Skip to content

Commit babfec6

Browse files
h-m0r1h-m0r1
authored
Merge pull request #17 from h-m0r1/positive_only_PR
Added Positive_only
2 parents cd4ff1f + 09e1eee commit babfec6

File tree

5 files changed

+4655
-1262
lines changed

5 files changed

+4655
-1262
lines changed

‎.github/workflows/test.yml

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -12,6 +12,8 @@ jobs:
1212
- uses: actions/checkout@v2
1313
- name: Set up python
1414
uses: actions/setup-python@v2
15+
with:
16+
python-version: '3.9'
1517
- name: Install dependencies
1618
run: |
1719
sudo apt-get --yes install ccache
@@ -32,6 +34,8 @@ jobs:
3234
- uses: actions/checkout@v2
3335
- name: Set up python
3436
uses: actions/setup-python@v2
37+
with:
38+
python-version: '3.9'
3539
- name: Set up repo
3640
run: |
3741
wget -O- https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor | sudo tee /usr/share/keyrings/oneapi-archive-keyring.gpg > /dev/null

‎mk_preset.py

Lines changed: 183 additions & 133 deletions
Original file line numberDiff line numberDiff line change
@@ -71,160 +71,205 @@ def run(nlambda_list, ndigit_list):
7171
ndigit_str = ' '.join(map(str, ndigit_list))
7272
print(
7373
f"""\
74-
! Precomputed data of IR for a parameter matrix of
75-
! nlambda = {nlambda_str} and ndigit = {ndigit_str}.
76-
! This file was automatically generated:
77-
! python3 mk_preset.py --nlambda {nlambda_str} --ndigit {ndigit_str} > sparse_ir_preset.f90
78-
! Do not edit this file manually.
79-
!
80-
module sparse_ir_preset
81-
use sparse_ir
82-
implicit none
83-
""")
74+
!! Precomputed data of IR for a parameter matrix of
75+
!! nlambda = {nlambda_str} and ndigit = {ndigit_str}.
76+
!! This file was automatically generated:
77+
!! python3 mk_preset.py --nlambda {nlambda_str} --ndigit {ndigit_str} > sparse_ir_preset.f90
78+
!! Do not edit this file manually.
79+
!!
80+
!-----------------------------------------------------------------------
81+
MODULE sparse_ir_preset
82+
!-----------------------------------------------------------------------
83+
USE sparse_ir
84+
!
85+
IMPLICIT NONE
86+
!
87+
PRIVATE
88+
!
89+
INTEGER, PARAMETER :: DP = KIND(0d0)
90+
!
91+
PUBLIC :: mk_ir_preset
92+
!
93+
""", end="")
8494
for nlambda in nlambda_list:
8595
for ndigit in ndigit_list:
8696
b = bases[(nlambda, ndigit)]
8797
sig = f"nlambda{nlambda}_ndigit{ndigit}"
8898
print(
8999
f"""\
90-
double precision :: s_{sig}({b.size})
91-
double precision :: tau_{sig}({b.ntau})
92-
integer :: freq_f_{sig}({b.nfreq_f})
93-
integer :: freq_b_{sig}({b.nfreq_b})
94-
double precision :: omega_{sig}({b.nomega})
95-
double precision :: u_r_{sig}({b.ntau_reduced} * {b.size})
96-
double precision :: uhat_f_r_{sig}({b.nfreq_f_reduced} * {b.size})
97-
double precision :: uhat_b_r_{sig}({b.nfreq_b_reduced} * {b.size})
98-
double precision :: v_r_{sig}({b.nomega_reduced} * {b.size})
99-
"""
100-
)
100+
REAL(KIND = DP) :: s_{sig}({b.size})
101+
REAL(KIND = DP) :: tau_{sig}({b.ntau})
102+
INTEGER :: freq_f_{sig}({b.nfreq_f})
103+
INTEGER :: freq_b_{sig}({b.nfreq_b})
104+
REAL(KIND = DP) :: omega_{sig}({b.nomega})
105+
REAL(KIND = DP) :: u_r_{sig}({b.ntau_reduced} * {b.size})
106+
REAL(KIND = DP) :: uhat_f_r_{sig}({b.nfreq_f_reduced} * {b.size})
107+
REAL(KIND = DP) :: uhat_b_r_{sig}({b.nfreq_b_reduced} * {b.size})
108+
REAL(KIND = DP) :: v_r_{sig}({b.nomega_reduced} * {b.size})
109+
""", end="")
101110

