- simple device printf support added (supports %u, %d, %f)

- add support to call_impl for passing values in through local param space (need to add support for return)
- track local framesize during ptx parsing / update local stack-pointer during call/return
- detect appropriate param space for ld.param (and st.param) depending upon address symbol's scope
- putting back the crazy math for local mem allocation (seems to be required)
- bug fix for global byte array initialization
- force ld and st w/o space specifier to be generic_space
- make type_decode method of type_info_key
- adjust debug printing so listing of code and CFG information requires higher numbers (100 and 50)
- local param ld/st accesses treated like local ld/st for timing model

[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 6878]

1 files changed, 3 insertions, 10 deletions

diff --git a/src/cuda-sim/ptx_sim.cc b/src/cuda-sim/ptx_sim.cc
index 85e6185..dffc128 100644
--- a/src/cuda-sim/ptx_sim.cc
+++ b/src/cuda-sim/ptx_sim.cc
@@ -358,6 +358,7 @@ void ptx_thread_info::callstack_push( unsigned pc, unsigned rpc, const symbol *r
    m_last_was_call = true;
    m_callstack.push_back( stack_entry(m_symbol_table,m_func_info,pc,rpc,return_var_src,return_var_dst,call_uid) );
    m_regs.push_back( reg_map_t() );
+   m_local_mem_stack_pointer += m_func_info->local_mem_framesize(); 
 }
 
 #define POST_DOMINATOR 1 /* must match definition in shader.h */
@@ -368,7 +369,9 @@ extern void set_operand_value( const symbol *dst, const ptx_reg_t &data );
 bool ptx_thread_info::callstack_pop()
 {
    assert( !m_callstack.empty() );
+   assert( m_local_mem_stack_pointer >= m_func_info->local_mem_framesize() );
    m_func_info = m_callstack.back().m_func_info;
+   m_local_mem_stack_pointer -= m_func_info->local_mem_framesize(); 
    m_symbol_table = m_callstack.back().m_symbol_table;
    m_NPC = m_callstack.back().m_PC;
    m_RPC_updated = true;
@@ -448,16 +451,6 @@ void ptx_thread_info::dump_modifiedregs()
    }
 }
 
-void ptx_thread_info::push_local_mem_stack() 
-{ 
-   m_local_mem_stack_pointer += m_func_info->local_mem_framesize(); 
-}
-
-void ptx_thread_info::pop_local_mem_stack() 
-{ 
-   m_local_mem_stack_pointer -= m_func_info->local_mem_framesize(); 
-}
-
 void ptx_thread_info::set_npc( const function_info *f )
 {
    m_NPC = f->get_start_PC();