-
Notifications
You must be signed in to change notification settings - Fork 0
/
disp.py
389 lines (342 loc) · 15.9 KB
/
disp.py
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
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
#!/usr/bin/env python
"""
my second torch script, use to calculate disp-disp interaction
"""
from functools import partial
import torch
from dmff_torch.pairwise import distribute_v3, distribute_scalar, distribute_dispcoeff
from dmff_torch.pme import setup_ewald_parameters
from dmff_torch.recip import pme_recip
from dmff_torch.utils import jit_condition
from functorch import vmap
import numpy as np
from dmff_torch.nblist import build_covalent_map
class ADMPDispPmeForce:
'''
This is a convenient wrapper for dispersion PME calculations
It wrapps all the environment parameters of multipolar PME calculation
The so called "environment paramters" means parameters that do not need to be differentiable
'''
def __init__(self, box, rc, ethresh, pmax, lpme=True):
self.rc = rc
self.ethresh = ethresh
self.pmax = pmax
# Need a different function for dispersion ??? Need tests
self.lpme = lpme
self.device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
if lpme:
kappa, K1, K2, K3 = setup_ewald_parameters(rc, ethresh, box)
self.kappa = kappa
self.K1 = K1
self.K2 = K2
self.K3 = K3
###############################################################################
# modify here, for the torch.jit purpose
pme_order = torch.tensor(6, dtype=torch.int32, device=self.device)
bspline_range = torch.arange(-pme_order//2, pme_order//2)
n_mesh = pme_order**3
shift_y,shift_x,shift_z = torch.meshgrid(bspline_range, bspline_range, bspline_range,indexing='ij')
shifts = torch.stack((shift_x,shift_y,shift_z)).transpose(0,3).reshape((1,n_mesh,3))
self.n_mesh = torch.tensor(n_mesh, dtype=torch.int32, device=self.device)
self.shifts = shifts.to(torch.float32)
##############################################################################
else:
self.kappa = torch.tensor(0.0, dtype=torch.float32, device=self.device)
self.K1 = torch.tensor(0, dtype=torch.int32, device=self.device)
self.K2 = torch.tensor(0, dtype=torch.int32, device=self.device)
self.K3 = torch.tensor(0, dtype=torch.int32, device=self.device)
self.n_mesh = None
self.shifts = None
# setup calculators
self.refresh_calculators()
return
#def generate_get_energy(self):
# energy_disp_pme(positions, box, pairs,
# c_list, mScales,
# self.kappa, self.K1, self.K2, self.K3, self.pmax,
# self.n_mesh, self.shifts, lpme=self.lpme)
def update_env(self, attr, val):
'''
Update the environment of the calculator
'''
setattr(self, attr, val)
self.refresh_calculators()
def refresh_calculators(self):
'''
refresh the energy and force calculator according to the current environment
'''
self.get_energy = energy_disp_pme
return
def energy_disp_pme(positions, box, pairs,
c_list, mScales, kappa, K1, K2, K3, pmax,
n_mesh, shifts, aex, aes, apol, adisp, adhf,
b, ldmp, lpme=True):
'''
Top level wrapper for dispersion pme
Input:
positions:
Na * 3: positions
box:
3 * 3: box, axes arranged in row
pairs:
Np * 3: interacting pair indices and topology distance
c_list:
Na * (pmax-4)/2: atomic dispersion coefficients
mScales:
(Nexcl,): permanent multipole-multipole interaction exclusion scalings: 1-2, 1-3 ...
covalent_map:
Na * Na: topological distances between atoms, if i, j are topologically distant, then covalent_map[i, j] == 0
disp_pme_recip_fn:
function: the reciprocal calculator, see recip.py
kappa:
float: kappa in A^-1
K1, K2, K3:
int: max K for reciprocal calculations
pmax:
int array: maximal exponents (p) to compute, e.g., (6, 8, 10)
lpme:
bool: whether do pme or not, useful when doing cluster calculations
Output:
energy: total dispersion pme energy
'''
if lpme is False:
kappa = torch.tensor(0, dtype=torch.int32, device=positions.device)
ene_real = disp_pme_real(positions, box, pairs, c_list, mScales, kappa, pmax, ldmp, aex, aes, apol, adisp, adhf, b)
if lpme:
ene_recip = pme_recip(torch.tensor(6, dtype=torch.int32, device=positions.device), kappa, True, K1, K2, K3, positions, box, c_list[:, 0, None], n_mesh, shifts, torch.tensor(0, dtype=torch.int32, device=positions.device))
if pmax >= torch.tensor(8, dtype=torch.int32, device=positions.device):
ene_recip += pme_recip(torch.tensor(8, dtype=torch.int32, device=positions.device), kappa, True, K1, K2, K3, positions, box, c_list[:, 1, None], n_mesh, shifts, torch.tensor(0, dtype=torch.int32, device=positions.device))
if pmax >= torch.tensor(10, dtype=torch.int32, device=positions.device):
ene_recip += pme_recip(torch.tensor(10, dtype=torch.int32, device=positions.device), kappa, True, K1, K2, K3, positions, box, c_list[:, 2, None], n_mesh, shifts, torch.tensor(0, dtype=torch.int32, device=positions.device))
ene_self = disp_pme_self(c_list, kappa, pmax)
return ene_real + ene_recip + ene_self
else:
return ene_real
def disp_pme_real(positions, box, pairs,
c_list,
mScales,
kappa, pmax, ldmp, aex,
aes, apol, adisp, adhf, b):
'''
This function calculates the dispersion real space energy
It expands the atomic parameters to pairwise parameters
Input:
positions:
Na * 3: positions
box:
3 * 3: box, axes arranged in row
pairs:
Np * 3: interacting pair indices and topology distance
c_list:
Na * (pmax-4)/2: atomic dispersion coefficients
mScales:
(Nexcl,): permanent multipole-multipole interaction exclusion scalings: 1-2, 1-3 ...
covalent_map:
Na * Na: topological distances between atoms, if i, j are topologically distant, then covalent_map[i, j] == 0
kappa:
float: kappa in A^-1
pmax:
int array: maximal exponents (p) to compute, e.g., (6, 8, 10)
Output:
ene: dispersion pme realspace energy
'''
@vmap
@jit_condition()
def regularize_pairs(p):
# using vmap; we view 2-d array with only its element (1-d array, exampe p = p[m]), but dp is same as p[:,0] - p[:,1]
dp = p[1] - p[0]
dp = torch.where(dp > torch.tensor(0, dtype=torch.int32, device=dp.device), torch.tensor(0, dtype=torch.int32, device=dp.device), torch.tensor(1, dtype=torch.int32, device=dp.device))
# vmap don't support .item on a Tensor, for nopbc system, no buffer atoms
#dp_vec = torch.tensor([dp, 2 * dp])
p[0] = p[0] - dp
p[1] = p[1] - dp * 2
return p
@vmap
@jit_condition()
def pair_buffer_scales(p):
dp = p[0] - p[1]
return torch.where(dp < torch.tensor(0, dtype=torch.int32, device=dp.device), torch.tensor(1, dtype=torch.int32, device=dp.device), torch.tensor(0, dtype=torch.int32, device=dp.device))
@partial(vmap, in_dims=(0, None, None), out_dims=0)
@jit_condition()
def v_pbc_shift(drvecs, box, box_inv):
unshifted_dsvecs = torch.matmul(drvecs, box_inv)
dsvecs = unshifted_dsvecs - torch.floor(unshifted_dsvecs + 0.5)
return torch.matmul(dsvecs, box)
pairs[:,:2] = regularize_pairs(pairs[:,:2])
box_inv = torch.linalg.inv(box)
ri = distribute_v3(positions.T, pairs[:, 0]).T
rj = distribute_v3(positions.T, pairs[:, 1]).T
nbonds = pairs[:, 2]
indices = (nbonds + (mScales.shape[0] - 1)) % mScales.shape[0]
mscales = distribute_scalar(mScales, indices)
buffer_scales = pair_buffer_scales(pairs[:, :2])
mscales = mscales * buffer_scales
ci = distribute_dispcoeff(c_list.T, pairs[:, 0]).T
cj = distribute_dispcoeff(c_list.T, pairs[:, 1]).T
dr = ri - rj
dr = v_pbc_shift(dr, box, box_inv)
norm_dr = torch.linalg.norm(dr, dim=-1)
if ldmp == True:
aexi = distribute_scalar(aex, pairs[:,0])
aexj = distribute_scalar(aex, pairs[:,1])
aesi = distribute_scalar(aes, pairs[:,0])
aesj = distribute_scalar(aes, pairs[:,1])
apoli = distribute_scalar(apol, pairs[:,0])
apolj = distribute_scalar(apol, pairs[:,1])
adispi = distribute_scalar(adisp, pairs[:,0])
adispj = distribute_scalar(adisp, pairs[:,1])
adhfi = distribute_scalar(adhf, pairs[:,0])
adhfj = distribute_scalar(adhf, pairs[:,1])
bi = distribute_scalar(b, pairs[:,0])
bj = distribute_scalar(b, pairs[:,1])
#################################################################################
# final disp interaction is - (E_lr - E_sr) here, we give E_lr_real - E_sr
#################################################################################
ene_real = - torch.sum(
disp_dmp_kernel(norm_dr, ci, cj, aexi, aexj, aesi, aesj, apoli, apolj, adispi, adispj, adhfi, adhfj, bi, bj, mscales) * buffer_scales) + torch.sum(disp_pme_real_kernel(norm_dr, ci, cj, box, box_inv, mscales, kappa, pmax) * buffer_scales)
#ene_real = torch.sum((disp_dmp_kernel(norm_dr, ci, cj, aexi, aexj, aesi, aesj, apoli, apolj, adispi, adispj, adhfi, adhfj, bi, bj, mscales) +
# disp_pme_real_kernel(norm_dr, ci, cj, box, box_inv, mscales, kappa, pmax))
# * buffer_scales
# )
else:
ene_real = torch.sum(
disp_pme_real_kernel(norm_dr, ci, cj, box, box_inv, mscales, kappa, pmax)
* buffer_scales
)
return torch.sum(ene_real)
@partial(vmap, in_dims=(0, None, None), out_dims=0)
def pbc_shift(drvecs, box, box_inv):
unshifted_dsvecs = torch.matmul(drvecs, box_inv)
dsvecs = unshifted_dsvecs - torch.floor(unshifted_dsvecs + 0.5)
return torch.matmul(dsvecs, box)
@jit_condition()
def g_p(x2, pmax):
'''
Compute the g(x, p) function
Inputs:
x:
float: the input variable
pmax:
int: the maximal powers of dispersion, here we assume evenly spacing even powers starting from 6
e.g., (6,), (6, 8) or (6, 8, 10)
Outputs:
g:
(p-4)//2: g(x, p)
'''
x4 = x2 * x2
x8 = x4 * x4
exp_x2 = torch.exp(-x2)
g6 = (1 + x2 + 0.5 * x4) * exp_x2
g8 = torch.where(pmax >= torch.tensor(8, dtype=torch.int32, device=x2.device), g6 + (x4 * x2 / 6) * exp_x2, torch.zeros_like(g6))
g10 = torch.where(pmax >= torch.tensor(10, dtype=torch.int32, device=x2.device), g8 + (x8 / 24) * exp_x2, torch.zeros_like(g6))
g = [g6, g8, g10]
return g
@vmap
#@jit_condition()
def disp_dmp_kernel(dr, ci, cj, aexi, aexj, aesi, aesj, apoli, apolj, adispi, adispj, adhfi, adhfj, bi, bj, m):
a_ex = (aexi * aexj)
a_es = (aesi * aesj)
a_pol = (apoli * apolj)
a_disp = (adispi * adispj)
a_dhf = (adhfi * adhfj)
b = torch.sqrt(bi * bj)
c6 = ci[0] * cj[0]
c8 = ci[1] * cj[1]
c10 = ci[2] * cj[2]
br = b * dr
br2 = br * br
exp_br = torch.exp(-br)
P = 1/3 * br2 + br + 1
x = br - (2 * br2 + 3 * br) / (br2 + 3 * br + 3)
s6 = 1 + x + x**2/2 + x**3/6 + x**4/24 + x**5/120 + x**6/720
s8 = s6 + x**7/5040 + x**8/40320
s10 = s8 + x**9/362880 + x**10/3628800
exp_x = torch.exp(-x)
f6 = exp_x * s6
f8 = exp_x * s8
f10 = exp_x * s10
f = (a_ex + a_es + a_pol + a_disp + a_dhf) * P * exp_br /2625.5 + (f6*c6/dr**6 + f8*c8/dr**8 + f10*c10/dr**10)
expdmp = torch.where(dr < torch.tensor(2.5, dtype=torch.float32, device=dr.device), torch.tensor(0., dtype=torch.float32, device=dr.device), torch.exp(-(dr-torch.tensor(2.5, dtype=torch.float32, device=dr.device))**3))
