gem5/configs/nvm/sweep_hybrid.py

# Copyright (c) 2020 ARM Limited
# All rights reserved.
#
# The license below extends only to copyright in the software and shall
# not be construed as granting a license to any other intellectual
# property including but not limited to intellectual property relating
# to a hardware implementation of the functionality of the software
# licensed hereunder.  You may use the software subject to the license
# terms below provided that you ensure that this notice is replicated
# unmodified and in its entirety in all distributions of the software,
# modified or unmodified, in source code or in binary form.
#
# 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.

import math
import argparse

import m5
from m5.objects import *
from m5.util import addToPath
from m5.stats import periodicStatDump

addToPath('../')

from common import ObjectList
from common import MemConfig

# this script is helpful to sweep the efficiency of a specific memory
# controller configuration, by varying the number of banks accessed,
# and the sequential stride size (how many bytes per activate), and
# observe what bus utilisation (bandwidth) is achieved

parser = argparse.ArgumentParser()

hybrid_generators = {
    "HYBRID" : lambda x: x.createHybrid,
}

# Use a single-channel DDR3-1600 x64 (8x8 topology) by default
parser.add_argument("--nvm-type", default="NVM_2400_1x64",
                    choices=ObjectList.mem_list.get_names(),
                    help = "type of memory to use")

parser.add_argument("--mem-type", default="DDR4_2400_16x4",
                    choices=ObjectList.mem_list.get_names(),
                    help = "type of memory to use")

parser.add_argument("--nvm-ranks", "-n", type=int, default=1,
                    help = "Number of ranks to iterate across")

parser.add_argument("--mem-ranks", "-r", type=int, default=2,
                    help = "Number of ranks to iterate across")

parser.add_argument("--rd-perc", type=int, default=100,
                    help = "Percentage of read commands")

parser.add_argument("--nvm-perc", type=int, default=100,
                    help = "Percentage of NVM commands")

parser.add_argument("--mode", default="HYBRID",
                    choices=hybrid_generators.keys(),
                    help = "Hybrid: Random DRAM + NVM traffic")

parser.add_argument("--addr-map",
                    choices=ObjectList.dram_addr_map_list.get_names(),
                    default="RoRaBaCoCh", help = "NVM address map policy")

args = parser.parse_args()

# at the moment we stay with the default open-adaptive page policy,
# and address mapping

# start with the system itself, using a multi-layer 2.0 GHz
# crossbar, delivering 64 bytes / 3 cycles (one header cycle)
# which amounts to 42.7 GByte/s per layer and thus per port
system = System(membus = IOXBar(width = 32))
system.clk_domain = SrcClockDomain(clock = '2.0GHz',
                                   voltage_domain =
                                   VoltageDomain(voltage = '1V'))

# set 2 ranges, the first, smaller range for DDR
# the second, larger (1024) range for NVM
# the NVM range starts directly after the DRAM range
system.mem_ranges = [AddrRange('128MB'),
                     AddrRange(Addr('128MB'), size ='1024MB')]

# do not worry about reserving space for the backing store
system.mmap_using_noreserve = True

# force a single channel to match the assumptions in the DRAM traffic
# generator
args.mem_channels = 1
args.external_memory_system = 0
args.hybrid_channel = True
MemConfig.config_mem(args, system)

# the following assumes that we are using the native controller
# with NVM and DRAM interfaces, check to be sure
if not isinstance(system.mem_ctrls[0], m5.objects.HeteroMemCtrl):
    fatal("This script assumes the controller is a HeteroMemCtrl subclass")
if not isinstance(system.mem_ctrls[0].dram, m5.objects.DRAMInterface):
    fatal("This script assumes the first memory is a DRAMInterface subclass")
if not isinstance(system.mem_ctrls[0].nvm, m5.objects.NVMInterface):
    fatal("This script assumes the second memory is a NVMInterface subclass")

# there is no point slowing things down by saving any data
system.mem_ctrls[0].dram.null = True
system.mem_ctrls[0].nvm.null = True

# Set the address mapping based on input argument
system.mem_ctrls[0].dram.addr_mapping = args.addr_map
system.mem_ctrls[0].nvm.addr_mapping = args.addr_map