102111
print(
103112
"""\
104-
contains
105-
106-
function mk_ir_preset(nlambda, ndigit, beta) result(obj)
107-
integer, intent(in) :: nlambda, ndigit
108-
double precision, intent(in) :: beta
109-
type(IR) :: obj
110-
""")
113+
!
114+
CONTAINS
115+
!
116+
!-----------------------------------------------------------------------
117+
FUNCTION mk_ir_preset(nlambda, ndigit, beta, positive_only) result(obj)
118+
!-----------------------------------------------------------------------
119+
!
120+
INTEGER, INTENT(IN) :: nlambda, ndigit
121+
REAL(KIND = DP), INTENT(IN) :: beta
122+
LOGICAL, INTENT(IN), OPTIONAL :: positive_only
123+
TYPE(IR) :: obj
124+
!
125+
""", end="")
111126

112127
for nlambda in nlambda_list:
113128
for ndigit in ndigit_list:
114129
print(
115130
f"""\
116-
if (nlambda == {nlambda} .and. ndigit == {ndigit}) then
117-
obj = mk_nlambda{nlambda}_ndigit{ndigit}(beta)
118-
return
119-
end if
120-
"""
121-
)
131+
IF (nlambda == {nlambda} .and. ndigit == {ndigit}) THEN
132+
IF ((.NOT. PRESENT(positive_only))) THEN
133+
obj = mk_nlambda{nlambda}_ndigit{ndigit}(beta)
134+
ELSE
135+
obj = mk_nlambda{nlambda}_ndigit{ndigit}(beta, positive_only)
136+
ENDIF
137+
RETURN
138+
ENDIF
139+
!
140+
""", end="")
122141

123142

124143
print(
125144
"""\
126-
stop "Invalid parameters"
127-
end function
128-
"""
129-
)
145+
STOP "Invalid parameters"
146+
!
147+
!-----------------------------------------------------------------------
148+
END FUNCTION mk_ir_preset
149+
!-----------------------------------------------------------------------
150+
!
151+
""", end="")
130152

131153
for nlambda in nlambda_list:
132154
for ndigit in ndigit_list:
133155
print_data(nlambda, ndigit, bases[(nlambda, ndigit)])
134156

135-
print("end")
157+
print(
158+
"""\
159+
!-----------------------------------------------------------------------
160+
END MODULE sparse_ir_preset
161+
!-----------------------------------------------------------------------
162+
""", end="")
136163

