diff --git a/main.cpp b/main.cpp
index f6e4c40334aa49c4738c082961c859677b829df8..065823b290b5418d76da7c8e736369b281b5a1a9 100644
--- a/main.cpp
+++ b/main.cpp
@@ -5,6 +5,113 @@
#include <CLI/CLI.hpp>
+inline double e(double rho, double p, double gamma)
+{
+ return p/(rho*(gamma-1));
+}
+
+inline double p(double rho, double e, double gamma)
+{
+ return (gamma-1)*rho*e;
+}
+
+struct ReduceMin {
+private:
+ Kokkos::View<double*> x_;
+
+public:
+ typedef Kokkos::View<double*>::value_type value_type;
+
+ ReduceMin(const Kokkos::View<double*>& x) : x_ (x) {}
+
+ typedef Kokkos::View<double*>::size_type size_type;
+
+ KOKKOS_INLINE_FUNCTION void
+ operator() (const size_type i, value_type& update) const
+ {
+ if (x_(i) < update) {
+ update = x_(i);
+ }
+ }
+
+ KOKKOS_INLINE_FUNCTION void
+ join (volatile value_type& dst,
+ const volatile value_type& src) const
+ {
+ if (src < dst) {
+ dst = src;
+ }
+ }
+
+ KOKKOS_INLINE_FUNCTION void
+ init (value_type& dst) const
+ { // The identity under max is -Inf.
+ dst= Kokkos::reduction_identity<value_type>::min();
+ }
+};
+
+
+double acoustic_dt(const Kokkos::View<double*>& Vj,
+ const Kokkos::View<double*>& cj)
+{
+ const size_t nj = Vj.size();
+ double dt = std::numeric_limits<double>::max();
+
+ Kokkos::View<double*> Vj_cj("Vj_cj", nj);
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ Vj_cj[j] = Vj[j]/cj[j];
+ });
+
+ Kokkos::parallel_reduce(nj, ReduceMin(Vj_cj), dt);
+
+ // Kokkos::parallel_reduce(n, KOKKOS_LAMBDA(const long i, long& lcount) {
+ // lcount += (i % 2) == 0;
+ // }, count);
+ return dt;
+}
+
+
+void computeExplicitFluxes(const Kokkos::View<double*>& xr,
+ const Kokkos::View<double*>& xj,
+ const Kokkos::View<double*>& rhoj,
+ const Kokkos::View<double*>& uj,
+ const Kokkos::View<double*>& pj,
+ const Kokkos::View<double*>& cj,
+ const Kokkos::View<double*>& Vj,
+ Kokkos::View<double*>& ur,
+ Kokkos::View<double*>& pr)
+{
+ // calcul de ur
+ ur[0]=0;
+ const size_t nr = ur.size();
+ const size_t nj = uj.size();
+
+ Kokkos::parallel_for(nj-1, KOKKOS_LAMBDA(const int& j) {
+ const int r = j+1;
+ const int k = r;
+ const double ujr = uj[j];
+ const double ukr = uj[k];
+ const double pjr = pj[j];
+ const double pkr = pj[k];
+
+ ur[r]=(rhoj[j]*cj[j]*ujr + rhoj[k]*cj[k]*ukr + pjr-pkr)/(rhoj[j]*cj[j]+rhoj[k]*cj[k]);
+ });
+ ur[nr-1]=0;
+
+ // calcul de pr
+ pr[0] = pj[0] + rhoj[0]*cj[0]*(ur[0] - uj[0]);
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j) {
+ const int r = j+1;
+
+ const double ujr = uj[j];
+ const double pjr = pj[j];
+
+ pr[r]=pjr+rhoj[j]*cj[j]*(ujr-ur[r]);
+ });
+}
+
+
int main(int argc, char *argv[])
{
CLI::App app{"Pastis help"};
@@ -40,39 +147,7 @@ int main(int argc, char *argv[])
Kokkos::initialize(argc, argv);
Kokkos::DefaultExecutionSpace::print_configuration(std::cout);
- // if (argc < 2) {
- // fprintf(stderr, "Usage: %s [<kokkos_options>] <size>\n", argv[0]);
- // Kokkos::finalize();
- // exit(1);
- // }
-
- const long& n = number;
- std::cout << "Number of even integers from 0 to " << n - 1 << '\n';
-
- Kokkos::Timer timer;
- timer.reset();
-
- // Compute the number of even integers from 0 to n-1, in parallel.
- long count = 0;
- Kokkos::parallel_reduce(n, KOKKOS_LAMBDA(const long i, long& lcount) {
- lcount += (i % 2) == 0;
- }, count);
-
- double count_time = timer.seconds();
- std::cout << " Parallel: " << count << " " << rang::style::bold << count_time << rang::style::reset << '\n';
-
- timer.reset();
-
- // Compare to a sequential loop.
