{SPAM}? Re: Re: [molpro-user] gmainv1 failure
csong
csong at ustc.edu.cn
Mon Mar 3 09:19:26 GMT 2008
Dear Krik and Li Xi
Thanks for your reply
As you suggested, I have decreased the size of the shmmax kernel parameter, but it
still can not works. The molpro was compiled with I8 and swap we used was about
10GB, similar to the physical memory.
(use free command under linux give message:
total used free shared buffers cached
Mem: 12334648 418200 11916448 0 106504 128548
-/+ buffers/cache: 183148 12151500
Swap: 10482404 4180 10478224)
So is there any other possible solution with this bothersome problem?
Song
>-------------------------------------------------------------
>发件人：Kirk Peterson
>发送日期：2008-03-03 16:27:22
>收件人：csong
>抄送：molpro-user molpro-user
>主题：Re: [molpro-user] gmainv1 failure
>
>Song,
>
>I assume you've compiled molpro with I*8 integers? I would recommend
>decreasing the size of the shmmax kernel parameter. This is a segment
>size, not a total allowed limit.
>
>as root:
>
>echo '33554432' > /proc/sys/kernel/shmmax
>
>How much swap do you have on this system? I think Linux can be fussy
>if you don't have at least as much swap as you have physical memory.
>
>-Kirk
>
>On Mar 2, 2008, at 5:59 PM, csong wrote:
>
>> Dear andy
>>
>> Thanks for your reply
>>
>> I have run molpro with lower memory as you suggested, and the
>> details are as
>> following:
>>
>> if memory option adopt default value
>> Variable memory set to 8000000 words, buffer space 230000 words
>> the calculation stops in a few minutes at HF calculation
>> error in output file:
>> "NOT ENOUGH MEMORY IN SUBROUTINE AOSORT.
>> ERROR EXIT
>> CURRENT STACK: MAIN"
>>
>> if memory,480,M option was used
>> the calculation stops at CCSD step
>> error in output file:
>> "Multipassing necessary in transformation. To avoid, increase memory
>> by
>>
>> 1167114599 words
>> NPASS IN TRSOR3 3
>> insufficient memory available - require 96835362 have
>> 84660125
>> the request was for real words
>> ERROR EXIT
>> CURRENT STACK: MAIN"
>>
>> if memory,300,M option was used
>> the calculation stops at CCSD step
>> error in output file:
>> "Multipassing necessary in transformation. To avoid, increase memory
>> by
>> 1347114599 words
>> NPASS IN TRSOR3 5
>> insufficient memory available - require 193670724 have
>> 98359007
>> the request was for real words
>> ERROR EXIT
>> CURRENT STACK: MAIN
>> "
>>
>> so apparently, our ccsd calculation need more memory and the memory
>> option must be
>> defined
>> but the system have 12GB ~98304MW memory
>>
>> When we define more than 480MW memory for the calculation
>> run molpro will give "gmainv1 failure to allocate XXXXXX" in linux
>> and the outpue file show error:
>> "Failure in attempting memory allocation of XXXXXX words (XXXX Mbyte)
>> This error has been generated by the operating system,
>> and may be the result of insufficient system memory or paging space
>> In order to avoid the problem in the MOLPRO context,
>> consider also reducing the requested memory through
>> the MEMORY input command, or the -m command line option
>> gmainv1 failure to allocate XXXXXX
>>
>> ERROR EXIT
>> CURRENT STACK: MAIN
>>
>>
>>
>> **********************************************************************************
>> *
>> *******************************
>> insufficient memory available - require XX
>> have 0
>> the request was for real words
>>
>> RECURSIVE CALL OF FEHLER. EMERGENCY STOP.
