LCPQ/qp2
9
1
Fork 0
mirror of https://github.com/QuantumPackage/qp2.git synced 2024-12-22 19:43:32 +01:00
Code Releases Activity

Disk-based cholesky

Browse Source
This commit is contained in:
Anthony Scemama 2024-06-04 16:15:09 +02:00
parent 2a9b8c56a1
commit c95a0b2d87
3 changed files with 96 additions and 96 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

178
src/ao_two_e_ints/cholesky.irp.f
View File

@ -16,6 +16,7 @@ END_PROVIDER
BEGIN_PROVIDER [ integer, cholesky_ao_num ]
&BEGIN_PROVIDER [ double precision, cholesky_ao, (ao_num, ao_num, 1) ]
use mmap_module
implicit none
BEGIN_DOC
! Cholesky vectors in AO basis: (ik|a):
@ -30,19 +31,19 @@ END_PROVIDER
integer*8 :: ndim8
integer :: rank
double precision :: tau, tau2
double precision, pointer :: L(:,:), L_old(:,:)
double precision, pointer :: L(:,:)
double precision :: s
double precision :: dscale, dscale_tmp
double precision, allocatable :: D(:), Delta(:,:), Ltmp_p(:,:), Ltmp_q(:,:)
double precision, allocatable :: D(:), Delta(:,:), Ltmp_p(:,:), Ltmp_q(:,:), D_sorted(:)
integer, allocatable :: addr1(:), addr2(:)
integer*8, allocatable :: Lset(:), Dset(:), addr3(:)
logical, allocatable :: computed(:)
integer :: i,j,k,m,p,q, dj, p2, q2
integer*8 :: i8, j8, p8, qj8
integer :: N, np, nq, npmax
integer*8 :: i8, j8, p8, qj8, rank_max, np8
integer :: N, np, nq
double precision :: Dmax, Dmin, Qmax, f
double precision, external :: get_ao_two_e_integral
@ -61,12 +62,12 @@ END_PROVIDER
ndim8 = ao_num*ao_num*1_8
double precision :: wall0,wall1
type(c_ptr) :: c_pointer(2)
integer :: fd(2)
call wall_time(wall0)
deallocate(cholesky_ao)
! TODO : Save L() to disk
if (read_ao_cholesky) then
print *, 'Reading Cholesky vectors from disk...'
iunit = getUnitAndOpen(trim(ezfio_work_dir)//'cholesky_ao', 'R')
@ -81,6 +82,16 @@ END_PROVIDER
PROVIDE nucl_coord ao_two_e_integral_schwartz
call set_multiple_levels_omp(.False.)
rank_max = min(ndim8,274877906944_8/1_8/ndim8)
call mmap(trim(ezfio_work_dir)//'cholesky_ao_tmp', (/ ndim8, rank_max /), 8, fd(1), .False., c_pointer(1))
call c_f_pointer(c_pointer(1), L, (/ ndim8, rank_max /))
!print *, 'rank_max/ndim8', dble(rank_max) / dble(ndim8)
! Deleting the file while it is open makes the file invisible on the filesystem,
! and automatically deleted, even if the program crashes
iunit = getUnitAndOpen(trim(ezfio_work_dir)//'cholesky_ao_tmp', 'R')
close(iunit,status='delete')
if (do_direct_integrals) then
if (ao_two_e_integral(1,1,1,1) < huge(1.d0)) then
! Trigger providers inside ao_two_e_integral
@ -98,9 +109,6 @@ END_PROVIDER
call print_memory_usage()
allocate(L(ndim8,1))
!print *, 'allocate : (L(ndim8,1))', memory_of_double8(ndim8)
print *, ''
print *, 'Cholesky decomposition of AO integrals'
print *, '======================================'
@ -112,7 +120,7 @@ END_PROVIDER
rank = 0
allocate( D(ndim8), Lset(ndim8), Dset(ndim8) )
allocate( D(ndim8), Lset(ndim8), Dset(ndim8), D_sorted(ndim8))
allocate( addr1(ndim8), addr2(ndim8), addr3(ndim8) )
!print *, 'allocate : (D(ndim8))', memory_of_int8(ndim8)
!print *, 'allocate : (Lset(ndim8))', memory_of_int8(ndim8)
@ -132,44 +140,52 @@ END_PROVIDER
if (do_direct_integrals) then
!$OMP PARALLEL DO DEFAULT(SHARED) PRIVATE(i8) SCHEDULE(dynamic,16)
do i8=1,ndim8
do i8=ndim8,1,-1
D(i8) = ao_two_e_integral(addr1(i8), addr2(i8), &
addr1(i8), addr2(i8))
enddo
!$OMP END PARALLEL DO
else
!$OMP PARALLEL DO DEFAULT(SHARED) PRIVATE(i8) SCHEDULE(dynamic,16)
do i8=1,ndim8
do i8=ndim8,1,-1
D(i8) = get_ao_two_e_integral(addr1(i8), addr1(i8), &
addr2(i8), addr2(i8), &
ao_integrals_map)
enddo
!$OMP END PARALLEL DO
endif
D_sorted(:) = -D(:)
call dsort_noidx_big(D_sorted,ndim8)
D_sorted(:) = dabs(D_sorted(:))
Dmax = maxval(D)
Dmax = D_sorted(1)
! 2.
npmax = huge(1_4)*1_8
np = npmax
dscale = 1.d0
dscale_tmp = Dmax
do while (np == npmax)
np=0
dscale = tau2/Dmax
do i8=1,ndim8
if (D_sorted(i8) <= dscale) exit
enddo
mem = qp_max_mem+1
do while ( (mem > qp_max_mem).and.(i8>1_8) )
dscale = min(1.d0,dsqrt(tau2/(D_sorted(i8)*Dmax)))
dscale_tmp = dscale*dscale*Dmax
! print *, 'dscale = ', dscale, dble(i8)/dble(ndim8)
np8=0_8
do p8=1,ndim8
if ( dscale_tmp*D(p8) > tau2 ) then
np = np+1
Lset(np) = p8
if (np == npmax) then
! Overflow detected
dscale = dscale*0.1d0
dscale_tmp = dscale*dscale*Dmax
!print *, 'Overflow detected '
