Skip to content

Commit

Permalink
Answers for SUNDIALS#447 (Fortran 32-bit Interface) (#33)
Browse files Browse the repository at this point in the history 
Updates for LLNL/sundials#447 (Fortran 32-bit Interface)

---------

Co-authored-by: Daniel R. Reynolds <[email protected]>
Co-authored-by: David Gardner <[email protected]>
  • Loading branch information
@balos1 @drreynolds @gardner48
3 people authored Jun 12, 2024
1 parent 96a64d9 commit dd4aebc
Show file tree
Hide file tree
Showing 5 changed files with 302 additions and 0 deletions.
80 changes: 80 additions & 0 deletions linux-ubuntu20.04-x86_64/gcc-9.4.0/double/cv_diag_kry_bbd_f2003.out
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,80 @@

Diagonal test problem:
neq = 40
nlocal = 10
nprocs = 4
rtol = 1.00E-05
atol = 1.00E-10
alpha = 1.00E+01
ydot_i = -alpha*i * y_i (i = 1,...,neq)
Method is BDF/NEWTON/SPGMR
Precond is band-block-diagonal, using CVBBDPRE

Preconditioning on left:
t steps fe
--------------------------------
0.100000 221 261
0.200000 265 307
0.300000 290 333
0.400000 306 350
0.500000 319 364
0.600000 329 374
0.700000 339 385
0.800000 345 391
0.900000 352 398
1.000000 359 405
--------------------------------
Max. absolute error is 2.83E-09

Final Solver Statistics:
Internal solver steps = 359
Total RHS evals = 405
Total preconditioner setups = 7
Total preconditioner solves = 728
Total nonlinear iterations = 402
Total linear iterations = 364
Average Krylov subspace dimension = 0.9055
Total Convergence Failures - Nonlinear = 0
- Linear = 0
Total number of error test failures = 5
Main solver real/int workspace sizes = 609 144
Linear solver real/int workspace sizes = 534 126
BBD preconditioner real/int workspace sizes = 160 72
Total number of g evals = 14



Preconditioning on right:
t steps fe
--------------------------------
0.100000 221 261
0.200000 265 307
0.300000 290 333
0.400000 306 350
0.500000 319 364
0.600000 329 374
0.700000 339 385
0.800000 345 391
0.900000 352 398
1.000000 359 405
--------------------------------
Max. absolute error is 2.83E-09

Final Solver Statistics:
Internal solver steps = 359
Total RHS evals = 405
Total preconditioner setups = 7
Total preconditioner solves = 728
Total nonlinear iterations = 402
Total linear iterations = 364
Average Krylov subspace dimension = 0.9055
Total Convergence Failures - Nonlinear = 0
- Linear = 0
Total number of error test failures = 5
Main solver real/int workspace sizes = 609 144
Linear solver real/int workspace sizes = 534 126
BBD preconditioner real/int workspace sizes = 160 72
Total number of g evals = 14



76 changes: 76 additions & 0 deletions linux-ubuntu20.04-x86_64/gcc-9.4.0/double/cv_diag_kry_f2003.out
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,76 @@

Diagonal test problem:
neq = 40
nlocal = 10
nprocs = 4
rtol = 1.00E-05
atol = 1.00E-10
alpha = 1.00E+01
ydot_i = -alpha*i * y_i (i = 1,...,neq)
Method is BDF/NEWTON/SPGMR
Diagonal preconditioner uses approximate Jacobian

Preconditioning on left:
t steps fe
--------------------------------
0.100000 221 400
0.200000 265 452
0.300000 290 481
0.400000 306 500
0.500000 319 517
0.600000 329 528
0.700000 339 538
0.800000 346 548
0.900000 351 555
1.000000 355 563
--------------------------------
Max. absolute error is 1.46E-08

