path: root/src/mem/ruby/tester/SyntheticDriver.cc

/*
 * Copyright (c) 1999-2008 Mark D. Hill and David A. Wood
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions are
 * met: redistributions of source code must retain the above copyright
 * notice, this list of conditions and the following disclaimer;
 * redistributions in binary form must reproduce the above copyright
 * notice, this list of conditions and the following disclaimer in the
 * documentation and/or other materials provided with the distribution;
 * neither the name of the copyright holders nor the names of its
 * contributors may be used to endorse or promote products derived from
 * this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

/*
 * $Id$
 *
 */

#include "Global.hh"
#include "System.hh"
#include "SyntheticDriver.hh"
#include "EventQueue.hh"
//#ifndef XACT_MEM
#include "RequestGenerator.hh"
//#endif
//#include "XactAbortRequestGenerator.hh"
//#include "XactRequestGenerator.hh"
#include "SubBlock.hh"
#include "Chip.hh"

SyntheticDriver::SyntheticDriver(System* sys_ptr)
{
  cout << "SyntheticDriver::SyntheticDriver" << endl;
  if (g_SIMICS) {
    ERROR_MSG("g_SIMICS should not be defined.");
  }

  m_finish_time = 0;
  m_done_counter = 0;

  m_last_progress_vector.setSize(RubyConfig::numberOfProcessors());
  for (int i=0; i<m_last_progress_vector.size(); i++) {
    m_last_progress_vector[i] = 0;
  }

  m_lock_vector.setSize(g_synthetic_locks);
  for (int i=0; i<m_lock_vector.size(); i++) {
    m_lock_vector[i] = -1;  // No processor last held it
  }

  m_request_generator_vector.setSize(RubyConfig::numberOfProcessors());
  for (int i=0; i<m_request_generator_vector.size(); i++) {
    if(XACT_MEMORY){
      //m_request_generator_vector[i] = new XactRequestGenerator(i, *this);
    } else {
      m_request_generator_vector[i] = new RequestGenerator(i, *this);
    }
  }

  // add the tester consumer to the global event queue
  g_eventQueue_ptr->scheduleEvent(this, 1);
}

SyntheticDriver::~SyntheticDriver()
{
  for (int i=0; i<m_last_progress_vector.size(); i++) {
    delete m_request_generator_vector[i];
  }
}

void SyntheticDriver::hitCallback(NodeID proc, SubBlock& data, CacheRequestType type, int thread)
{
  DEBUG_EXPR(TESTER_COMP, MedPrio, data);
  //cout << "  " << proc << " in S.D. hitCallback" << endl;
  if(XACT_MEMORY){
    //XactRequestGenerator* reqGen = static_cast<XactRequestGenerator*>(m_request_generator_vector[proc]);
    //reqGen->performCallback(proc, data);
  } else {
    m_request_generator_vector[proc]->performCallback(proc, data);
  }

  // Mark that we made progress
  m_last_progress_vector[proc] = g_eventQueue_ptr->getTime();
}

void SyntheticDriver::abortCallback(NodeID proc, SubBlock& data, CacheRequestType type, int thread)
{
  //cout << "SyntheticDriver::abortCallback" << endl;
  DEBUG_EXPR(TESTER_COMP, MedPrio, data);

  if(XACT_MEMORY){
    //XactRequestGenerator* reqGen = static_cast<XactRequestGenerator*>(m_request_generator_vector[proc]);
    //reqGen->abortTransaction();
    //reqGen->performCallback(proc, data);
  } else {
    m_request_generator_vector[proc]->performCallback(proc, data);
  }

  // Mark that we made progress
  m_last_progress_vector[proc] = g_eventQueue_ptr->getTime();
}

// For Transactional Memory
/*
// called whenever we send a nack
void SyntheticDriver::notifySendNack( int proc, const Address & addr, uint64 remote_timestamp, const MachineID & remote_id ){
  if(XACT_MEMORY){
    //XactRequestGenerator* reqGen = static_cast<XactRequestGenerator*>(m_request_generator_vector[proc]);
    //reqGen->notifySendNack(addr, remote_timestamp, remote_id);
  }
  else{
    cout << "notifySendNack NOT USING TM" << endl;
    ASSERT(0);
  }
}

// called whenever we receive a NACK
//  Either for a demand request or log store
void SyntheticDriver::notifyReceiveNack( int proc, const Address & addr, uint64 remote_timestamp, const MachineID & remote_id ){
  if(XACT_MEMORY){
    //XactRequestGenerator* reqGen = static_cast<XactRequestGenerator*>(m_request_generator_vector[proc]);
    //reqGen->notifyReceiveNack(addr, remote_timestamp, remote_id);
  }
  else{
    cout << "notifyReceiveNack NOT USING TM" << endl;
    ASSERT(0);
  }
}

