diff --git a/compiler/modules/hierarchical_decoder.py b/compiler/modules/hierarchical_decoder.py
index 0e70db55..4648519a 100644
--- a/compiler/modules/hierarchical_decoder.py
+++ b/compiler/modules/hierarchical_decoder.py
@@ -10,6 +10,7 @@ import debug
 import design
 from math import log
 from math import sqrt
+from math import ceil
 import math
 import contact
 from sram_factory import factory
@@ -31,7 +32,7 @@ class hierarchical_decoder(design.design):
 
         self.cell_height = height        
         self.rows = rows
-        self.num_inputs = int(math.log(self.rows, 2))
+        self.num_inputs = math.ceil(math.log(self.rows, 2))
         (self.no_of_pre2x4,self.no_of_pre3x8)=self.determine_predecodes(self.num_inputs)
 
         self.create_netlist()
@@ -338,14 +339,15 @@ class hierarchical_decoder(design.design):
             for i in range(len(self.predec_groups[0])):
                 for j in range(len(self.predec_groups[1])):
                     row = len(self.predec_groups[0])*j + i
-                    name = self.NAND_FORMAT.format(row)
-                    self.nand_inst.append(self.add_inst(name=name,
-                                                        mod=self.nand2))
-                    pins =["out_{0}".format(i),
-                           "out_{0}".format(j + len(self.predec_groups[0])),
-                           "Z_{0}".format(row),
-                           "vdd", "gnd"]
-                    self.connect_inst(pins)
+                    if (row < self.rows):
+                        name = self.NAND_FORMAT.format(row)
+                        self.nand_inst.append(self.add_inst(name=name,
+                                                            mod=self.nand2))
+                        pins =["out_{0}".format(i),
+                               "out_{0}".format(j + len(self.predec_groups[0])),
+                               "Z_{0}".format(row),
+                               "vdd", "gnd"]
+                        self.connect_inst(pins)
 
 
         # Row Decoder NAND GATE array for address inputs >5.
@@ -356,16 +358,17 @@ class hierarchical_decoder(design.design):
                         row = (len(self.predec_groups[0])*len(self.predec_groups[1])) * k \
                             + len(self.predec_groups[0])*j + i
 
-                        name = self.NAND_FORMAT.format(row)
-                        self.nand_inst.append(self.add_inst(name=name,
-                                                            mod=self.nand3))
+                        if (row < self.rows):
+                            name = self.NAND_FORMAT.format(row)
+                            self.nand_inst.append(self.add_inst(name=name,
+                                                                mod=self.nand3))
                         
-                        pins = ["out_{0}".format(i),
-                                "out_{0}".format(j + len(self.predec_groups[0])),
-                                "out_{0}".format(k + len(self.predec_groups[0]) + len(self.predec_groups[1])),
-                                "Z_{0}".format(row),
-                                "vdd", "gnd"]
-                        self.connect_inst(pins)
+                            pins = ["out_{0}".format(i),
+                                    "out_{0}".format(j + len(self.predec_groups[0])),
+                                    "out_{0}".format(k + len(self.predec_groups[0]) + len(self.predec_groups[1])),
+                                    "Z_{0}".format(row),
+                                    "vdd", "gnd"]
+                            self.connect_inst(pins)
 
 
     def create_decoder_inv_array(self):
@@ -527,10 +530,11 @@ class hierarchical_decoder(design.design):
             for index_B in self.predec_groups[1]:
                 for index_A in self.predec_groups[0]:
                     # FIXME: convert to connect_bus?
-                    predecode_name = "predecode_{}".format(index_A)
-                    self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("A"))
-                    predecode_name = "predecode_{}".format(index_B)                    
-                    self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("B"))
+                    if (row_index < self.rows):
+                        predecode_name = "predecode_{}".format(index_A)
+                        self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("A"))
+                        predecode_name = "predecode_{}".format(index_B)                    
+                        self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("B"))
                     row_index = row_index + 1
 
