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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252
|
<?xml version="1.0" encoding="UTF-8"?>
<simulation xmds-version="2">
<testing>
<xsil_file name="bug_adaptive_timestep_hang.xsil" expected="bug_adaptive_timestep_hang_expected.xsil" absolute_tolerance="1e-7" relative_tolerance="1e-5">
<moment_group number="1" absolute_tolerance="1e-4" relative_tolerance="1e-3" />
<moment_group number="2" absolute_tolerance="1e-4" relative_tolerance="1e-6" />
<moment_group number="3" absolute_tolerance="1e-4" relative_tolerance="1e-6" />
</xsil_file>
</testing>
<name>bug_adaptive_timestep_hang</name>
<author>Mattias Johnsson</author>
<description>
GPE simulation of an Rb85 BEC in a harmonic trap. In the ground state
(F=2, mf=2) we have U=0. We then flip a portion into the F=3 state.
Here the scattering length is ~400a0. No one knows that the inter-state
scattering length is. Wait a while, then flip it back.
What happens?
We make use of cylindrical trapping frequencies, so assume omega_x=omega_y = omega_r.
That is, system is invariant under rotation around z. To do this we
use Hankel transforms instead of Fourier transforms, and describe the
(x,y) coords in (r, theta) = (r). That is, use a Bessel basis rather than a
Catesian one.
This means our 2D sim is fully accurate and there's no dimensional reductional
crap to deal with, i.e. we we don't have to scale the nonlinearities.
Note that the equations are done in a rotating frame to remove the Uab/2/dV
vacuum correction, i.e. I'm using the tilde variables in
a = atilde exp(i t/hbar Uab/2/dV )
so keep this in mind when looking at expectation value terms that don't consist
of equal numbers of daggered and non-daggered operators. See p1596 for details.
</description>
<features>
<benchmark />
<bing />
<fftw plan="patient" />
<auto_vectorise />
<globals>
<![CDATA[
const double hbar = 1.054e-34;
const double amu = 1.660539e-27;
const double mass = 84.912 * amu; // Rb 85
//const double mass = 86.909 * amu; // Rb 87
const double Omega = 2.0 * M_PI * 1.25e5; // Coupling Rabi frequency, move 1/2 population in 1us
//const double Omega = 3.06278e5; // Coupling Rabi frequency, move 1/11 population in 1us
//const double Omega = 3.2175e5; // Coupling Rabi frequency, move 1/10 population in 1us
const double omega_r = 2.0 * M_PI * 50.0;
const double omega_z = 2.0 * M_PI * 30.0;
const double sr = sqrt(hbar/mass/omega_r);
const double sz = sqrt(hbar/mass/omega_z);
const double N0 = 5e4;
const double a0 = 5.2918e-11; // Bohr radius
const double a3d_initial = 400.0 * a0;
const double a3d_00 = 400.0 * a0;
const double a3d_10 = 0.0 * a0;
const double a3d_11 = 400.0 * a0;
const double U3d_initial = 4.0*M_PI*hbar*hbar*a3d_initial/mass;
const double U3d_00 = 4.0*M_PI*hbar*hbar*a3d_00/mass;
const double U3d_10 = 4.0*M_PI*hbar*hbar*a3d_10/mass;
const double U3d_11 = 4.0*M_PI*hbar*hbar*a3d_11/mass;
const double BSPhase = 0.2;
const double mu3d_initial = pow(15.0 * pow(mass,1.5) * omega_r*omega_r*omega_z*N0*U3d_initial/16.0/sqrt(2.0)/M_PI, 0.4);
]]>
</globals>
</features>
<!-- <driver name="mpi-multi-path" paths="1000" /> -->
<geometry>
<propagation_dimension> t </propagation_dimension>
<transverse_dimensions>
<!-- For TF ground state -->
<dimension name="r" lattice="32" domain="(0, 16e-6)" transform="bessel" volume_prefactor="2*M_PI" />
<dimension name="z" lattice="64" domain="(-24e-6, 24e-6)" />
</transverse_dimensions>
</geometry>
<noise_vector name="initnoise" kind="gaussian" method="dsfmt" type="complex" seed="667813208 116090193 819982544">
<components>n_1 n_2</components>
</noise_vector>
<vector name="wavefunction" type="complex" dimensions="r z">
<components> psi0 psi1 </components>
<!-- Initialisation by function -->
<initialisation>
<dependencies>initnoise</dependencies>
<![CDATA[
// ground state HO
//psi0 = sqrt(N0) * 1/sqrt(sr*sr*sz) * pow(M_PI,-0.75) * exp(-0.5*(r*r/sr/sr + z*z/sz/sz)) + n_1/sqrt(2.0);
// Ground state Thomas-Fermi
if (mu3d_initial-0.5*mass*(omega_r*omega_r*r*r + omega_z*omega_z*z*z) >=0 ) {
psi0 = sqrt(2)*sqrt(mu3d_initial-0.5*mass*(omega_r*omega_r*r*r + omega_z*omega_z*z*z))/sqrt(U3d_initial) + n_1/sqrt(2.0);
} else {
psi0 = n_1/sqrt(2.0);
}
psi1 = n_2 /sqrt(2.0);
//printf("mu=%e; U3d_initial=%e; ",mu3d_initial,U3d_initial);
]]>
</initialisation>
</vector>
<vector name="potential" type="real" dimensions="r z">
<components> V </components>
<initialisation>
<![CDATA[
V = 0.5 * mass * (omega_r*omega_r*r*r + omega_z*omega_z*z*z);
]]>
</initialisation>
</vector>
<computed_vector name="particle_number" dimensions="" type="real">
<!-- The point of this computed vector is that we need access to the
quantity number^2. A standard moment group will give us number by
integrating the density field, but it can't then square the result.
