OpenRAM/compiler/modules/bitcell_array.py

227 lines
9.0 KiB
Python

# See LICENSE for licensing information.
#
# Copyright (c) 2016-2019 Regents of the University of California and The Board
# of Regents for the Oklahoma Agricultural and Mechanical College
# (acting for and on behalf of Oklahoma State University)
# All rights reserved.
#
import debug
import design
from tech import drc, spice
from vector import vector
from globals import OPTS
from sram_factory import factory
import logical_effort
class bitcell_array(design.design):
"""
Creates a rows x cols array of memory cells. Assumes bit-lines
and word line is connected by abutment.
Connects the word lines and bit lines.
"""
def __init__(self, cols, rows, name):
design.design.__init__(self, name)
debug.info(1, "Creating {0} {1} x {2}".format(self.name, rows, cols))
self.add_comment("rows: {0} cols: {1}".format(rows, cols))
self.column_size = cols
self.row_size = rows
self.create_netlist()
if not OPTS.netlist_only:
self.create_layout()
# We don't offset this because we need to align
# the replica bitcell in the control logic
#self.offset_all_coordinates()
def create_netlist(self):
""" Create and connect the netlist """
self.add_modules()
self.add_pins()
self.create_instances()
def create_layout(self):
# We increase it by a well enclosure so the precharges don't overlap our wells
self.height = self.row_size*self.cell.height
self.width = self.column_size*self.cell.width
xoffset = 0.0
for col in range(self.column_size):
yoffset = 0.0
for row in range(self.row_size):
name = "bit_r{0}_c{1}".format(row, col)
if row % 2:
tempy = yoffset + self.cell.height
dir_key = "MX"
else:
tempy = yoffset
dir_key = ""
self.cell_inst[row,col].place(offset=[xoffset, tempy],
mirror=dir_key)
yoffset += self.cell.height
xoffset += self.cell.width
self.add_layout_pins()
self.add_boundary()
self.DRC_LVS()
def add_pins(self):
row_list = self.cell.list_all_wl_names()
column_list = self.cell.list_all_bitline_names()
for col in range(self.column_size):
for cell_column in column_list:
self.add_pin(cell_column+"_{0}".format(col))
for row in range(self.row_size):
for cell_row in row_list:
self.add_pin(cell_row+"_{0}".format(row))
self.add_pin("vdd")
self.add_pin("gnd")
def add_modules(self):
""" Add the modules used in this design """
self.cell = factory.create(module_type="bitcell")
self.add_mod(self.cell)
def list_bitcell_pins(self, col, row):
""" Creates a list of connections in the bitcell,
indexed by column and row, for instance use in bitcell_array """
bitcell_pins = []
pin_names = self.cell.list_all_bitline_names()
for pin in pin_names:
bitcell_pins.append(pin+"_{0}".format(col))
pin_names = self.cell.list_all_wl_names()
for pin in pin_names:
bitcell_pins.append(pin+"_{0}".format(row))
bitcell_pins.append("vdd")
bitcell_pins.append("gnd")
return bitcell_pins
def create_instances(self):
""" Create the module instances used in this design """
self.cell_inst = {}
for col in range(self.column_size):
for row in range(self.row_size):
name = "bit_r{0}_c{1}".format(row, col)
self.cell_inst[row,col]=self.add_inst(name=name,
mod=self.cell)
self.connect_inst(self.list_bitcell_pins(col, row))
def add_layout_pins(self):
""" Add the layout pins """
row_list = self.cell.list_all_wl_names()
column_list = self.cell.list_all_bitline_names()
for col in range(self.column_size):
for cell_column in column_list:
bl_pin = self.cell_inst[0,col].get_pin(cell_column)
self.add_layout_pin(text=cell_column+"_{0}".format(col),
layer=bl_pin.layer,
offset=bl_pin.ll().scale(1,0),
width=bl_pin.width(),
height=self.height)
for row in range(self.row_size):
for cell_row in row_list:
wl_pin = self.cell_inst[row,0].get_pin(cell_row)
self.add_layout_pin(text=cell_row+"_{0}".format(row),
layer=wl_pin.layer,
offset=wl_pin.ll().scale(0,1),
width=self.width,
height=wl_pin.height())
# For every second row and column, add a via for gnd and vdd
for row in range(self.row_size):
for col in range(self.column_size):
inst = self.cell_inst[row,col]
for pin_name in ["vdd", "gnd"]:
for pin in inst.get_pins(pin_name):
self.add_power_pin(name=pin_name, loc=pin.center(), vertical=True, start_layer=pin.layer)
def analytical_delay(self, corner, slew, load):
"""Returns relative delay of the bitline in the bitcell array"""
from tech import parameter
#The load being driven/drained is mostly the bitline but could include the sense amp or the column mux.