         elif (self.num_inputs > 5):
@@ -538,12 +542,13 @@ class hierarchical_decoder(design.design):
                 for index_B in self.predec_groups[1]:
                     for index_A in self.predec_groups[0]:
                         # FIXME: convert to connect_bus?
-                        predecode_name = "predecode_{}".format(index_A) 
-                        self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("A"))
-                        predecode_name = "predecode_{}".format(index_B) 
-                        self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("B"))
-                        predecode_name = "predecode_{}".format(index_C) 
-                        self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("C"))
+                        if (row_index < self.rows):
+                            predecode_name = "predecode_{}".format(index_A) 
+                            self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("A"))
+                            predecode_name = "predecode_{}".format(index_B) 
+                            self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("B"))
+                            predecode_name = "predecode_{}".format(index_C) 
+                            self.route_predecode_rail(predecode_name, self.nand_inst[row_index].get_pin("C"))
                         row_index = row_index + 1
 
     def route_vdd_gnd(self):
diff --git a/compiler/pgates/pwrite_driver.py b/compiler/pgates/pwrite_driver.py
new file mode 100644
index 00000000..a808737b
--- /dev/null
+++ b/compiler/pgates/pwrite_driver.py
@@ -0,0 +1,304 @@
+# See LICENSE for licensing information.
+#
+#Copyright (c) 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 design
+from tech import drc, parameter, spice
+import debug
+import math
+from tech import drc
+from vector import vector
+from globals import OPTS
+from sram_factory import factory
+
+class pwrite_driver(design.design):
+    """
+    The pwrite_driver is two tristate inverters that drive the bitlines.
+    The data input is first inverted before one tristate. 
+    The inverted enable is also generated to control one tristate.
+    """
+    def __init__(self, name, size=0):
+        debug.error("pwrite_driver not implemented yet.", -1)
+        debug.info(1, "creating pwrite_driver {}".format(name))
+        design.design.__init__(self, name)
+        self.size = size
+        self.beta = parameter["beta"]
+        self.pmos_width = self.beta*self.size*parameter["min_tx_size"]
+        self.nmos_width = self.size*parameter["min_tx_size"]
+
+        # The tech M2 pitch is based on old via orientations
+        self.m2_pitch = self.m2_space + self.m2_width
+
+        # Width is matched to the bitcell,
+        # Height will be variable
+        self.bitcell = factory.create(module_type="bitcell")
+        self.width = self.bitcell.width
+
+        # Creates the netlist and layout
+        # Since it has variable height, it is not a pgate.
+        self.create_netlist()
+        if not OPTS.netlist_only:        
+            self.create_layout()
+            self.DRC_LVS()
+
+
+
+    def create_netlist(self):
+        self.add_pins()
+        self.add_modules()
+        self.create_insts()
+
+    def create_layout(self):
+        self.place_modules()
+        self.route_wires()
+        self.route_supplies()
+
+        
+        
+    def add_pins(self):
+        self.add_pin("din", "INPUT")
+        self.add_pin("bl", "OUTPUT")
+        self.add_pin("br", "OUTPUT")
+        self.add_pin("en", "INPUT")
+        self.add_pin("vdd", "POWER")
+        self.add_pin("gnd", "GROUND")
+
+        
+    def add_modules(self):
+
+        # Tristate inverter
+        self.tri = factory.create(module_type="ptristate_inv", height="min")
+        self.add_mod(self.tri)
+        debug.check(self.tri.width<self.width,"Could not create tristate inverter to match bitcell width")
+
+        #self.tbuf = factory.create(module_type="ptristate_buf", height="min")
+        #self.add_mod(self.tbuf)
+        #debug.check(self.tbuf.width<self.width,"Could not create tristate buffer to match bitcell width")
+        
+        # Inverter for din and en
+        self.inv = factory.create(module_type="pinv", under_rail_vias=True)
+        self.add_mod(self.inv)
+
+    def create_insts(self):
+        # Enable inverter
+        self.en_inst = self.add_inst(name="en_inv", mod=self.inv)
+        self.connect_inst(["en", "en_bar", "vdd", "gnd"])
+
+        # Din inverter
+        self.din_inst = self.add_inst(name="din_inv", mod=self.inv)
+        self.connect_inst(["din", "din_bar", "vdd", "gnd"])
+        
+        # Bitline tristate
+        self.bl_inst = self.add_inst(name="bl_tri", mod=self.tri)
+        self.connect_inst(["din_bar", "bl", "en", "en_bar", "vdd", "gnd"])
