Re: [Apbs-users] Memory allocation error
Biomolecular electrostatics software
Brought to you by:
sobolevnrm
From: Baker, N. <Nat...@pn...> - 2011-06-03 14:37:17
|
Hello – Is this 240 GB of shared memory? If so, are you using a 64-bit executable that can address all of that memory? Thanks, __________________________________________________ Nathan Baker Pacific Northwest National Laboratory Tel: 509-375-3997 http://kdi.pnnl.gov/bios/baker.stm From: LUQUE BUZO, DANIEL [mailto:dl...@cn...] Sent: Thursday, June 02, 2011 11:56 PM To: Baker, Nathan Cc: apb...@li... Subject: Re: [Apbs-users] Memory allocation error Hello- Yes, the machine has >250 Gb of memory that is much more that the 24 Gb required. Thanks. Dani Quoting "Baker, Nathan": > Hello – > > The more relevant number (24372.754 MB) is in the output below. The > 600487976 B was simply the last thing the program tried to allocate > before it ran out of memory. Does your machine have >24 GB of memory? > > Thanks, > __________________________________________________ > Nathan Baker > Pacific Northwest National Laboratory > Tel: 509-375-3997 > http://kdi.pnnl.gov/bios/baker.stm > > From: LUQUE BUZO, DANIEL [mailto:dl...@cn...]<mailto:[mailto:dl...@cn...]> > Sent: Thursday, June 02, 2011 11:15 AM > To: apb...@li...<mailto:apb...@li...> > Subject: [Apbs-users] Memory allocation error > > Hello, > > I am trying to calculate the electrostatic potential of a quite huge > pqr file and when I try to run apbs I receive a memory allocation > error: > > Unable to allocate memory! > (This often means you don't have enough memory available for this > calculation.) > FATAL ERROR: Unable to allocate enough memory for problem size. > FATAL ERROR: You requested 600487976 bytes of RAM. > FATAL ERROR: Check that your system has that much memory. > VASSERT: ASSERTION FAILURE! filename vmem.c, line 252, (ram != ((void *)0)) > Abort > > The surprising thing is that 600487976 bytes corresponds to ~572 Mb, > that seems to be not a really hight number. In order to be sure of > the requested memory I have ran psize.py with the pqr file with this > result: > > ################# MOLECULE INFO #################### > Number of ATOM entries = 163619 > Number of HETATM entries (ignored) = 0 > Total charge = -16.006 e > Dimensions = 249.520 x 285.122 x 169.285 A > Center = 36.079 x 0.000 x 76.197 A > Lower corner = -88.681 x -142.561 x -8.446 A > Upper corner = 160.839 x 142.561 x 160.839 A > > ############## GENERAL CALCULATION INFO ############# > Coarse grid dims = 424.184 x 484.707 x 287.784 A > Fine grid dims = 269.520 x 305.122 x 189.285 A > Num. fine grid pts. = 545 x 609 x 385 > Parallel solve required (24372.754 MB > 400.000 MB) > Total processors required = 288 > Proc. grid = 8 x 6 x 6 > Grid pts. on each proc. = 97 x 129 x 129 > Fine mesh spacing = 0.416569 x 0.476009 x 0.295296 A > Estimated mem. required for parallel solve = 307.880 MB/proc. > Number of focusing operations = 3 > > ################# ESTIMATED REQUIREMENTS #################### > Memory per processor = 307.880 MB > Grid storage requirements (ASCII) = 42561.308 MB > > > I understand that I need ~24 Gb of RAM memory to run the process. The > surprising thing is that I am running the process in a machine with > >250 Gb of shared memory and, thus, this requirement of 24 Gb of > memory sound reasonable. > > I have tried to re-compile apbs with the maximum memory option (in > the configure step) set to 200 Gb with the same results of memory > allocation error. > > Is there any awy to fix this problem without run it in parallel mode? > > Kind Regards, > > Dani > |