diff options
Diffstat (limited to 'src/gpgpu-sim/shader.cc')
| -rw-r--r-- | src/gpgpu-sim/shader.cc | 14 |
1 files changed, 8 insertions, 6 deletions
diff --git a/src/gpgpu-sim/shader.cc b/src/gpgpu-sim/shader.cc index 6f68283..4ae0f62 100644 --- a/src/gpgpu-sim/shader.cc +++ b/src/gpgpu-sim/shader.cc @@ -2416,8 +2416,10 @@ void pipelined_simd_unit::cycle() { if (!m_dispatch_reg->dispatch_delay()) { int start_stage = m_dispatch_reg->latency - m_dispatch_reg->initiation_interval; - move_warp(m_pipeline_reg[start_stage], m_dispatch_reg); - active_insts_in_pipeline++; + if(m_pipeline_reg[start_stage]->empty()) { + move_warp(m_pipeline_reg[start_stage], m_dispatch_reg); + active_insts_in_pipeline++; + } } } occupied >>= 1; @@ -3080,7 +3082,7 @@ void warp_inst_t::print(FILE *fout) const { fprintf(fout, "bubble\n"); return; } else - fprintf(fout, "0x%04x ", pc); + fprintf(fout, "0x%04llx ", pc); fprintf(fout, "w%02d[", m_warp_id); for (unsigned j = 0; j < m_config->warp_size; j++) fprintf(fout, "%c", (active(j) ? '1' : '0')); @@ -3266,7 +3268,7 @@ void shader_core_ctx::display_pipeline(FILE *fout, int print_mem, if (!m_inst_fetch_buffer.m_valid) fprintf(fout, "bubble\n"); else { - fprintf(fout, "w%2u : pc = 0x%x, nbytes = %u\n", + fprintf(fout, "w%2u : pc = 0x%llx, nbytes = %u\n", m_inst_fetch_buffer.m_warp_id, m_inst_fetch_buffer.m_pc, m_inst_fetch_buffer.m_nbytes); } @@ -3932,7 +3934,7 @@ bool shd_warp_t::waiting() { void shd_warp_t::print(FILE *fout) const { if (!done_exit()) { - fprintf(fout, "w%02u npc: 0x%04x, done:%c%c%c%c:%2u i:%u s:%u a:%u (done: ", + fprintf(fout, "w%02u npc: 0x%04llx, done:%c%c%c%c:%2u i:%u s:%u a:%u (done: ", m_warp_id, m_next_pc, (functional_done() ? 'f' : ' '), (stores_done() ? 's' : ' '), (inst_in_pipeline() ? ' ' : 'i'), (done_exit() ? 'e' : ' '), n_completed, m_inst_in_pipeline, @@ -4008,7 +4010,7 @@ void opndcoll_rfu_t::init(unsigned num_banks, shader_core_ctx *shader) { sub_core_model = shader->get_config()->sub_core_model; m_num_warp_scheds = shader->get_config()->gpgpu_num_sched_per_core; - unsigned reg_id; + unsigned reg_id = 0; if (sub_core_model) { assert(num_banks % shader->get_config()->gpgpu_num_sched_per_core == 0); assert(m_num_warp_scheds <= m_cu.size() && |