return f * m * expdmp
@partial(vmap, in_dims=(0, 0, 0, None, None, 0, None, None), out_dims=(0))
def disp_pme_real_kernel(dr, ci, cj, box, box_inv, mscales, kappa, pmax):
'''
The kernel to calculate the realspace dispersion energy
Inputs:
ri:
Np * 3: position i
rj:
Np * 3: position j
ci:
Np * (pmax-4)/2: dispersion coeffs of i, c6, c8, c10 etc
cj:
Np * (pmax-4)/2: dispersion coeffs of j, c6, c8, c10 etc
kappa:
float: kappa
pmax:
int: largest p in 1/r^p, assume starting from 6 with increment of 2
Output:
energy:
float: the dispersion pme energy
'''
@jit_condition()
def calc_e(dr, ci, cj, box, box_inv, mscales, kappa, pmax):
dr2 = dr * dr
#dr2 = torch.matmul(dr, dr)
x2 = kappa * kappa * dr2
g = g_p(x2, pmax)
dr6 = dr2 * dr2 * dr2
ene = (mscales + g[0] - 1) * ci[0] * cj[0] / dr6
dr8 = dr6 * dr2; dr10 = dr8 * dr2
ene8 = torch.where(pmax >= torch.tensor(8, dtype=torch.int32, device=mscales.device), (mscales + g[1] - 1) * ci[1] * cj[1] / dr8, torch.zeros_like(ene))
ene10 = torch.where(pmax >= torch.tensor(10, dtype=torch.int32, device=mscales.device), (mscales + g[2] - 1) * ci[2] * cj[2] / dr10, torch.zeros_like(ene))
ene = ene + ene8 + ene10
#expdmp = torch.where(dr < torch.tensor(2.5, dtype=torch.float32, device=dr.device), torch.tensor(0., dtype=torch.float32, device=dr.device), torch.exp(-(dr-torch.tensor(2.5, dtype=torch.float32, device=dr.device))**3))
expdmp = torch.tensor(1., dtype=torch.float32, device=dr.device)
ene = ene * expdmp
return ene
ene = calc_e(dr, ci, cj, box, box_inv, mscales, kappa, pmax)
return ene
@jit_condition()
def disp_pme_self(c_list, kappa, pmax):
'''
This function calculates the dispersion self energy
Inputs:
c_list:
Na * 3: dispersion susceptibilities C_6, C_8, C_10
kappa:
float: kappa used in dispersion
Output:
ene_self:
float: the self energy
'''
E_6 = -kappa**6/12 * torch.sum(c_list[:, 0]**2)
E_8 = torch.where(pmax >= torch.tensor(8, dtype=torch.int32, device=E_6.device), -kappa**8/48 * torch.sum(c_list[:, 1]**2), torch.zeros_like(E_6))
E_10 = torch.where(pmax >= torch.tensor(10, dtype=torch.int32, device=E_6.device), -kappa**10/240 * torch.sum(c_list[:, 2]**2), torch.zeros_like(E_6))
E_6 = E_6 + E_8 + E_10
return E_6
if __name__ == '__main__':
# first we define the cov_map from the topo infor
data = np.load('100K_properties.npz',allow_pickle=True)
num_idx = 26
positions = data['coord'][num_idx]; bonds = data['topo'][num_idx]
c6 = data['c6'][num_idx]; c8 = data['c8'][num_idx]; c10 = data['c10'][num_idx]
atoms = {'positions':positions,'bonds':bonds,'c6':c6,'c8':c8,'c10':c10}
cov_map = build_covalent_map(atoms, 6)
pair_full = []
for na in range(len(atoms['positions'])):
for nb in range(na + 1, len(atoms['positions'])):
pair_full.append([na, nb, 0])
pair_full = np.array(pair_full, dtype=int)
pair_full[:,2] = cov_map[pair_full[:,0], pair_full[:,1]]
pairs = torch.tensor(pair_full,requires_grad=False)
mscales = torch.tensor([0., 0., 0., 0., 1., 1.], dtype=torch.float32, requires_grad=False)
box = torch.tensor([[50.,0.,0.],[0.,50.,0.],[0.,0.,50.]], dtype=torch.float32, requires_grad=False)
positions = torch.tensor(positions, dtype=torch.float32, requires_grad=True)
c6 = torch.tensor(c6, dtype=torch.float32, requires_grad=False)
c8 = torch.tensor(c8, dtype=torch.float32, requires_grad=False)
c10 = torch.tensor(c10, dtype=torch.float32, requires_grad=False)
c_list = torch.vstack((c6,c8,c10)).T
e = energy_disp_pme(positions, box, pairs, c_list, mscales,
None, None, None, None, 10, None, None, None, lpme=False)
grad = torch.autograd.grad(outputs=e,inputs=positions)