+
+        # Bitline bar tristate
+        self.br_inst = self.add_inst(name="br_tri", mod=self.tri)
+        self.connect_inst(["din", "br", "en", "en_bar", "vdd", "gnd"])
+
+
+    def place_modules(self):
+        
+        # Add enable to the right
+        self.din_inst.place(vector(0, 0))
+
+        # Add BR tristate above
+        self.br_inst.place(vector(0, self.en_inst.uy()+self.br_inst.height), mirror="MX")
+
+        # Add BL tristate buffer
+        #print(self.bl_inst.width,self.width)
+        self.bl_inst.place(vector(self.width,self.br_inst.uy()), mirror="MY")
+        
+        # Add din to the left
+        self.en_inst.place(vector(self.width, self.bl_inst.uy()+self.en_inst.height), rotate=180)
+        
+        self.height = self.en_inst.uy()
+        
+        
+    def route_bitlines(self):
+        """ 
+        Route the bitlines to the spacing of the bitcell 
+        ( even though there may be a column mux  ) 
+        """
+
+        # Second from left track and second from right track
+        right_x = self.width + self.m2_pitch
+        left_x = -self.m2_pitch
+        
+        bl_xoffset = left_x
+        bl_out=vector(bl_xoffset, self.height)
+        bl_in=self.bl_inst.get_pin("out").center()
+        self.add_via_center(layers=("metal1","via1","metal2"),
+                            offset=bl_in)
+
+        bl_mid = vector(bl_out.x,bl_in.y)
+        self.add_path("metal2", [bl_in, bl_mid, bl_out])
+        
+        self.add_layout_pin_rect_center(text="bl",
+                                        layer="metal2",
+                                        offset=bl_out)
+        
+        br_xoffset = right_x
+        br_out=vector(br_xoffset, self.height)
+        br_in=self.br_inst.get_pin("out").center()
+        self.add_via_center(layers=("metal1","via1","metal2"),
+                            offset=br_in)
+
+        br_mid = vector(br_out.x,br_in.y)
+        self.add_path("metal2", [br_in, br_mid, br_out])
+        self.add_layout_pin_rect_center(text="br",
+                                        layer="metal2",
+                                        offset=br_out)
+        
+        #br_xoffset = b.get_pin("br".cx()
+        #self.br_inst.get_pin("br")
+
+    def route_din(self):
+        
+        # Left
+        track_xoff = self.get_m2_track(1)
+        
+        din_loc = self.din_inst.get_pin("A").center()
+        self.add_via_stack("metal1", "metal2", din_loc)
+        din_track = vector(track_xoff,din_loc.y)
+        
+        br_in = self.br_inst.get_pin("in").center()
+        self.add_via_stack("metal1", "metal2", br_in)
+        br_track = vector(track_xoff,br_in.y)
+        
+        din_in = vector(track_xoff,0)
+        
+        self.add_path("metal2", [din_in,  din_track, din_loc, din_track, br_track, br_in])
+
+        self.add_layout_pin_rect_center(text="din",
+                                        layer="metal2",
+                                        offset=din_in)
+
+    def route_din_bar(self):
+        
+        # Left
+        track_xoff = self.get_m4_track(self.din_bar_track)
+        
+        din_bar_in = self.din_inst.get_pin("Z").center()
+        self.add_via_stack("metal1", "metal3", din_bar_in)
+        din_bar_track = vector(track_xoff,din_bar_in.y)
+        
+        bl_in = self.bl_inst.get_pin("in").center()
+        self.add_via_stack("metal1", "metal3", bl_in)
+        bl_track = vector(track_xoff,bl_in.y)
+        
+        din_in = vector(track_xoff,0)
+        
+        self.add_wire(("metal3","via3","metal4"), [din_bar_in, din_bar_track, bl_track, bl_in])
+
+        self.add_layout_pin_rect_center(text="din",
+                                        layer="metal4",
+                                        offset=din_in)
+        
+
+    def route_en_bar(self):
+        # Enable in track 
+        track_xoff = self.get_m4_track(self.en_bar_track)
+        
+        # This M2 pitch is a hack since the A and Z pins align horizontally
+        en_bar_loc = self.en_inst.get_pin("Z").uc()
+        en_bar_track = vector(track_xoff, en_bar_loc.y)
+        self.add_via_stack("metal1", "metal3", en_bar_loc)
+        
+        # This is a U route to the right down then left
+        bl_en_loc = self.bl_inst.get_pin("en_bar").center()
+        bl_en_track = vector(track_xoff, bl_en_loc.y)
+        self.add_via_stack("metal1", "metal3", bl_en_loc)
+        br_en_loc = self.br_inst.get_pin("en_bar").center()
+        br_en_track = vector(track_xoff, bl_en_loc.y)
+        self.add_via_stack("metal1", "metal3", br_en_loc)
+
+        
+        # L shape
+        self.add_wire(("metal3","via3","metal4"),
+                      [en_bar_loc, en_bar_track, bl_en_track])
+        # U shape
+        self.add_wire(("metal3","via3","metal4"),