137164

138165
def print_data(nlambda, ndigit, b):
139166
sig = f"nlambda{nlambda}_ndigit{ndigit}"
140167
print(
141-
f"""
142-
function mk_nlambda{nlambda}_ndigit{ndigit}(beta) result(obj)
143-
double precision, intent(in) :: beta
144-
type(IR) :: obj
145-
complex(kind(0d0)), allocatable :: u(:, :), uhat_f(:, :), uhat_b(:, :), v(:, :), dlr(:, :)
146-
integer, parameter :: size = {b.size}, ntau = {b.ntau}, nfreq_f = {b.nfreq_f}, nfreq_b = {b.nfreq_b}, nomega = {b.nomega}
147-
integer, parameter :: nlambda = {nlambda}, ndigit = {ndigit}
148-
integer, parameter :: ntau_reduced = ntau/2+1, nfreq_f_reduced = nfreq_f/2+1, nfreq_b_reduced = nfreq_b/2+1
149-
integer, parameter :: nomega_reduced = nomega/2+1
150-
double precision, parameter :: lambda = 1.d{nlambda}, eps = 1.d-{ndigit}
151-
152-
integer :: itau, l, ifreq, iomega
153-
""")
168+
f"""\
169+
!-----------------------------------------------------------------------
170+
FUNCTION mk_nlambda{nlambda}_ndigit{ndigit}(beta, positive_only) result(obj)
171+
!-----------------------------------------------------------------------
172+
!
173+
REAL(KIND = DP), INTENT(IN) :: beta
174+
LOGICAL, INTENT(IN), OPTIONAL :: positive_only
175+
TYPE(IR) :: obj
176+
REAL(KIND = DP), ALLOCATABLE :: u(:, :), v(:, :), dlr(:, :)
177+
COMPLEX(KIND = DP), ALLOCATABLE :: uhat_f(:, :), uhat_b(:, :)
178+
INTEGER, PARAMETER :: size = {b.size}, ntau = {b.ntau}, nfreq_f = {b.nfreq_f}, nfreq_b = {b.nfreq_b}, nomega = {b.nomega}
179+
INTEGER, PARAMETER :: nlambda = {nlambda}, ndigit = {ndigit}
180+
INTEGER, PARAMETER :: ntau_reduced = ntau/2+1
181+
INTEGER, PARAMETER :: nfreq_f_reduced = nfreq_f/2+1, nfreq_b_reduced = nfreq_b/2+1
182+
INTEGER, PARAMETER :: nomega_reduced = nomega/2+1
183+
REAL(KIND = DP), PARAMETER :: lambda = 1.d{nlambda}, eps = 1.d-{ndigit}
184+
!
185+
INTEGER :: itau, l, ifreq, iomega
186+
!
187+
""", end="")
154188
for varname in ["s", "tau", "freq_f", "freq_b", "u_r", "uhat_f_r", "uhat_b_r", "omega", "v_r"]:
155-
print(8*" " + f"call init_{varname}_{sig}()")
189+
print(2*" " + f"CALL init_{varname}_{sig}()")
156190

