1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
|
# More testing BBSaveState.
from neuron import h
pc = h.ParallelContext()
cvode = h.CVode()
class Model1:
def __init__(self, ncell, n):
# gid -> [n NetCons] -> gid+1
cells = {gid: h.Follower() for gid in range(pc.id(), ncell, pc.nhost())}
for gid in cells:
pc.set_gid2node(gid, pc.id())
pc.cell(gid, h.NetCon(cells[gid], None))
netcons = {}
for gid in cells:
netcons[gid] = [
pc.gid_connect(gid - 1, cells[gid]) for _ in range(n) if gid > 0
]
# Stimulate gid 0
if 0 in cells:
# stim cell[0] with fast burst of 5 spikes
ns = h.NetStim()
ns.start = 0.9999
ns.number = 5
ns.interval = 0.1
nsnc = h.NetCon(ns, cells[0])
nsnc.delay = 0
nsnc.weight[0] = 0
for gid, cell in cells.items():
# cells generate spike on every input of weight > 1
# cell.refrac = 0
for i, nc in enumerate(netcons[gid]):
nc.weight[0] = 0.001 * i + h.dt
nc.delay = (1.0 + 0.01 * i) * h.dt
self.cells = cells
self.netcons = netcons
self.ns = ns
self.nsnc = nsnc
class Model2:
"""NetStim -> Cell -> N Cells -> Cell"""
# exercise PreSyn on queue with use_min_delay_ by means of
# several NetCon with same PreSyn and same delay
def __init__(self, n):
self.ncell_per_layer = [1, n, 2]
self.cells = {}
# make cells and associate with gid (round robin distribution)
for ilayer, ncell in enumerate(self.ncell_per_layer):
for icell in range(ncell):
gid = self.info2gid(ilayer, icell)
if (gid % pc.nhost()) == pc.id():
self.cells[gid] = h.Follower()
pc.set_gid2node(gid, pc.id())
pc.cell(gid, h.NetCon(self.cells[gid], None))
# make connections (all to all from layer i-1 to layer i)
self.netcons = {}
for gid, cell in self.cells.items():
ilayer, icell = self.gid2info(gid)
if ilayer == 0:
continue
srclayer = ilayer - 1
for isrc in range(self.ncell_per_layer[srclayer]):
srcgid = self.info2gid(srclayer, isrc)
nc = pc.gid_connect(srcgid, cell)
nc.weight[0] = 0.1
nc.delay = 1.0
self.netcons[(srcgid, gid)] = nc
# Stimulate layer 0 with NetStim
self.netstims = {}
self.netstim_con = {}
for gid in range(self.ncell_per_layer[0]):
if gid in self.cells:
ns = h.NetStim()
self.netstims[gid] = ns
ns.start = 0.9999
ns.number = 5
ns.interval = 0.1
nc = h.NetCon(ns, self.cells[gid])
nc.delay = 0.0
nc.weight[0] = 0.1
self.netstim_con[gid] = nc
self.ns = self.netstims[0]
self.nsnc = self.netstim_con[0]
print(self.cells)
print(self.netcons)
print(self.netstim_con)
def info2gid(self, ilayer, icell):
return ilayer * 100 + icell
def gid2info(self, gid):
return (int(gid / 100)), gid % 100
def test_bbss():
print("focus on BinQ initialization Issue #1444")
ncell = 3
n = 5
model = Model2(5)
# pc.gid_clear()
# model = Model1(ncell, n)
spiketime = h.Vector()
spikegid = h.Vector()
pc.spike_record(-1, spiketime, spikegid)
def run(tstop):
pc.set_maxstep(10)
h.finitialize()
pc.psolve(tstop)
spiketime_std = spiketime.c()
spikegid_std = spikegid.c()
def set_stdspikes():
spiketime_std.copy(spiketime)
spikegid_std.copy(spikegid)
tstop = 5
run(tstop)
set_stdspikes()
def prspikes():
print("prspikes")
for i, gid in enumerate(spikegid):
print("%g %d" % (spiketime[i], gid))
def compare_spikes():
x = list(zip(spiketime_std, spikegid_std))
x = sorted(x, key=lambda e: e[1])
y = list(zip(spiketime, spikegid))
y = sorted(y, key=lambda e: e[1])
if len(x) != len(y):
print(len(x), len(y))
# assert len(x) == len(y)
if x != y:
q = (x, y) if len(x) <= len(y) else (y, x)
for i, a in enumerate(q[0]):
b = q[1][i]
if a != b:
# assert a[1] == b[1]
z = abs(a[0] - b[0])
if z >= 1e-9:
print(x[i], y[i])
assert z < 1e-9
def srun(tsave, tstop):
qm = cvode.queue_mode()
print(
"srun tsave=%g tstop=%g start=%g delay=%g %s"
% (tsave, tstop, model.ns.start, model.nsnc.delay, "binq" if qm else "")
)
run(tsave)
st = spiketime.c()
sg = spikegid.c()
bbss = h.BBSaveState()
bbss.save("allcell_bbss.dat")
qm = 1
if qm:
print("after save")
cvode.print_event_queue()
h.finitialize(-65)
spiketime.resize(0)
spikegid.resize(0)
bbss.restore("allcell_bbss.dat")
if qm:
print("after restore")
cvode.print_event_queue()
pc.psolve(tstop)
st.append(spiketime)
sg.append(spikegid)
spiketime.copy(st)
spikegid.copy(sg)
print(" %d spikes" % len(spiketime))
srun(1.1, tstop)
compare_spikes()
cvode.queue_mode(1)
run(tstop)
assert len(spiketime_std) == len(spiketime)
spiketime_std = spiketime.c()
spikegid_std = spikegid.c()
srun(1.1, tstop)
compare_spikes()
for binq in [0, 1]:
cvode.queue_mode(binq)
parms = [
(1, 0),
(1e-10, 0),
(0, 0),
(0.25 * h.dt, 0),
(0.5 * h.dt, 0),
(0.75 * h.dt, 0),
]
for parm in parms:
if 0 in model.cells:
model.ns.start, model.nsnc.delay = parm
for tsave in [0.0, h.dt]:
run(tstop)
set_stdspikes()
srun(tsave, tstop)
compare_spikes()
cvode.queue_mode(0)
pc.gid_clear()
if __name__ == "__main__":
test_bbss()
|