- long seq_count = 0;
- for (long i = 0; i < n; ++i) {
- seq_count += (i % 2) == 0;
- }
-
- count_time = timer.seconds();
- std::cout << "Sequential: " << seq_count << ' ' << rang::style::bold << count_time << rang::style::reset << '\n';
-
- const int nj=n;
+ const long& nj=number;
Kokkos::View<double*> xj("xj", nj);
Kokkos::View<double*> rhoj("rhoj", nj);
@@ -86,26 +161,133 @@ int main(int argc, char *argv[])
Kokkos::View<double*> gammaj("gammaj", nj);
Kokkos::View<double*> cj("cj", nj);
Kokkos::View<double*> mj("mj", nj);
+ Kokkos::View<double*> inv_mj("inv_mj", nj);
const int nr=nj+1;
Kokkos::View<double*> xr("xr", nr);
const double delta_x = 1./nj;
-
+ Kokkos::Timer timer;
timer.reset();
Kokkos::parallel_for(nr, KOKKOS_LAMBDA(const int& r){
xr[r] = r*delta_x;
});
- count_time = timer.seconds();
- std::cout << " Parallel: " << count << " " << rang::style::bold << count_time << rang::style::reset << '\n';
- for (int r=0; r<nr; ++r) {
- std::cout << xr[r] << '\n';
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ xj[j] = 0.5*(xr[j]+xr[j+1]);
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ Vj[j] = xr[j+1]-xr[j];
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ if (xj[j]<0.5) {
+ rhoj[j]=1;
+ } else {
+ rhoj[j]=0.125;
+ }
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ if (xj[j]<0.5) {
+ pj[j]=1;
+ } else {
+ pj[j]=0.1;
+ }
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ gammaj[j] = 1.4;
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ ej[j] = e(rhoj[j],pj[j],gammaj[j]);
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ Ej[j] = ej[j]+0.5*uj[j]*uj[j];
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ cj[j] = std::sqrt(gammaj[j]*pj[j]/rhoj[j]);
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ mj[j] = rhoj[j] * Vj[j];
+ });
+
+ const double tmax=0.2;
+ double t=0;
+
+ int itermax=std::numeric_limits<int>::max();
+ int iteration=0;
+
+ while((t<tmax) and (iteration<itermax)) {
+ double dt = 0.4*acoustic_dt(Vj, cj);
+ if (t+dt<tmax) {
+ t+=dt;
+ } else {
+ dt=tmax-t;
+ t=tmax;
+ }
+
+ if (iteration%100 == 0) {
+ std::cout << "dt=" << dt << "t=" << t << " i=" << iteration << '\n';
+ }
+
+ Kokkos::View<double*> ur("ur", nr);
+ Kokkos::View<double*> pr("pr", nr);
+
+ computeExplicitFluxes(xr, xj,
+ rhoj, uj, pj, cj, Vj,
+ ur, pr);
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ int rm=j;
+ int rp=j+1;
+
+ uj[j] += dt/mj[j]*(pr[rm]-pr[rp]);
+ Ej[j] += dt/mj[j]*(pr[rm]*ur[rm]-pr[rp]*ur[rp]);
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ ej[j] = Ej[j] - 0.5 * uj[j]*uj[j];
+ });
+
+ Kokkos::parallel_for(nr, KOKKOS_LAMBDA(const int& r){
+ xr[r] += dt*ur[r];
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ xj[j] = 0.5*(xr[j]+xr[j+1]);
+ Vj[j] = xr[j+1]-xr[j];
+ });
+
+ Kokkos::parallel_for(nj, KOKKOS_LAMBDA(const int& j){
+ rhoj[j] = mj[j]/Vj[j];
+ pj[j] = p(rhoj[j], ej[j], gammaj[j]);
+ cj[j] = std::sqrt(gammaj[j]*pj[j]/rhoj[j]); // inv_mj*vj
+ });
+
+ ++iteration;
}
+ std::cout << "* " << rang::style::underline << "Final time" << rang::style::reset
+ << ": " << rang::fgB::green << t << rang::fg::reset << " (" << iteration << " iterations)\n";
+ double count_time = timer.seconds();
+ std::cout << "* Execution time: " << rang::style::bold << count_time << rang::style::reset << '\n';
+
+ {
+ std::ofstream fout("rho");
+
+ for (int j=0; j<nj; ++j) {
+ fout << xj[j] << ' ' << rhoj[j] << '\n';
+ }
+ }
Kokkos::finalize();
return 0;