157191
print(
158-
f"""
159-
allocate(u(ntau, size))
160-
allocate(uhat_f(nfreq_f, size))
161-
allocate(uhat_b(nfreq_b, size))
162-
allocate(v(nomega, size))
163-
allocate(dlr(nomega, size))
164-
165-
! Use the fact U_l(tau) is even/odd for even/odd l-1.
166-
do l = 1, size
167-
do itau = 1, ntau_reduced
168-
u(itau, l) = u_r_{sig}(itau + {b.ntau_reduced}*(l-1))
169-
u(ntau-itau+1, l) = (-1)**(l-1) * u_r_{sig}(itau + {b.ntau_reduced}*(l-1))
170-
end do
171-
end do
172-
173-
! Use the fact U^F_l(iv) is pure imaginary/real for even/odd l-1.
174-
do l = 1, size, 2
175-
do ifreq = 1, nfreq_f_reduced
176-
uhat_f(ifreq, l) = cmplx(0.0, uhat_f_r_{sig}(ifreq + {b.nfreq_f_reduced}*(l-1)), kind(0d0))
177-
end do
178-
end do
179-
do l = 2, size, 2
180-
do ifreq = 1, nfreq_f_reduced
181-
uhat_f(ifreq, l) = cmplx(uhat_f_r_{sig}(ifreq + {b.nfreq_f_reduced}*(l-1)), 0.0, kind(0d0))
182-
end do
183-
end do
184-
do l = 1, size
185-
do ifreq = 1, nfreq_f
186-
uhat_f(nfreq_f-ifreq+1, l) = conjg(uhat_f(ifreq, l))
187-
end do
188-
end do
189-
190-
! Use the fact U^B_l(iv) is pure real/imaginary for even/odd l-1
191-
do l = 1, size, 2
192-
do ifreq = 1, nfreq_b_reduced
193-
uhat_b(ifreq, l) = cmplx(uhat_b_r_{sig}(ifreq + {b.nfreq_b_reduced}*(l-1)), 0.0d0, kind(0d0))
194-
end do
195-
end do
196-
do l = 2, size, 2
197-
do ifreq = 1, nfreq_b_reduced
198-
uhat_b(ifreq, l) = cmplx(0.0d0, uhat_b_r_{sig}(ifreq + {b.nfreq_b_reduced}*(l-1)), kind(0d0))
199-
end do
200-
end do
201-
do l = 1, size
202-
do ifreq = 1, nfreq_b
203-
uhat_b(nfreq_b-ifreq+1, l) = conjg(uhat_b(ifreq, l))
204-
end do
205-
end do
206-
207-
! Use the fact V_l(omega) is even/odd for even/odd l-1.
208-
do l = 1, size
209-
do iomega = 1, nomega_reduced
210-
v(iomega, l) = v_r_{sig}(iomega + {b.nomega_reduced}*(l-1))
211-
v(nomega-iomega+1, l) = (-1)**(l-1) * v_r_{sig}(iomega + {b.nomega_reduced}*(l-1))
212-
end do
213-
do iomega = 1, nomega
214-
dlr(iomega, l) = - s_{sig}(l) * v(iomega, l)
215-
end do
216-
end do
217-
218-
call init_ir(obj, beta, lambda, eps,&
219-
s_{sig}, tau_{sig},&
220-
freq_f_{sig}, freq_b_{sig},&
221-
u, uhat_f, uhat_b, omega_{sig},&
222-
v, dlr, 1d-20)
223-
224-
deallocate(u, uhat_f, uhat_b, v, dlr)
225-
end function
226-
"""
227-
)
192+
f"""\
193+
ALLOCATE(u(ntau, size))
194+
ALLOCATE(uhat_f(nfreq_f, size))
195+
ALLOCATE(uhat_b(nfreq_b, size))
196+
ALLOCATE(v(nomega, size))
197+
ALLOCATE(dlr(nomega, size))
198+
!
199+
! Use the fact U_l(tau) is even/odd for even/odd l-1.
200+
DO l = 1, size
201+
DO itau = 1, ntau_reduced
202+
u(itau, l) = u_r_{sig}(itau + {b.ntau_reduced}*(l-1))
203+
u(ntau-itau+1, l) = (-1)**(l-1) * u_r_{sig}(itau + {b.ntau_reduced}*(l-1))
204+
ENDDO
205+
ENDDO
206+
!
207+
! Use the fact U^F_l(iv) is pure imaginary/real for even/odd l-1.
208+
DO l = 1, size, 2
209+
DO ifreq = 1, nfreq_f_reduced
210+
uhat_f(ifreq, l) = CMPLX(0.0d0, uhat_f_r_{sig}(ifreq + {b.nfreq_f_reduced}*(l-1)), KIND = DP)
211+
ENDDO
212+
ENDDO
213+
DO l = 2, size, 2
214+
DO ifreq = 1, nfreq_f_reduced
215+
uhat_f(ifreq, l) = CMPLX(uhat_f_r_{sig}(ifreq + {b.nfreq_f_reduced}*(l-1)), 0.0, KIND = DP)
216+
ENDDO
217+
ENDDO
218+
DO l = 1, size
219+
DO ifreq = 1, nfreq_f
220+
uhat_f(nfreq_f-ifreq+1, l) = CONJG(uhat_f(ifreq, l))
221+
ENDDO
222+
ENDDO
223+
!
224+
! Use the fact U^B_l(iv) is pure real/imaginary for even/odd l-1
225+
DO l = 1, size, 2
226+
DO ifreq = 1, nfreq_b_reduced
227+
uhat_b(ifreq, l) = CMPLX(uhat_b_r_{sig}(ifreq + {b.nfreq_b_reduced}*(l-1)), 0.0d0, KIND = DP)
228+
ENDDO
229+
ENDDO
230+
DO l = 2, size, 2
231+
DO ifreq = 1, nfreq_b_reduced
232+
uhat_b(ifreq, l) = CMPLX(0.0d0, uhat_b_r_{sig}(ifreq + {b.nfreq_b_reduced}*(l-1)), KIND = DP)
233+
ENDDO
234+
ENDDO
235+
DO l = 1, size
236+
DO ifreq = 1, nfreq_b
237+
uhat_b(nfreq_b-ifreq+1, l) = CONJG(uhat_b(ifreq, l))
238+
ENDDO
239+
ENDDO
240+
!
241+
! Use the fact V_l(omega) is even/odd for even/odd l-1.
242+
DO l = 1, size
243+
DO iomega = 1, nomega_reduced
244+
v(iomega, l) = v_r_{sig}(iomega + {b.nomega_reduced}*(l-1))
245+
v(nomega-iomega+1, l) = (-1)**(l-1) * v_r_{sig}(iomega + {b.nomega_reduced}*(l-1))
246+
ENDDO
247+
DO iomega = 1, nomega
248+
dlr(iomega, l) = - s_{sig}(l) * v(iomega, l)
249+
ENDDO
250+
ENDDO
251+
!
252+
IF ((.NOT. PRESENT(positive_only))) THEN
253+
CALL init_ir(obj, beta, lambda, eps,&
254+
s_{sig}, tau_{sig},&
255+
freq_f_{sig}, freq_b_{sig},&
256+
u, uhat_f, uhat_b, omega_{sig},&
257+
v, dlr, 1d-20)
258+
ELSE
259+
CALL init_ir(obj, beta, lambda, eps,&
260+
s_{sig}, tau_{sig},&
261+
freq_f_{sig}, freq_b_{sig},&
262+
u, uhat_f, uhat_b, omega_{sig},&
263+
v, dlr, 1d-20, positive_only)
264+
ENDIF
265+
!
266+
DEALLOCATE(u, uhat_f, uhat_b, v, dlr)
267+
!
268+
!-----------------------------------------------------------------------
269+
END FUNCTION mk_nlambda{nlambda}_ndigit{ndigit}
270+
!-----------------------------------------------------------------------
271+
!
272+
""", end="")
228273