+                      [bl_en_loc, bl_en_track, br_en_track, br_en_loc])
+
+    
+    def route_en(self):
+        
+        # Enable in track 
+        track_xoff = self.get_m4_track(self.en_track)
+
+        # The en pin will be over the vdd rail
+        vdd_yloc = self.en_inst.get_pin("vdd").cy()
+        self.add_layout_pin_segment_center(text="en",
+                                           layer="metal3",
+                                           start=vector(0,vdd_yloc),
+                                           end=vector(self.width,vdd_yloc))
+
+        en_loc = self.en_inst.get_pin("A").center()
+        en_rail = vector(en_loc.x, vdd_yloc)        
+        self.add_via_stack("metal1", "metal2", en_loc)
+        self.add_path("metal2", [en_loc, en_rail])
+        self.add_via_stack("metal2", "metal3", en_rail)
+
+        # Start point in the track on the pin rail
+        en_track = vector(track_xoff, vdd_yloc)
+        self.add_via_stack("metal3", "metal4", en_track)
+        
+        # This is a U route to the right down then left
+        bl_en_loc = self.bl_inst.get_pin("en").center()
+        bl_en_track = vector(track_xoff, bl_en_loc.y)
+        self.add_via_stack("metal1", "metal3", bl_en_loc)
+        br_en_loc = self.br_inst.get_pin("en").center()
+        br_en_track = vector(track_xoff, bl_en_loc.y)
+        self.add_via_stack("metal1", "metal3", br_en_loc)
+
+        # U shape
+        self.add_wire(("metal3","via3","metal4"),
+                      [en_track, bl_en_track, bl_en_loc, bl_en_track, br_en_track, br_en_loc])
+
+
+    def get_m4_track(self,i):
+        return 0.5*self.m4_space + i*(self.m4_width+self.m4_space)
+    def get_m3_track(self,i):
+        return 0.5*self.m3_space + i*(self.m3_width+self.m3_space)
+    def get_m2_track(self,i):
+        return 0.5*self.m2_space + i*(self.m2_width+self.m2_space)
+    
+    def route_wires(self):
+        # M4 tracks
+        self.din_bar_track = 2
+        self.en_track = 0
+        self.en_bar_track = 1
+        
+        self.route_bitlines()
+        self.route_din()
+        self.route_din_bar()        
+        self.route_en()
+        self.route_en_bar()
+             
+    def route_supplies(self):
+        for inst in [self.en_inst, self.din_inst, self.bl_inst, self.br_inst]:
+            # Continous vdd rail along with label.
+            vdd_pin=inst.get_pin("vdd")
+            self.add_layout_pin(text="vdd",
+                                layer="metal1",
+                                offset=vdd_pin.ll().scale(0,1),
+                                width=self.width,
+                                height=vdd_pin.height())
+        
+            # Continous gnd rail along with label.
+            gnd_pin=inst.get_pin("gnd")
+            self.add_layout_pin(text="gnd",
+                                layer="metal1",
+                                offset=gnd_pin.ll().scale(0,1),
+                                width=self.width,
+                                height=vdd_pin.height())
+
+        
+    def get_w_en_cin(self):
+        """Get the relative capacitance of a single input"""
+        # This is approximated from SCMOS. It has roughly 5 3x transistor gates.
+        return 5*3
diff --git a/compiler/tests/04_pwrite_driver_test.py b/compiler/tests/04_pwrite_driver_test.py
new file mode 100644
index 00000000..5e1bf621
--- /dev/null
+++ b/compiler/tests/04_pwrite_driver_test.py
@@ -0,0 +1,36 @@
+#!/usr/bin/env python3
+# See LICENSE for licensing information.
+#
+# Copyright (c) 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 unittest
+from testutils import header, openram_test
+import sys
+import os
+sys.path.append(os.getenv("OPENRAM_HOME"))
+import globals
+from globals import OPTS
+from sram_factory import factory
+import debug
+
+@unittest.skip("SKIPPING 04_pwrite_driver_test")
+class pwrite_driver_test(openram_test):
+
+    def runTest(self):
+        globals.init_openram("config_{0}".format(OPTS.tech_name))
+
+        debug.info(2, "Checking 1x pwrite_driver")
+        tx = factory.create(module_type="pwrite_driver", size=1)
+        self.local_check(tx)
+
+        globals.end_openram()        
+
+# run the test from the command line
+if __name__ == "__main__":
+    (OPTS, args) = globals.parse_args()
+    del sys.argv[1:]
+    header(__file__, OPTS.tech_name)
+    unittest.main()
diff --git a/compiler/tests/06_hierarchical_decoder_test.py b/compiler/tests/06_hierarchical_decoder_test.py
index d4d68017..01d9c0e7 100755
--- a/compiler/tests/06_hierarchical_decoder_test.py
+++ b/compiler/tests/06_hierarchical_decoder_test.py
@@ -35,14 +35,30 @@ class hierarchical_decoder_test(openram_test):
         a = factory.create(module_type="hierarchical_decoder", rows=16)
         self.local_check(a)
 