!print *, 'dscale = ', dscale
exit
endif
np8 = np8+1_8
Lset(np8) = p8
endif
enddo
i8 = i8*3_8/4_8
if (np8 > huge(1_4)/64_8) cycle
np = np8
! print *, 'np = ', np
call resident_memory(mem)
mem = mem &
+ 0.1d0*np*memory_of_double(np) ! Delta(np,nq)
enddo
! 3.
@ -179,7 +195,7 @@ END_PROVIDER
i = 0
! 5.
do while ( (Dmax > tau).and.(rank < min(ndim8,huge(1_4))) )
do while ( (Dmax > tau).and.(rank*1_8 < min(ndim8,rank_max)) )
! a.
i = i+1
@ -202,12 +218,12 @@ END_PROVIDER
enddo
call total_memory(mem)
call resident_memory(mem)
mem = mem &
+ np*memory_of_double(nq) &! Delta(np,nq)
+ (rank+nq)* memory_of_double8(ndim8) &! L(ndim8,rank+nq)
+ (np+nq)*memory_of_double(block_size) ! Ltmp_p(np,block_size) + Ltmp_q(nq,block_size)
!print *, 'mem = ', mem
if (mem > qp_max_mem) then
s = s*2.d0
else
@ -217,7 +233,7 @@ END_PROVIDER
if ((s > 1.d0).or.(nq == 0)) then
call print_memory_usage()
print *, 'Required peak memory: ', mem, 'Gb'
call total_memory(mem)
call resident_memory(mem)
print *, 'Already used memory: ', mem, 'Gb'
print *, 'Not enough memory. Reduce cholesky threshold'
stop -1
@ -227,26 +243,6 @@ END_PROVIDER
! d., e.
L_old => L
allocate(L(ndim8,rank+nq), stat=ierr)
!print *, 'allocate : L(ndim8,rank+nq)', memory_of_double8(ndim8*(rank+nq))
if (ierr /= 0) then
call print_memory_usage()
print *, irp_here, ': allocation failed : (L(ndim8,rank+nq))'
stop -1
endif
!$OMP PARALLEL DO PRIVATE(k,j8)
do k=1,rank
do j8=1,ndim8
L(j8,k) = L_old(j8,k)
enddo
enddo
!$OMP END PARALLEL DO
deallocate(L_old)
allocate(Delta(np,nq), stat=ierr)
!print *, 'allocate : Delta(np,nq)', memory_of_double8(np*nq*1_8)
@ -280,25 +276,16 @@ END_PROVIDER
!print *, 'N, rank, block_size', N, rank, block_size
!print *, 'p1'
!$OMP PARALLEL DEFAULT(SHARED) PRIVATE(k,p,q,j)
!print *, 'computed'
!$OMP DO
!$OMP PARALLEL DO DEFAULT(SHARED) PRIVATE(p,q,j)
do q=1,nq
computed(q) = .False.
enddo
!$OMP ENDDO NOWAIT
!print *, 'Delta'
!$OMP DO
do q=1,nq
do j=1,np
Delta(j,q) = 0.d0
enddo
enddo
!$OMP ENDDO NOWAIT
!$OMP END PARALLEL DO
!print *, 'Ltmp_p'
!$OMP PARALLEL DEFAULT(SHARED) PRIVATE(k,p,q,j)
do k=1,N
!$OMP DO
do p=1,np
@ -312,7 +299,6 @@ END_PROVIDER
enddo
!$OMP END DO NOWAIT
enddo
!$OMP BARRIER
!$OMP END PARALLEL
@ -338,7 +324,7 @@ END_PROVIDER
rank = N+j
if (iblock == block_size) then
!print *, 'dgemm'
!print *, 'dgemm', np, nq
call dgemm('N','T',np,nq,block_size,-1.d0, &
Ltmp_p, np, Ltmp_q, nq, 1.d0, Delta, np)
@ -401,7 +387,6 @@ END_PROVIDER
! iii.
f = 1.d0/dsqrt(Qmax)
!print *, 'p4'
!$OMP PARALLEL PRIVATE(p,q) DEFAULT(shared)
!$OMP DO
do p=1,np
@ -441,29 +426,41 @@ END_PROVIDER
Dmax = max(Dmax, D(Lset(p)))
enddo
np = npmax
dscale = 1.d0
dscale_tmp = Dmax
do while (np == npmax)
np=0
mem = qp_max_mem+1
do while ( (mem > qp_max_mem).and.(i8>1_8) )
dscale = min(1.d0,dsqrt(tau2/(D_sorted(i8)*Dmax)))
dscale_tmp = dscale*dscale*Dmax
!print *, 'dscale = ', dscale, dble(i8)/dble(ndim8)
np8=0_8
do p8=1,ndim8
if ( dscale_tmp*D(p8) > tau2 ) then
np = np+1
Lset(np) = p8
if (np == npmax) then
! Overflow detected
dscale = dscale*0.5d0
dscale_tmp = dscale*dscale*Dmax
!print *, 'Overflow detected '
!print *, 'dscale = ', dscale
exit
endif
np8 = np8+1_8
Lset(np8) = p8
endif
enddo
i8 = i8*3_8/4_8
if (np8 > huge(1_4)/64_8) cycle
np = np8
!print *, 'np = ', np
call resident_memory(mem)
mem = mem &
+ 0.1d0*np*memory_of_double(np) ! Delta(np,nq)
enddo
if (np == 0) then
call print_memory_usage()
print *, 'Required peak memory: ', mem, 'Gb'
call resident_memory(mem)
print *, 'Already used memory: ', mem, 'Gb'
print *, 'Not enough memory. Reduce cholesky threshold'
stop -1
endif
enddo
enddo
print *, '============ ============='
print *, ''
allocate(cholesky_ao(ao_num,ao_num,rank), stat=ierr)
!print *, 'allocate : cholesky_ao(ao_num,ao_num,rank)', memory_of_double8(ao_num*ao_num*rank*1_8)
@ -473,18 +470,19 @@ END_PROVIDER
print *, irp_here, ': Allocation failed'
stop -1
endif
!$OMP PARALLEL DO PRIVATE(k)
!$OMP PARALLEL DO PRIVATE(k,j)
do k=1,rank
do j=1,ao_num
cholesky_ao(1:ao_num,j,k) = L((j-1)*ao_num+1:j*ao_num,k)
enddo
enddo
!$OMP END PARALLEL DO
deallocate(L)
cholesky_ao_num = rank
print *, '============ ============='
print *, ''
call munmap( (/ ndim8, ndim8 /), 8, fd(1), c_pointer(1) )
cholesky_ao_num = rank
if (write_ao_cholesky) then
print *, 'Writing Cholesky vectors to disk...'