229274
print_vector_data(b.s, f"s_{sig}")
230275
print_vector_data(b.x, f"tau_{sig}")
@@ -253,25 +298,30 @@ def print_vector_data(vec, var_name):
253298
vec = np.asfortranarray(vec)
254299
vec = vec.T.ravel()
255300
print(
256-
f"""
257-
subroutine init_{var_name}()
258-
"""
259-
)
301+
f"""\
302+
!-----------------------------------------------------------------------
303+
SUBROUTINE init_{var_name}()
304+
!-----------------------------------------------------------------------
305+
!
306+
""", end="")
260307
start = 0
261308
while start < vec.size:
262309
end = min(start + NLINE, vec.size)
263310
sub_vec = vec[start:end]
264-
print(2*4*" ", f"{var_name}({start+1}:{end}) = (/ &")
311+
print(2*" " + f"{var_name}({start+1}:{end}) = (/ &")
265312
end_str = "&"
266313
print(", &\n".join(_array_to_strings(sub_vec)) + end_str)
267-
print(2*4*" " + "/)")
314+
print(2*" " + "/)")
268315
start = end
269316

270317
print(
271-
f"""
272-
end subroutine
273-
"""
274-
)
318+
f"""\
319+
!
320+
!-----------------------------------------------------------------------
321+
END SUBROUTINE init_{var_name}
322+
!-----------------------------------------------------------------------
323+
!
324+
""", end="")
275325

276326

277327
if __name__ == '__main__':

0 commit comments

Comments
 (0)