- integrate changes from fermi-test (CL's under that path in range 7261-7418).

(add scoreboard logic from ptxplus branch and modified operand collector
with parallel ALU/SFU pipelines)

passing regressions

[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 7419]

1 files changed, 3 insertions, 3 deletions

diff --git a/src/cuda-sim/instructions.cc b/src/cuda-sim/instructions.cc
index 47b3ebb..cbf4c44 100644
--- a/src/cuda-sim/instructions.cc
+++ b/src/cuda-sim/instructions.cc
@@ -2639,7 +2639,7 @@ void st_impl( const ptx_instruction *pI, ptx_thread_info *thread )
          thread->get_vector_operand_values(src1, ptx_regs, 2); 
          mem->write(addr,size/8,&ptx_regs[0].s64,thread,pI);
          mem->write(addr+size/8,size/8,&ptx_regs[1].s64,thread,pI);
-         free(ptx_regs);
+         delete [] ptx_regs;
       }
       if (vector_spec == V3_TYPE) {
          ptx_reg_t* ptx_regs = new ptx_reg_t[3]; 
@@ -2647,7 +2647,7 @@ void st_impl( const ptx_instruction *pI, ptx_thread_info *thread )
          mem->write(addr,size/8,&ptx_regs[0].s64,thread,pI);
          mem->write(addr+size/8,size/8,&ptx_regs[1].s64,thread,pI);
          mem->write(addr+2*size/8,size/8,&ptx_regs[2].s64,thread,pI);
-         free(ptx_regs);
+         delete [] ptx_regs;
       }
       if (vector_spec == V4_TYPE) {
          ptx_reg_t* ptx_regs = new ptx_reg_t[4]; 
@@ -2656,7 +2656,7 @@ void st_impl( const ptx_instruction *pI, ptx_thread_info *thread )
          mem->write(addr+size/8,size/8,&ptx_regs[1].s64,thread,pI);
          mem->write(addr+2*size/8,size/8,&ptx_regs[2].s64,thread,pI);
          mem->write(addr+3*size/8,size/8,&ptx_regs[3].s64,thread,pI);
-         free(ptx_regs);
+         delete [] ptx_regs;
       }
    }
    thread->m_last_effective_address = addr;