4
src/ao_two_e_ints/two_e_integrals.irp.f
View File

@ -460,8 +460,8 @@ BEGIN_PROVIDER [ double precision, ao_two_e_integral_schwartz, (ao_num, ao_num)
!$OMP PARALLEL DO PRIVATE(i,k) &
!$OMP DEFAULT(NONE) &
!$OMP SHARED (ao_num,ao_two_e_integral_schwartz) &
!$OMP SCHEDULE(guided)
do i=1,ao_num
!$OMP SCHEDULE(dynamic)
do i=ao_num,1,-1
do k=1,i
ao_two_e_integral_schwartz(i,k) = dsqrt(ao_two_e_integral(i,i,k,k))
ao_two_e_integral_schwartz(k,i) = ao_two_e_integral_schwartz(i,k)

8
src/ezfio_files/get_unit_and_open.irp.f
View File

@ -47,11 +47,13 @@ integer function getUnitAndOpen(f,mode)
endif
open(unit=getUnitAndOpen,file=f,status='OLD',action='READ',form='UNFORMATTED')
else if (mode.eq.'W') then
open(unit=getUnitAndOpen,file=new_f,status='UNKNOWN',action='WRITE',form='UNFORMATTED')
open(unit=getUnitAndOpen,file=new_f,status='UNKNOWN',action='READWRITE',form='UNFORMATTED')
else if (mode.eq.'A') then
open(unit=getUnitAndOpen,file=new_f,status='UNKNOWN',action='READWRITE',position='APPEND',form='UNFORMATTED')
else if (mode.eq.'w') then
open(unit=getUnitAndOpen,file=new_f,status='UNKNOWN',action='WRITE',form='FORMATTED')
open(unit=getUnitAndOpen,file=new_f,status='UNKNOWN',action='READWRITE',form='FORMATTED')
else if (mode.eq.'a') then
open(unit=getUnitAndOpen,file=new_f,status='UNKNOWN',action='WRITE',position='APPEND',form='FORMATTED')
open(unit=getUnitAndOpen,file=new_f,status='UNKNOWN',action='READWRITE',position='APPEND',form='FORMATTED')
else if (mode.eq.'x') then
open(unit=getUnitAndOpen,file=new_f,form='FORMATTED')
endif