[mpich-discuss] error :Assertion failed in file ch3u_buffer.c at line 77: 0 memcpy argument memory ranges overlap

[loc duong ding]

Thank you for your suggestions. Finally, I succeeded to compile mpich with configure option: configure --prefix=/home/users/dinhloc1984/mpich2 --enable-fc --enable-f77 --with-device=ch3:nemesis:m FC=gfortran F77=gfortran CC=gcc.

But I have problem when I run the mpiexec. I got the error: 

MX:node1:mx_init:querying driver:error 5(errno=2):No MX device entry in /dev.


I try to looking for some instruction in manual, but I can not see it. 

> 
> As for the distinct buffers, just don't try to send and 
receive from the same buffer at the same time, it simply doesn't make 
sense (the receive can corrupt the send data). 

For the buffer problem, I can not manage the code in the OPENMX package. Is there possible to solve this problem by mpich2 set up? 


Thank you so much for your help. 


Best,

-----------------------------------------------
Loc Duong Dinh
Ms-Ph.D Student
Sungkyunkwan Advanced Institute of Nanotechnology,
Sungkyunkwan University,
Suwon, 440-746, Korea 
Email: mambom1902 at yahoo.com, ddloc at skku.edu


>________________________________
> From: Dave Goodell <goodell at mcs.anl.gov>
>To: loc duong ding <mambom1902 at yahoo.com> 
>Cc: "mpich-discuss at mcs.anl.gov" <mpich-discuss at mcs.anl.gov> 
>Sent: Wednesday, June 27, 2012 1:17 PM
>Subject: Re: [mpich-discuss] error :Assertion failed in file ch3u_buffer.c at line 77: 0 memcpy argument memory ranges overlap
> 
>MPD is deprecated and no longer built by default: http://wiki.mcs.anl.gov/mpich2/index.php/Frequently_Asked_Questions#Q:_I_don.27t_like_.3CWHATEVER.3E_about_mpd.2C_or_I.27m_having_a_problem_with_mpdboot.2C_can_you_fix_it.3F
>
>Pass "FC=gfortran" and "F77=gfortran" to your configure command to use it instead of fort.
>
>Alternatively, consider using one of the many binary distributions of MPICH2.  It does not sound like you are comfortable working with MPICH2 or its build system, and the binary distributions are usually easier to install and use.
>
>-Dave
>
>On Jun 27, 2012, at 12:46 PM MDT, loc duong ding wrote:
>
>> I tried to install mpich2-1.4p1 with configure --prefix=/home/users/dinhloc1984/mpich2 --enable-fc --enable-f77 --with-device=ch3:nemesis:mx
>> THere is no error after make and make install. However, I can not find mpd, mpdboot, ... in  /home/users/dinhloc1984/mpich2/bin. Is this strange? 
>> 
>> Another thing is I would like to compile mpif90 by gfortran but I can not set up. IT is always recognize ifort. Please give me some hints to solve this. 
>>  
>> Best regards, 
>> -----------------------------------------------
>> Loc Duong Dinh
>> Ms-Ph.D Student
>> Sungkyunkwan Advanced Institute of Nanotechnology,
>> Sungkyunkwan University,
>> Suwon, 440-746, Korea 
>> Email: mambom1902 at yahoo.com, ddloc at skku.edu
>> From: Dave Goodell <goodell at mcs.anl.gov>
>> To: loc duong ding <mambom1902 at yahoo.com> 
>> Cc: mpich <mpich-discuss at mcs.anl.gov> 
>> Sent: Wednesday, June 27, 2012 8:45 AM
>> Subject: Re: [mpich-discuss] error :Assertion failed in file ch3u_buffer.c at line 77: 0 memcpy argument memory ranges overlap
>> 
>> [please keep mpich-discuss@ CCed]
>> 
>> Please update to mpich2-1.4.1p1 or later and retry.  Chasing down bugs in old versions is usually a big waste of time for everybody.
>> 
>> As for the distinct buffers, just don't try to send and receive from the same buffer at the same time, it simply doesn't make sense (the receive can corrupt the send data).
>> 
>> -Dave
>> 
>> On Jun 27, 2012, at 9:40 AM MDT, loc duong ding wrote:
>> 
>> > Dear Dave,
>> > 
>> > I am using version mpich2-1.3.1 and compiled with gcc 4.1. 
>> > I installed this version quite long time ago. I can not remember I used the option you mentioned or not. Could you give more detail instructions how to use distinct buffer? 
>> > 
>> > I will try to compile the newest one with your recommendation. 
>> > 
>> > Thank you for your instructions. 
>> > 
>> > Best wishes, 
>> > -----------------------------------------------
>> > Loc Duong Dinh
>> > Ms-Ph.D Student
>> > Sungkyunkwan Advanced Institute of Nanotechnology,
>> > Sungkyunkwan University,
>> > Suwon, 440-746, Korea 
>> > Email: mambom1902 at yahoo.com, ddloc at skku.edu
>> > From: Dave Goodell <goodell at mcs.anl.gov>
>> > To: loc duong ding <mambom1902 at yahoo.com>; mpich-discuss at mcs.anl.gov 
>> > Sent: Wednesday, June 27, 2012 8:18 AM
>> > Subject: Re: [mpich-discuss] error :Assertion failed in file ch3u_buffer.c at line 77: 0 memcpy argument memory ranges overlap
>> > 
>> > What version of MPICH2 are you using?
>> > 
>> > Are you configuring with "--with-device=ch3:nemesis:mx"?
>> > 
>> > It looks like you are doing a send/recv to/from the same process with the same buffer. It's OK to send/recv to/from the same process, but you must use distinct buffers.
>> > 
>> > -Dave
>> > 
>> > On Jun 27, 2012, at 5:49 AM MDT, loc duong ding wrote:
>> > 
>> > > Dear MPICH2 developers,
>> > > 
>> > > I have the problem when using MPICH2. I tried compile with OPENMX code and it is fine in compiling. However, there is a problem when I run the package using mpich2: 
>> > > 
>> > > Assertion failed in file ch3u_buffer.c at line 77: 0
>> > > memcpy argument memory ranges overlap
>> > >  
>> > > I try to search the problem from internet but I can not find the solution for this. 
>> > > 
>> > > Could you give some instructions to solve this problem?
>> > > 
>> > > Thank you for all your help. 
>> > > 
>> > > With best wishes, 
>> > > -----------------------------------------------
>> > > Loc Duong Dinh
>> > > Ms-Ph.D Student
>> > > Sungkyunkwan Advanced Institute of Nanotechnology,
>> > > Sungkyunkwan University,
>> > > Suwon, 440-746, Korea 
>> > > Email: mambom1902 at yahoo.com, ddloc at skku.edu
>> > > _______________________________________________
>> > > mpich-discuss mailing list    mpich-discuss at mcs.anl.gov
>> > > To manage subscription options or unsubscribe:
>> > > https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
>> > 
>> > 
>> > 
>> 
>> 
>> 
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20120628/b540d2f5/attachment.html>