Final Solver Statistics:
Internal solver steps = 355
Total RHS evals = 563
Total preconditioner setups = 0
Total preconditioner solves = 727
Total nonlinear iterations = 560
Total linear iterations = 367
Average Krylov subspace dimension = 0.6554
Total Convergence Failures - Nonlinear = 0
- Linear = 0
Total number of error test failures = 5
Main solver real/int workspace sizes = 609 144
Linear solver real/int workspace sizes = 534 126



Preconditioning on right:
t steps fe
--------------------------------
0.100000 221 400
0.200000 265 452
0.300000 290 481
0.400000 306 500
0.500000 319 517
0.600000 329 528
0.700000 339 538
0.800000 345 547
0.900000 352 554
1.000000 358 560
--------------------------------
Max. absolute error is 2.09E-09

Final Solver Statistics:
Internal solver steps = 358
Total RHS evals = 560
Total preconditioner setups = 0
Total preconditioner solves = 730
Total nonlinear iterations = 557
Total linear iterations = 367
Average Krylov subspace dimension = 0.6589
Total Convergence Failures - Nonlinear = 0
- Linear = 0
Total number of error test failures = 5
Main solver real/int workspace sizes = 609 144
Linear solver real/int workspace sizes = 534 126



30 changes: 30 additions & 0 deletions linux-ubuntu20.04-x86_64/gcc-9.4.0/double/cv_diag_non_p_f2003.out
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,30 @@

Diagonal test problem:
neq = 8
nlocal = 2
nprocs = 4
rtol = 1.00E-05
atol = 1.00E-10
alpha = 1.25E+00
ydot_i = -alpha*i * y_i (i = 1,...,neq)
Method is ADAMS/FIXED-POINT

t steps fe
----------------------------
0.100000 14 31
0.200000 22 39
0.300000 29 46
0.400000 37 54
0.500000 44 61
0.600000 51 68
0.700000 59 77
0.800000 66 84
0.900000 74 92
1.000000 81 99
--------------------------------
Max. absolute error is 4.07E-08

Final Solver Statistics:
Internal solver steps = 81
Total RHS evals = 99
Total number of error test failures = 3
68 changes: 68 additions & 0 deletions linux-ubuntu20.04-x86_64/gcc-9.4.0/double/ida_heat2D_kry_bbd_f2003.out
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,68 @@

2D Heat PDE test problem:
nprocs = 4
nx = 100
ny = 100
kx = 1.00
ky = 1.00
rtol = 0.00E+00
atol = 1.00E-03
nxl (proc 0) = 50
nyl (proc 0) = 50


Case 1
Difference quotient half-bandwidths = 50
Retained matrix half-bandwidths = 1

Output Summary
umax = max-norm of solution
nre = nre + nreLS (total number of RES evals.)

t ||u||_max k nst nni nli nre nge h npe nps
------------------------------------------------------------------------------------
0.010000 8.46213E-01 2 15 20 32 20+ 32 816 2.56E-03 8 52
0.020000 7.05517E-01 2 18 24 47 24+ 47 918 5.12E-03 9 71
0.040000 4.83694E-01 3 22 29 73 29+ 73 918 5.12E-03 9 102
0.080000 2.15365E-01 3 28 38 143 38+143 918 6.97E-03 9 181
0.160000 6.15447E-02 2 36 48 211 48+211 1020 1.39E-02 10 259
0.320000 1.63800E-02 2 48 69 393 69+393 1428 1.88E-02 14 462
0.640000 1.06183E-02 1 84 125 603 125+603 2958 5.52E-02 29 728
1.280000 1.98532E-02 1 88 131 621 131+621 3162 2.21E-01 31 752
2.560000 3.00701E-02 1 92 138 651 138+651 3264 3.52E-01 32 789
5.120000 2.46999E-02 1 96 147 670 147+670 3570 1.07E+00 35 817
10.240000 1.82570E-02 1 98 150 682 150+682 3672 2.14E+00 36 832
------------------------------------------------------------------------------------

Final Solver Statistics:
Total number of error test failures = 1
Total number of nonlinear conv. failures = 3
Total number of linear conv. failures = 2