+        debug.info(1, "Testing 17 row sample for hierarchical_decoder")
+        a = factory.create(module_type="hierarchical_decoder", rows=17)
+        self.local_check(a)
+
+        debug.info(1, "Testing 23 row sample for hierarchical_decoder")
+        a = factory.create(module_type="hierarchical_decoder", rows=23)
+        self.local_check(a)
+
         debug.info(1, "Testing 32 row sample for hierarchical_decoder")
         a = factory.create(module_type="hierarchical_decoder", rows=32)
         self.local_check(a)
 
+        debug.info(1, "Testing 65 row sample for hierarchical_decoder")
+        a = factory.create(module_type="hierarchical_decoder", rows=65)
+        self.local_check(a)
+
         debug.info(1, "Testing 128 row sample for hierarchical_decoder")
         a = factory.create(module_type="hierarchical_decoder", rows=128)
         self.local_check(a)
 
+        debug.info(1, "Testing 341 row sample for hierarchical_decoder")
+        a = factory.create(module_type="hierarchical_decoder", rows=341)
+        self.local_check(a)
+
         debug.info(1, "Testing 512 row sample for hierarchical_decoder")
         a = factory.create(module_type="hierarchical_decoder", rows=512)
         self.local_check(a)
@@ -58,14 +74,34 @@ class hierarchical_decoder_test(openram_test):
         a = factory.create(module_type="hierarchical_decoder", rows=16)
         self.local_check(a)
 
+        factory.reset()
+        debug.info(1, "Testing 17 row sample for hierarchical_decoder (multi-port case)")
+        a = factory.create(module_type="hierarchical_decoder", rows=17)
+        self.local_check(a)
+
+        factory.reset()
+        debug.info(1, "Testing 23 row sample for hierarchical_decoder (multi-port case)")
+        a = factory.create(module_type="hierarchical_decoder", rows=23)
+        self.local_check(a)
+
         debug.info(1, "Testing 32 row sample for hierarchical_decoder (multi-port case)")
         a = factory.create(module_type="hierarchical_decoder", rows=32)
         self.local_check(a)
 
+        factory.reset()
+        debug.info(1, "Testing 65 row sample for hierarchical_decoder (multi-port case)")
+        a = factory.create(module_type="hierarchical_decoder", rows=65)
+        self.local_check(a)
+
         debug.info(1, "Testing 128 row sample for hierarchical_decoder (multi-port case)")
         a = factory.create(module_type="hierarchical_decoder", rows=128)
         self.local_check(a)
 
+        factory.reset()
+        debug.info(1, "Testing 341 row sample for hierarchical_decoder (multi-port case)")
+        a = factory.create(module_type="hierarchical_decoder", rows=341)
+        self.local_check(a)
+
         debug.info(1, "Testing 512 row sample for hierarchical_decoder (multi-port case)")
         a = factory.create(module_type="hierarchical_decoder", rows=512)
         self.local_check(a)