_t callback;
    op_type op = NO_OP;
-   register int is_write;
    int tid;
    int i1, i2, i3, i4, o1, o2, o3, o4; //4 outputs needed for texture fetches in cuda-sim
    int i;
@@ -1950,16 +1948,6 @@ void shader_decode( shader_core_ctx_t *shader,
       shader->pipeline_reg[IF_ID][i].out[2] = o3;
       shader->pipeline_reg[IF_ID][i].out[3] = o4;
 
-      if ( op == STORE_OP ) {
-         is_write = TRUE;
-      }
-
-      if ( op == BRANCH_OP ) {
-         int taken=0;
-         assert( gpgpu_cuda_sim );
-         taken = ptx_branch_taken(shader->thread[tid].ptx_thd_info);
-      }
-
       // go to the next instruction 
       // - done implicitly in ptx_exec_inst()
       
diff --git a/src/gpgpu-sim/shader.h b/src/gpgpu-sim/shader.h
index 8333202..1567d6a 100644
--- a/src/gpgpu-sim/shader.h
+++ b/src/gpgpu-sim/shader.h
@@ -139,7 +139,7 @@ typedef struct {
    int reg_bank_access_pending;
    int reg_bank_conflict_stall_checked; // flag to turn off register bank conflict checker to avoid double stalling
 
-   unsigned char inst_type;
+   unsigned inst_type;
 
    unsigned priority;
 
diff --git a/src/gpgpu-sim/stack.h b/src/gpgpu-sim/stack.h
index b00e25b..9ddab6d 100644
--- a/src/gpgpu-sim/stack.h
+++ b/src/gpgpu-sim/stack.h
@@ -69,7 +69,7 @@
 #ifndef _MY_STACK_
 #define _MY_STACK_
 
-#include "../util.h"
+#include "../abstract_hardware_model.h"
 
 typedef struct {
    address_type *v;
diff --git a/src/gpgpu-sim/stat-tool.cc b/src/gpgpu-sim/stat-tool.cc
index 2dc9c0b..9845e06 100644
--- a/src/gpgpu-sim/stat-tool.cc
+++ b/src/gpgpu-sim/stat-tool.cc
@@ -88,6 +88,7 @@
 #endif
 
 #include "histogram.h"
+#include "../abstract_hardware_model.h"
 
 binned_histogram::binned_histogram (std::string name, int nbins, int* bins) 
    : m_name(name), m_nbins(nbins), m_bins(NULL), m_bin_cnts(new int[m_nbins]), m_maximum(0)
@@ -178,7 +179,6 @@ void linear_histogram::add2bin (int sample) {
 #include <map>
 #include <algorithm>
 #include <string>
-#include "../util.h"
 
 #include "cflogger.h"
 
diff --git a/src/gpgpu-sim/warp_tracker.h b/src/gpgpu-sim/warp_tracker.h
index a4faa09..5375543 100644
--- a/src/gpgpu-sim/warp_tracker.h
+++ b/src/gpgpu-sim/warp_tracker.h
@@ -79,7 +79,7 @@
 
 #endif
 
-#include "../util.h"
+#include "../abstract_hardware_model.h"
 #include "shader.h"
 
 void init_warp_tracker( );