// called whenever we received ALL the NACKs. Take abort or retry action here
void SyntheticDriver::notifyReceiveNackFinal(int proc, const Address & addr){
  if(XACT_MEMORY){
    //XactRequestGenerator* reqGen = static_cast<XactRequestGenerator*>(m_request_generator_vector[proc]);
    //reqGen->notifyReceiveNackFinal(addr);
  }
  else{
    cout << "notifyReceiveNackFinal NOT USING TM" << endl;
    ASSERT(0);
  }
}

// called during abort handling
// void SyntheticDriver::notifyAbortStart( const Address & handlerPC ){

// }

// void SyntheticDriver::notifyAbortComplete( const Address & newPC ){

// }
*/

Address SyntheticDriver::pickAddress(NodeID node)
{
  // This methods picks a random lock that we were NOT that last
  // processor to acquire.  Why?  Without this change 2 and 4
  // processor runs, the odds of having the lock in your cache in
  // read/write state is 50% or 25%, respectively.  This effect can
  // make our 'throughput per processor' results look too strange.

  Address addr;
  // FIXME - make this a parameter of the workload
  bool done = false;
  int lock_number = 0;
  int counter = 0;
  while (1) {
    // Pick a random lock
    lock_number = random() % m_lock_vector.size();

    // Were we the last to acquire the lock?
    if (m_lock_vector[lock_number] != node) {
      break;
    }

    // Don't keep trying forever, since if there is only one lock, we're always the last to try to obtain the lock
    counter++;
    if (counter > 10) {
      break;
    }
  }

  // We're going to acquire it soon, so we can update the last
  // processor to hold the lock at this time
  m_lock_vector[lock_number] = node;

  // One lock per cache line
  addr.setAddress(lock_number * RubyConfig::dataBlockBytes());
  return addr;
}

void SyntheticDriver::reportDone()
{
  m_done_counter++;
  if (m_done_counter == RubyConfig::numberOfProcessors()) {
    m_finish_time = g_eventQueue_ptr->getTime();
  }
}

void SyntheticDriver::recordTestLatency(Time time)
{
  m_test_latency.add(time);
}

void SyntheticDriver::recordSwapLatency(Time time)
{
  m_swap_latency.add(time);
}

void SyntheticDriver::recordReleaseLatency(Time time)
{
  m_release_latency.add(time);
}

void SyntheticDriver::wakeup()
{
  //  checkForDeadlock();
  if (m_done_counter < RubyConfig::numberOfProcessors()) {
    g_eventQueue_ptr->scheduleEvent(this, g_DEADLOCK_THRESHOLD);
  }
}

void SyntheticDriver::checkForDeadlock()
{
  int size = m_last_progress_vector.size();
  Time current_time = g_eventQueue_ptr->getTime();
  for (int processor=0; processor<size; processor++) {
    if ((current_time - m_last_progress_vector[processor]) > g_DEADLOCK_THRESHOLD) {
      WARN_EXPR(processor);
      Sequencer* seq_ptr = g_system_ptr->getChip(processor/RubyConfig::numberOfProcsPerChip())->getSequencer(processor%RubyConfig::numberOfProcsPerChip());
      assert(seq_ptr != NULL);
      //     if (seq_ptr->isRequestPending()) {
      //       WARN_EXPR(seq_ptr->pendingAddress());
      //      }
      WARN_EXPR(current_time);
      WARN_EXPR(m_last_progress_vector[processor]);
      WARN_EXPR(current_time - m_last_progress_vector[processor]);
      ERROR_MSG("Deadlock detected.");
    }
  }
}

integer_t  SyntheticDriver::readPhysicalMemory(int procID, physical_address_t address,
                                               int len ){
  char buffer[8];
  ASSERT(len <= 8);
  Sequencer* seq = g_system_ptr->getChip(procID/RubyConfig::numberOfProcsPerChip())->getSequencer(procID%RubyConfig::numberOfProcsPerChip());
  assert(seq != NULL);
  bool found = seq->getRubyMemoryValue(Address(address), buffer, len );
  ASSERT(found);
  return *((integer_t *) buffer);
}

void  SyntheticDriver::writePhysicalMemory( int procID, physical_address_t address,
                                            integer_t value, int len ){
  char buffer[8];
  ASSERT(len <= 8);

  memcpy(buffer, (const void*) &value, len);
  DEBUG_EXPR(TESTER_COMP, MedPrio, "");
  Sequencer* seq = g_system_ptr->getChip(procID/RubyConfig::numberOfProcsPerChip())->getSequencer(procID%RubyConfig::numberOfProcsPerChip());
  assert(seq != NULL);
  bool found = seq->setRubyMemoryValue(Address(address), buffer, len );
  ASSERT(found);
  //return found;
}

void SyntheticDriver::printStats(ostream& out) const
{
  out << endl;
  out << "SyntheticDriver Stats" << endl;
  out << "---------------------" << endl;

  out << "synthetic_finish_time: " << m_finish_time << endl;
  out << "test_latency: " << m_test_latency << endl;
  out << "swap_latency: " << m_swap_latency << endl;
  out << "release_latency: " << m_release_latency << endl;
}

void SyntheticDriver::print(ostream& out) const
{
}