Re: Line 1.3 / minfx issue with 64 bit computers ? -- June 12, 2008

Hi,

This is a simple one.  Minfx-1.0.0 uses Numeric, whereas relax has
switched to numpy.  But I have also rewritten minfx to use numpy
instead of Numeric, but just haven't released minfx-1.0.1 with this
change yet.  If you use the repository version of minfx (see
https://gna.org/projects/minfx), specifically the '/trunk' directory,
then the problem will be solved.  I will try to release this new
version of minfx prior to releasing relax 1.3.1 because of this issue.

Regards,

Edward



On Thu, Jun 12, 2008 at 8:58 PM, Sébastien Morin
<sebastien.morin.1@xxxxxxxxx> wrote:

Hi,

I noticed something which could be a problem for 64 bits users... On x86
(PC), only one test does not pass for line 1.3 (r6355). However, on 64
bits Intel cpu, it seems that 3 tests don't pass, with the two
additional tests not passing as follows :

1.
ERROR: Constrained Newton opt, GMW Hessian mod, backtracking line search
{S2=0.970, te=2048, Rex=0.149}

2.
ERROR: Constrained Newton opt, GMW Hessian mod, More and Thuente line
search {S2=0.970, te=2048, Rex=0.149}

Version of different dependencies are as follows :

========================================
numpy-1.0.4-r2
numeric-24.2-r6
python-2.4.4-r9
scientificpython-2.6.1
========================================


The machine works on Gentoo Linux.

Finally, here is the traceback for the second problematic test (which
yields the same error) :


========================================

Traceback (most recent call last):
 File "/home/semor/relax-1.3/test_suite/system_tests/model_free.py",
line 453, in test_opt_constr_newton_gmw_mt_S2_0_970_te_2048_Rex_0_149
   self.relax.interpreter._Minimisation.minimise('newton', 'gmw', 'mt')
 File "/home/semor/relax-1.3/prompt/minimisation.py", line 358, in minimise
   minimise.minimise(min_algor=min_algor, min_options=min_options,
func_tol=func_tol, grad_tol=grad_tol, max_iterations=max_iterations,
constraints=constraints, scaling=scaling, verbosity=verbosity)
 File "/home/semor/relax-1.3/generic_fns/minimise.py", line 236, in
minimise
   minimise(min_algor=min_algor, min_options=min_options,
func_tol=func_tol, grad_tol=grad_tol, max_iterations=max_iterations,
constraints=constraints, scaling=scaling, verbosity=verbosity)
 File "/home/semor/relax-1.3/specific_fns/model_free/mf_minimise.py",
line 985, in minimise
   results = generic_minimise(func=self.mf.func, dfunc=self.mf.dfunc,
d2func=self.mf.d2func, args=(), x0=param_vector, min_algor=min_algor,
min_options=min_options, func_tol=func_tol, grad_tol=grad_tol,
maxiter=max_iterations, A=A, b=b, full_output=1, print_flag=verbosity)
 File "/home/semor/relax-1.3/minfx/generic.py", line 391, in
generic_minimise
   results = method_of_multipliers(func=func, dfunc=dfunc,
d2func=d2func, args=args, x0=x0, min_options=min_options, A=A, b=b, l=l,
u=u, c=c, dc=dc, d2c=d2c, func_tol=func_tol, grad_tol=grad_tol,
maxiter=maxiter, full_output=full_output, print_flag=print_flag)
 File "/home/semor/relax-1.3/minfx/method_of_multipliers.py", line 143,
in method_of_multipliers
   results = min.minimise()
 File "/home/semor/relax-1.3/minfx/method_of_multipliers.py", line 397,
in minimise
   results = self.generic_minimise(func=self.func_LA,
dfunc=self.func_dLA, d2func=self.func_d2LA, args=self.args, x0=self.xk,
min_algor=self.min_algor, min_options=self.min_options, func_tol=None,
grad_tol=self.tk, maxiter=maxiter, full_output=1,
print_flag=sub_print_flag, print_prefix="\t")
 File "/home/semor/relax-1.3/minfx/generic.py", line 322, in
generic_minimise
   results = newton(func=func, dfunc=dfunc, d2func=d2func, args=args,
x0=x0, min_options=Issue on min_options, func_tol=func_tol,
grad_tol=grad_tol, maxiter=maxiter, full_output=full_output,
print_flag=print_flag, print_prefix=print_prefix)
 File "/home/semor/relax-1.3/minfx/newton.py", line 45, in newton
   results = min.minimise()
 File "/home/semor/relax-1.3/minfx/base_classes.py", line 234, in minimise
   self.new_param_func()
 File "/home/semor/relax-1.3/minfx/newton.py", line 164, in new_param_func
   self.pk = self.get_pk()
 File "/home/semor/relax-1.3/minfx/base_classes.py", line 643, in gmw
   return gmw(self.dfk, self.d2fk, self.I, self.n, self.mach_acc,
self.print_prefix, self.print_flag, return_matrix)
 File "/home/semor/relax-1.3/minfx/hessian_mods/gmw81.py", line 151, in gmw
   y = solve_linear_equations(L, dfk)
 File "/usr/lib64/python2.4/site-packages/Numeric/LinearAlgebra.py",
line 88, in solve_linear_equations
   t =_commonType(a, b)
 File "/usr/lib64/python2.4/site-packages/Numeric/LinearAlgebra.py",
line 30, in _commonType
   t = a.typecode()
AttributeError: 'numpy.ndarray' object has no attribute 'typecode'

========================================


As these tests both use the minfx package, the issue probably comes from
minfx... If you think it is the case, please feel free to forward my
mail on their mailing list...


Thanks for help !


Séb  :)

_______________________________________________
relax (http://nmr-relax.com)

This is the relax-devel mailing list
relax-devel@xxxxxxx

To unsubscribe from this list, get a password
reminder, or change your subscription options,
visit the list information page at
https://mail.gna.org/listinfo/relax-devel

Re: Line 1.3 / minfx issue with 64 bit computers ?

Header

Content

Related Messages