#The load from the bitlines is due to the drain capacitances from all the other bitlines and wire parasitics.
drain_load = logical_effort.convert_farad_to_relative_c(parameter['bitcell_drain_cap'])
wire_unit_load = .05 * drain_load #Wires add 5% to this.
bitline_load = (drain_load+wire_unit_load)*self.row_size
return [self.cell.analytical_delay(corner, slew, load+bitline_load)]
def analytical_power(self, corner, load):
"""Power of Bitcell array and bitline in nW."""
from tech import drc, parameter
# Dynamic Power from Bitline
bl_wire = self.gen_bl_wire()
cell_load = 2 * bl_wire.return_input_cap()
bl_swing = OPTS.rbl_delay_percentage
freq = spice["default_event_rate"]
bitline_dynamic = self.calc_dynamic_power(corner, cell_load, freq, swing=bl_swing)
#Calculate the bitcell power which currently only includes leakage
cell_power = self.cell.analytical_power(corner, load)
#Leakage power grows with entire array and bitlines.
total_power = self.return_power(cell_power.dynamic + bitline_dynamic * self.column_size,
cell_power.leakage * self.column_size * self.row_size)
return total_power
def gen_wl_wire(self):
if OPTS.netlist_only:
width = 0
else:
width = self.width
wl_wire = self.generate_rc_net(int(self.column_size), width, drc("minwidth_metal1"))
wl_wire.wire_c = 2*spice["min_tx_gate_c"] + wl_wire.wire_c # 2 access tx gate per cell
return wl_wire
def gen_bl_wire(self):
if OPTS.netlist_only:
height = 0
else:
height = self.height
bl_pos = 0
bl_wire = self.generate_rc_net(int(self.row_size-bl_pos), height, drc("minwidth_metal1"))
bl_wire.wire_c =spice["min_tx_drain_c"] + bl_wire.wire_c # 1 access tx d/s per cell
return bl_wire
def output_load(self, bl_pos=0):
bl_wire = self.gen_bl_wire()
return bl_wire.wire_c # sense amp only need to charge small portion of the bl
# set as one segment for now
def input_load(self):
wl_wire = self.gen_wl_wire()
return wl_wire.return_input_cap()
def get_wordline_cin(self):
"""Get the relative input capacitance from the wordline connections in all the bitcell"""
#A single wordline is connected to all the bitcells in a single row meaning the capacitance depends on the # of columns
bitcell_wl_cin = self.cell.get_wl_cin()
total_cin = bitcell_wl_cin * self.column_size
return total_cin
def graph_exclude_bits(self, targ_row, targ_col):
"""Excludes bits in column from being added to graph except target"""
#Function is not robust with column mux configurations
for row in range(self.row_size):
for col in range(self.column_size):
if row == targ_row and col == targ_col:
continue
self.graph_inst_exclude.add(self.cell_inst[row,col])
def get_cell_name(self, inst_name, row, col):
"""Gets the spice name of the target bitcell."""
return inst_name+'.x'+self.cell_inst[row,col].name, self.cell_inst[row,col]