Case 2
Difference quotient half-bandwidths = 5
Retained matrix half-bandwidths = 1

Output Summary
umax = max-norm of solution
nre = nre + nreLS (total number of RES evals.)

t ||u||_max k nst nni nli nre nge h npe nps
------------------------------------------------------------------------------------
0.010000 8.46213E-01 2 15 20 32 20+ 32 96 2.56E-03 8 52
0.020000 7.05517E-01 2 18 24 47 24+ 47 108 5.12E-03 9 71
0.040000 4.83694E-01 3 22 29 73 29+ 73 108 5.12E-03 9 102
0.080000 2.15365E-01 3 28 38 143 38+143 108 6.97E-03 9 181
0.160000 6.15447E-02 2 36 48 211 48+211 120 1.39E-02 10 259
0.320000 1.63800E-02 2 48 69 393 69+393 168 1.88E-02 14 462
0.640000 1.06183E-02 1 84 125 603 125+603 348 5.52E-02 29 728
1.280000 1.98532E-02 1 88 131 621 131+621 372 2.21E-01 31 752
2.560000 3.00701E-02 1 92 138 651 138+651 384 3.52E-01 32 789
5.120000 2.46999E-02 1 96 147 670 147+670 420 1.07E+00 35 817
10.240000 1.82570E-02 1 98 150 682 150+682 432 2.14E+00 36 832
------------------------------------------------------------------------------------

Final Solver Statistics:
Total number of error test failures = 1
Total number of nonlinear conv. failures = 3
Total number of linear conv. failures = 2
48 changes: 48 additions & 0 deletions linux-ubuntu20.04-x86_64/gcc-9.4.0/double/kin_diagon_kry_f2003.out
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,48 @@

Example program kinDiagon_kry_f2003:
This FKINSOL example solves a 128 eqn diagonal algebraic system.
Its purpose is to demonstrate the use of the Fortran interface in
a parallel environment.

Solution method: KIN_none
Problem size: neq = 128
Number of procs: nprocs = 4

1 1.000000 2.000000 3.000000 4.000000
5 5.000000 6.000000 7.000000 8.000000
9 9.000000 10.000000 11.000000 12.000000
13 13.000000 14.000000 15.000000 16.000000
17 17.000000 18.000000 19.000000 20.000000
21 21.000000 22.000000 23.000000 24.000000
25 25.000000 26.000000 27.000000 28.000000
29 29.000000 30.000000 31.000000 32.000000
33 33.000000 34.000000 35.000000 36.000000
37 37.000000 38.000000 39.000000 40.000000
41 41.000000 42.000000 43.000000 44.000000
45 45.000000 46.000000 47.000000 48.000000
49 49.000000 50.000000 51.000000 52.000000
53 53.000000 54.000000 55.000000 56.000000
57 57.000000 58.000000 59.000000 60.000000
61 61.000000 62.000000 63.000000 64.000000
65 65.000000 66.000000 67.000000 68.000000
69 69.000000 70.000000 71.000000 72.000000
73 73.000000 74.000000 75.000000 76.000000
77 77.000000 78.000000 79.000000 80.000000
81 81.000000 82.000000 83.000000 84.000000
85 85.000000 86.000000 87.000000 88.000000
89 89.000000 90.000000 91.000000 92.000000
93 93.000000 94.000000 95.000000 96.000000
97 97.000000 98.000000 99.000000 100.000000
101 101.000000 102.000000 103.000000 104.000000
105 105.000000 106.000000 107.000000 108.000000
109 109.000000 110.000000 111.000000 112.000000
113 113.000000 114.000000 115.000000 116.000000
117 117.000000 118.000000 119.000000 120.000000
121 121.000000 122.000000 123.000000 124.000000
125 125.000000 126.000000 127.000000 128.000000

Final Statistics..

nni = 7 nli = 21
nfe = 8 npe = 2
nps = 28 nlcf = 0

0 comments on commit dd4aebc

Please sign in to comment.