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
|
#ifndef GyWue2EL4TyiAsNv0XegfFL4jk
#define GyWue2EL4TyiAsNv0XegfFL4jk
#include <map>
#include "sim_replay.hpp"
// discrete property Egal
struct Egal {
typedef uint8_t type;
typedef SpikeArrival quant;
typedef SpikeArrival::instance_ptr_t instance_ptr_t;
static const uint32_t size = 2;
static const char* const name;
};
const char* const Egal::name = "Egal";
namespace SimCausalityImpl {
using namespace boost::mpl;
using namespace boost;
namespace MC = ModelConsts;
template<typename PropList> struct SimCausality;
template<typename ReplayQuant, typename Quant>
struct MaybeTrackEvent {
template<typename Sim>
void operator() (Sim &sim) {
// we handle every event but only track those affecting neurons
auto &sg = static_cast<SimCausality<typename Sim::PropComp::properties_t>&>(sim);
if (is_same<Quant, SpikeArrival>::value) {
sg.handleSA();
}else if (is_same<Quant, RandomSpike>::value) {
sg.handleRand();
}else{
sim.template handleEvent<Quant>();
}
}
typedef MaybeTrackEvent<ReplayQuant, Quant> impl;
};
template<typename PropList>
struct SimCausality : public SimReplay<PropList, Neuron, MaybeTrackEvent> {
typedef SimReplay<PropList, Neuron, MaybeTrackEvent> Super;
using Super::queues;
using Super::pc;
using Super::ct;
using Super::handleEvent;
SimCausality(Time start)
: Super(IdList<uint16_t>((char*) "0-999"), start),
totalMarks{},
trueMarks{},
inactiveNeurons(1000),
hasFired{}
{
if (start == Time(0)) {
// special case: at t=0 all neurons are have zero membrane
// voltage and can be used for analysis before firing the first
// time
inactiveNeurons = 0;
for (int i=0; i<1000; i++)
hasFired[i] = true;
}
}
bool run(const Time until, uint64_t maxEvents) __attribute__((noinline)) {
return Super::run(until, maxEvents);
}
void handleRand() {
Ptr<Neuron> neuron(queues.get<RandomSpike>().minPayload().dst);
Ptr<SpikeArrival> fakeSA(Index<SpikeArrival>::nil());
handleDiff(neuron, fakeSA, MC::RandomSpikeWeight,
&SimCausality<PropList>::template handleEvent<RandomSpike>);
}
void handleSA() {
auto &event(queues.get<SpikeArrival>().minPayload());
Ptr<SpikeArrival> sa (event.src.get<0>());
Ptr<Synapse> syn (event.src.get<1>());
Ptr<Neuron> neuron(event.dst);
PLA_Get<Weight> getWeight (ct, syn);
Weight::type weight(pc.call(getWeight));
handleDiff(neuron, sa, weight,
&SimCausality<PropList>::template handleEvent<SpikeArrival>);
}
void handleDiff(Ptr<Neuron> neuron, Ptr<SpikeArrival> sa, Voltage::type diff,
bool (SimCausality<PropList>::*handleEvent)()) {
PLA_Get<Voltage> getVoltage (ct, neuron);
PLA_Get<IPCoeff1> getIPCoeff1(ct, neuron);
Voltage::type preEventVoltage(pc.call(getVoltage));
IPCoeff1::type ipCoeff1 (pc.call(getIPCoeff1));
bool evokedSpike = (this->*handleEvent)();
auto &openExc(this->openExc[neuron()]);
auto &openInh(this->openInh[neuron()]);
if (!hasFired[neuron()]) {
if (evokedSpike) {
hasFired[neuron()] = true;
inactiveNeurons--;
if (!inactiveNeurons)
lastNeuronActivation = ct;
}
return;
}
if (evokedSpike) {
// check exc. list against voltage diff
Time wntGap = ct + wntNorm(diff + preEventVoltage
- MC::FireThreshold - ipCoeff1);
for (auto i = openExc.lower_bound(wntGap);
i != openExc.end();) {
mark(i->second, 0);
openExc.erase(i++);
}
// mark own spike
mark(sa, 0);
// mark remainig inh/exc spikes as q>0, clear buffers
for (auto i : openExc) mark(i.second, 1); openExc.clear();
for (auto i : openInh) mark(i.second, 1); openInh.clear();
}else{
if (diff < 0) {
// add to inh. list
Time wnt = ct + wntNorm(- diff);
openInh.insert(std::make_pair(wnt, sa));
}else{
// add to exc. list
Time wntSelf = ct + wntNorm(diff);
openExc.insert(std::make_pair(wntSelf, sa));
// check inh. list against voltage gap
PLA_Get<Voltage> pla_get(ct, neuron);
Voltage::type voltage = pc.call(pla_get);
Time wntGap = ct + wntNorm(MC::FireThreshold + ipCoeff1 - voltage);
for (auto i = openInh.lower_bound(wntGap);
i != openInh.end();) {
mark(i->second, 0);
openInh.erase(i++);
}
}
}
}
void mark(Ptr<SpikeArrival> ptr, bool m) {
bool isSA = ptr() != Index<SpikeArrival>::nil();
totalMarks[isSA]++;
trueMarks[isSA] += !m;
if (isSA)
egalPC.cast(PLA_Set<Egal>(ptr, 1+m)); // unset values are encoded as zero
}
Time wntNorm(Voltage::type diff) {
return MC::Tau_Voltage * log(diff);
}
PropertyComposition<boost::mpl::list<
boost::mpl::pair<Egal, boost::mpl::bool_<true>>
>> egalPC;
uint64_t totalMarks[2],
trueMarks[2];
uint16_t inactiveNeurons;
Time lastNeuronActivation;
bool hasFired[maxNeurons];
std::multimap<Time, Ptr<SpikeArrival>> openInh[maxNeurons], openExc[maxNeurons];
};
} // NS
using SimCausalityImpl::SimCausality;
#endif // GyWue2EL4TyiAsNv0XegfFL4jk
|