>> FILE INFORMATION MAY BE LOST
>>
>> RECURSIVE CALL OF ERROR HANDLER, FILE INFORMATION MAY BE LOST
>> "
>>
>> but when the memory for molpro set to 11GB
>> the error in output file change to
>> "
>> GETMEM ERROR IN GETMEM iMap 1 ALLO
>>
>> Memory allocation map
>> ---------------------
>>
>> FldNam VarTyp ID Length iPos
>> - - - - - - - - - - - - - - - - - - - - - - - -
>> IGATIM REAL 377538531 20 188769266
>> IGASTAT REAL 377538573 7 188769287
>> INFO REAL 377538589 1832 188769295
>> HERR REAL 377542255 21 188771128
>> HERW REAL 377542299 21 188771150
>> TMAX REAL 377542343 3 188771172
>> DDX REAL 377542351 3 188771176
>> IMAP INTE 0 4851 0
>>
>> STACK TRACE:
>>
>> SEWARD 0.00 0.00
>> INPUT 0.00 0.00
>>
>> ERROR EXIT
>> CURRENT STACK: MAIN
>>
>> "
>>
>> So is this error still caused by the linux system or any unproper
>> set in molpro
>> parameters?
>>
>>
>> ------------------------------
>> Song
>>
>>> -------------------------------------------------------------
>>>
>>> Song,
>>>
>>> Can I just confirm that the program works if you don't specify a
>>> limit,
>>> for some simple calculation?
>>>
>>> Also, do you think the job will really need 500Mb of shared memory?
>>> What
>>> happens if you run the calculation with a lower amount?
>>>
>>> Best wishes,
>>>
>>> Andy
>>>
>>> csong wrote:
>>>> Dear Andy,
>>>> Many thanks for your quick reply.
>>>> I checked the value in /proc/sys/kernel/shmmax
>>>> $ more /proc/sys/kernel/shmmax
>>>> 68719476736
>>>>
>>>> It seems to be no limit. And I change it larger in /etc/
>>>> sysctl.conf,but it's also no use.
>>>>
>>>> Best wishes,
>>>> Song
>>>>
>>>>> -------------------------------------------------------------
>>>>> Andy May
>>>>>
>>>>> Song,
>>>>>
>>>>> I think the limit can be found in:
>>>>>
>>>>> /proc/sys/kernel/shmmax
>>>>>
>>>>> The standard limit appears to be 32 Mb.
>>>>>
>>>>> Best wishes,
>>>>>
>>>>> Andy
>>>>>
>>>>> csong wrote:
>>>>>> Hi,
>>>>>> I have a linux machine(RHEL5 system,Intel(R) Xeon(R) CPU E5335)
>>>>>> with 12GB of memory and I'm trying to execute a job but I
>>>>>> can't. The error message is the following:
>>>>>>
>>>>>> Failure in attempting memory allocation of 500230002 words (3816
>>>>>> Mbyte)
>>>>>> This error has been generated by the operating system,
>>>>>> and may be the result of insufficient system memory or paging
>>>>>> space
>>>>>> In order to avoid the problem in the MOLPRO context,
>>>>>> consider also reducing the requested memory through
>>>>>> the MEMORY input command, or the -m command line option
>>>>>> gmainv1 failure to allocate 500230002
>>>>>>
>>>>>> I already had done some search for similar issues and I had
>>>>>> found that a possible problem is the datasize limits, but the
>>>>>> problem stills even after changing these values to unlimited.
>>>>>> And I had update the kernel to the latest stable
>>>>>> version(2.6.24.3), but it is also no use.
>>>>>> # ulimit -a
>>>>>> core file size (blocks, -c) 0
>>>>>> data seg size (kbytes, -d) unlimited
>>>>>> max nice (-e) 0
>>>>>> file size (blocks, -f) unlimited
>>>>>> pending signals (-i) 106496
>>>>>> max locked memory (kbytes, -l) 32
>>>>>> max memory size (kbytes, -m) unlimited
>>>>>> open files (-n) 1024
>>>>>> pipe size (512 bytes, -p) 8
>>>>>> POSIX message queues (bytes, -q) 819200
>>>>>> max rt priority (-r) 0
>>>>>> stack size (kbytes, -s) 8192
>>>>>> cpu time (seconds, -t) unlimited
>>>>>> max user processes (-u) 106496
>>>>>> virtual memory (kbytes, -v) unlimited
>>>>>> file locks (-x) unlimited
>>>>>>
>>>>>>
>>>>>> Would you help me to solve it? Thanks.
>>>>>>
>>>>>> Song
>>>>>>
>>>>>>
>>>>
>>
>
More information about the Molpro-user
mailing list