To get around this we calcuate the number in a CV, and then square
that result in the samping_group output.
Also note that the stochatic averaging happens *after* this is done.
We stochatically average (number squared).
-->
<components>
computed_number0 computed_number1
</components>
<evaluation>
<dependencies>wavefunction</dependencies>
<![CDATA[
computed_number0 = mod2(psi0) - 0.5/dr/dz ;
computed_number1 = mod2(psi1) - 0.5/dr/dz ;
]]>
</evaluation>
</computed_vector>
<sequence>
<!-- first beam splitter -->
<integrate algorithm="ARK45" interval="1e-6" tolerance="1e-7">
<samples> 10 10 10 </samples>
<operators>
<integration_vectors>wavefunction</integration_vectors>
<dependencies>potential</dependencies>
<operator kind="ip" type="imaginary" >
<operator_names>Ltt</operator_names>
<![CDATA[
Ltt = -i*0.5*hbar*(kr*kr + kz*kz)/mass;
]]>
</operator>
<![CDATA[
dpsi0_dt = Ltt[psi0] - i/hbar * (V) * psi0 - i*Omega*psi1;
dpsi1_dt = Ltt[psi1] - i/hbar * (V) * psi1 - i*Omega*psi0;
]]>
</operators>
</integrate>
<!-- First hold time -->
<integrate algorithm="ARK45" interval="4e-4" tolerance="1e-7">
<samples> 10 10 10 </samples>
<operators>
<integration_vectors>wavefunction</integration_vectors>
<dependencies>potential</dependencies>
<operator kind="ip" constant="yes" type="imaginary" >
<operator_names>Ltt</operator_names>
<![CDATA[
Ltt = -i*0.5*hbar*(kr*kr + kz*kz)/mass;
]]>
</operator>
<![CDATA[
dpsi0_dt = Ltt[psi0] - i/hbar * (V + U3d_00*(mod2(psi0)-1.0/dr/dz) + U3d_10*(mod2(psi1)-0.5/dr/dz)) * psi0;
dpsi1_dt = Ltt[psi1] - i/hbar * (V + U3d_11*(mod2(psi1)-1.0/dr/dz) + U3d_10*(mod2(psi0)-0.5/dr/dz)) * psi1;
]]>
</operators>
</integrate>
<!-- And now the final beam splitter integration -->
<integrate algorithm="ARK45" interval="2e-6" tolerance="1e-7">
<samples> 10 10 50 </samples>
<operators>
<integration_vectors>wavefunction</integration_vectors>
<dependencies>potential</dependencies>
<operator kind="ip" constant="yes" type="imaginary" >
<operator_names>Ltt</operator_names>
<![CDATA[
Ltt = -i*0.5*hbar*(kr*kr + kz*kz)/mass;
]]>
</operator>
<![CDATA[
dpsi0_dt = Ltt[psi0] - i/hbar * (V ) * psi0 - i*exp(i*BSPhase)*Omega*psi1;
dpsi1_dt = Ltt[psi1] - i/hbar * (V ) * psi1 - i*exp(-i*BSPhase)*Omega*psi0;
]]>
</operators>
</integrate>
</sequence>
<output>
<sampling_group basis="r(0) z(0)" initial_sample="yes">
<moments>psi0real psi0imag psi1real psi1imag density0 density1 </moments>
<dependencies>wavefunction</dependencies>
<![CDATA[
psi0real = psi0.Re();
psi0imag = psi0.Im();
psi1real = psi1.Re();
psi1imag = psi1.Im();
density0 = mod2(psi0) - 0.5/dz/dr;
density1 = mod2(psi1) - 0.5/dz/dr;
]]>
</sampling_group>
<sampling_group basis="kr(0) kz(0)" initial_sample="yes">
<moments> density0k density1k </moments>
<dependencies>wavefunction</dependencies>
<![CDATA[
density0k = mod2(psi0)- 0.5/dkz/dkr;
density1k = mod2(psi1)- 0.5/dkz/dkr;
]]>
</sampling_group>
<sampling_group basis="r(1) z(1)" initial_sample="yes">
<moments> number0 number1 number0squared number1squared </moments>
<dependencies>wavefunction particle_number </dependencies>
<![CDATA[
number0 = computed_number0;
number1 = computed_number1;
number0squared = computed_number0 * computed_number0 - _lattice_r*_lattice_z/4.0;
number1squared = computed_number1 * computed_number1 - _lattice_r*_lattice_z/4.0;
]]>
</sampling_group>
</output>
</simulation>
|