shamilton/GeneticAlgo
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

add in xorshift rand function. begin reworking job queue to use cell entry pointers instead of just cells so results can be posted directly to the relevant entry

Browse Source
This commit is contained in:
Seth Hamilton
2025-08-16 00:41:28 -05:00
parent edda3761d1
commit 3265f045d1
3 changed files with 52 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

47
src/genetic.cpp
View File

@@ -21,17 +21,17 @@ namespace genetic {
template <class T> struct CellEntry {
float score;
T cell;
T *cell;
bool stale;
};
template <class T> struct CrossoverJob {
const ReadonlySpan<T> &parents;
const Span<T> &children_out;
Span<CellEntry<T> *> &parents;
Span<CellEntry<T> *> &children_out;
};
template <class T> struct FitnessJob {
const T &cell;
float &result_out;
CellEntry<T> *cell_entry;
};
template <class T> struct WorkQueue {
@@ -100,6 +100,8 @@ template <class T> struct WorkerThreadArgs {
bool *stop_flag;
};
template <class T> void do_crossover_job(CrossoverJob<T> cj) {}
template <class T> void *worker(void *args) {
WorkerThreadArgs<T> *work_args = (WorkerThreadArgs<T> *)args;
Strategy<T> &strat = work_args->strat;
@@ -107,7 +109,10 @@ template <class T> void *worker(void *args) {
bool *stop_flag = work_args->stop_flag;
auto JobDispatcher = overload{
[strat](FitnessJob<T> fj) { fj.result_out = strat.fitness(fj.cell); },
[strat](FitnessJob<T> fj) {
fj.cell_entry->result_out = strat.fitness(*(fj.cell_entry->cell));
fj.cell_entry->stale = true;
},
[strat](CrossoverJob<T> cj) {
strat.crossover(cj.parents, cj.children_out);
},
@@ -131,21 +136,20 @@ template <class T> void *worker(void *args) {
template <class T> Stats<T> run(Strategy<T> strat) {
Stats<T> stats;
WorkQueue<T> queue = make_work_queue<T>(strat.num_cells);
WorkQueue<T> work_queue = make_work_queue<T>(strat.num_cells);
vector<CellEntry<T>> cells_a, cells_b;
T cells[strat.num_cells];
// Using a vector so I can use the make_heap, push_heap, etc.
vector<CellEntry<T>> cell_queue;
for (int i = 0; i < strat.num_cells; i++) {
T cell = strat.make_default_cell();
cells_a.push_back({0, cell, true});
cells_b.push_back({0, cell, true});
cells[i] = strat.make_default_cell();
cell_queue.push_back({0, &cells[i], true});
}
std::vector<CellEntry<T>> &cur_cells = cells_a;
std::vector<CellEntry<T>> &next_cells = cells_b;
bool stop_flag = false;
WorkerThreadArgs<T> args = {
.strat = strat, .queue = queue, .stop_flag = &stop_flag};
.strat = strat, .queue = work_queue, .stop_flag = &stop_flag};
// spawn worker threads
pthread_t threads[strat.num_threads];
@@ -155,20 +159,27 @@ template <class T> Stats<T> run(Strategy<T> strat) {
for (int i = 0; i < strat.num_generations; i++) {
// generate fitness jobs
if (strat.test_all) {
} else {
}
// wait for fitness jobs to complete
// sort cells on performance
// generate crossover jobs
cur_cells = cur_cells == cells_a ? cells_b : cells_a;
next_cells = cur_cells == cells_a ? cells_b : cells_a;
}
// stop worker threads
stop_flag = true;
pthread_cond_broadcast(queue.jobs_available_cond);
pthread_cond_broadcast(work_queue.jobs_available_cond);
for (int i = 0; i < strat.num_threads; i++) {
pthread_join(threads[i], NULL);
}
}
template <class T> T &Span<T>::operator[](int i) {
assert(i >= 0 && i < len);
return _data[i];
}
} // namespace genetic