# stay in each state for 0.25 ms, long enough to warm things up, and
# short enough to avoid hitting a refresh
period = 250000000

# stay in each state as long as the dump/reset period, use the entire
# range, issue transactions of the right burst size, and match
# the maximum bandwidth to ensure that it is saturated

# get the number of banks
nbr_banks_dram = system.mem_ctrls[0].dram.banks_per_rank.value

# determine the burst length in bytes
burst_size_dram = int((system.mem_ctrls[0].dram.devices_per_rank.value *
                  system.mem_ctrls[0].dram.device_bus_width.value *
                  system.mem_ctrls[0].dram.burst_length.value) / 8)

# next, get the page size in bytes
page_size_dram = system.mem_ctrls[0].dram.devices_per_rank.value * \
    system.mem_ctrls[0].dram.device_rowbuffer_size.value

# get the number of regions
nbr_banks_nvm = system.mem_ctrls[0].nvm.banks_per_rank.value

# determine the burst length in bytes
burst_size_nvm = int((system.mem_ctrls[0].nvm.devices_per_rank.value *
                  system.mem_ctrls[0].nvm.device_bus_width.value *
                  system.mem_ctrls[0].nvm.burst_length.value) / 8)


burst_size = max(burst_size_dram, burst_size_nvm)

# next, get the page size in bytes
buffer_size_nvm = system.mem_ctrls[0].nvm.devices_per_rank.value * \
    system.mem_ctrls[0].nvm.device_rowbuffer_size.value

# match the maximum bandwidth of the memory, the parameter is in seconds
# and we need it in ticks (ps)
itt = min(system.mem_ctrls[0].dram.tBURST.value,
          system.mem_ctrls[0].nvm.tBURST.value) * 1000000000000

# assume we start at 0 for DRAM
max_addr_dram = system.mem_ranges[0].end
min_addr_nvm = system.mem_ranges[1].start
max_addr_nvm = system.mem_ranges[1].end

# use min of the page size and 512 bytes as that should be more than
# enough
max_stride = min(256, buffer_size_nvm, page_size_dram)

# create a traffic generator, and point it to the file we just created
system.tgen = PyTrafficGen()

# add a communication monitor
system.monitor = CommMonitor()

# connect the traffic generator to the bus via a communication monitor
system.tgen.port = system.monitor.cpu_side_port
system.monitor.mem_side_port = system.membus.cpu_side_ports

# connect the system port even if it is not used in this example
system.system_port = system.membus.cpu_side_ports

# every period, dump and reset all stats
periodicStatDump(period)

# run Forrest, run!
root = Root(full_system = False, system = system)
root.system.mem_mode = 'timing'

m5.instantiate()

def trace():
    addr_map = ObjectList.dram_addr_map_list.get(args.addr_map)
    generator = hybrid_generators[args.mode](system.tgen)
    for stride_size in range(burst_size, max_stride + 1, burst_size):
        num_seq_pkts_dram = int(math.ceil(float(stride_size) /
                                          burst_size_dram))
        num_seq_pkts_nvm = int(math.ceil(float(stride_size) / burst_size_nvm))
        yield generator(period,
                        0, max_addr_dram, burst_size_dram,
                        min_addr_nvm, max_addr_nvm, burst_size_nvm,
                        int(itt), int(itt),
                        args.rd_perc, 0,
                        num_seq_pkts_dram, page_size_dram,
                        nbr_banks_dram, nbr_banks_dram,
                        num_seq_pkts_nvm, buffer_size_nvm,
                        nbr_banks_nvm, nbr_banks_nvm,
                        addr_map, args.mem_ranks,
                        args.nvm_ranks, args.nvm_perc)

    yield system.tgen.createExit(0)

system.tgen.start(trace())

m5.simulate()

print("Hybrid DRAM + NVM sweep with max_stride: %d" % (max_stride))
print("NVM burst: %d, NVM banks: %d" % (burst_size_nvm, nbr_banks_nvm))
print("DRAM burst: %d, DRAM banks: %d" % (burst_size_dram, nbr_banks_dram))