Re: [phenixbb] runtime error in phenix.refine

Hi Nat and Pavel, Here is the output of "boost_adaptbx.show_platform_info". The inputs and logs I will send off-list. It's a reasonably large cell (119.84, 163.9, 178.39, 90, 94.2, 90) with over 64,000 atoms. Refinement proceeds fine in REFMAC and BUSTER, if that helps. __FILE__ = /net/anaconda/scratch1/phenix/phenix-1.8-1069/cctbx_project/boost_adaptbx/meta_ext.cpp __DATE__ = Jun 2 2012 __TIME__ = 07:34:31 __x86_64__ __linux __GNUC__ = 3 __GNUC_MINOR__ = 4 __GNUC_PATCHLEVEL__ = 2 boost::python::cxxabi_cxa_demangle_is_broken(): false __GXX_WEAK__ = 1 __LP64__ = 1 __VERSION__ = 3.4.2 20041017 (Red Hat 3.4.2-6.fc3) FE_INEXACT = 32 FE_DIVBYZERO = 4 FE_UNDERFLOW = 16 FE_OVERFLOW = 8 FE_INVALID = 1 FE_ALL_EXCEPT = 61 __SSE2__ = 1 BOOST_LIB_VERSION = 1_48 PY_VERSION = 2.7.2 PYTHON_API_VERSION = 1013 sizeof(bool) = 1 sizeof(short) = 2 sizeof(int) = 4 sizeof(long) = 8 sizeof(std::size_t) = 8 sizeof(void*) = 8 sizeof(long long) = 8 sizeof(float) = 4 sizeof(double) = 8 sizeof(long double) = 16 sizeof(boost::int32_t) = 4 sizeof(boost::uint32_t) = 4 sizeof(wchar_t) = 4 __PTRDIFF_TYPE__ Py_USING_UNICODE sizeof(PY_UNICODE_TYPE) = 2 gnu libc version: 2.12 BOOST_ADAPTBX_TYPE_ID_SIZE_T_EQ_UNSIGNED_LONG boost/TAG: 73458 os.name: posix sys.platform: linux2 sys.byteorder: little platform.platform(): Linux-2.6.32-220.13.1.el6.x86_64-x86_64-with-redhat-6.2-Santiago platform.architecture(): ('64bit', 'ELF') floating_point_exceptions.division_by_zero_trapped: False floating_point_exceptions.invalid_trapped: False floating_point_exceptions.overflow_trapped: False number of processors: 24 Memory total: 203,085,643,776 Memory free: 54,516,293,632 import thread: OK Division operator semantics: true division "hello" = ['h', 'e', 'l', 'l', 'o'] u"hello" = ['h', '\x00', 'e', '\x00', 'l', '\x00', 'l', '\x00', 'o', '\x00'] u"\u00C5" = ['\xc5', '\x00'] as utf-8 = ['\xc3', '\x85'] LATIN CAPITAL LETTER A WITH RING ABOVE = Å u+s,u,s: 0.14 0.10 0.04 micro-seconds/tick: 0.211 On 11 June 2012 20:05, Pavel Afonine wrote:

Hi Eugene,

if you send me the necessary inputs (off-list) I'll try to reproduce the problem and see what generates it ...

Pavel

 On 6/11/12 11:44 AM, Nathaniel Echols wrote:

...

On Sun, Jun 10, 2012 at 10:04 AM, Eugene Valkov  wrote:

...

I'm encountering the following during the second macro-cycle in phenix.refine in latest 1.8-1069 dev build for Linux 64-bit:

RuntimeError: Excessive number of cubicles:  This may be due to unreasonable parameters:    cubicle_edge=5    space_span=(8467.78,8440.85,8372.67)    n_cubicles=(1694,1689,1675)    max_number_of_bytes=101542821888

Any thoughts/suggestions much appreciated.

If I'm interpreting the code correctly, this means that the number of cubicles will lead to an integer format overflow - which I find hard to believe if you're using a 64-bit system.  Could you please send me the rest of the error message, along with the output of the command "boost_adaptbx.show_platform_info"?

thanks, Nat _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb