AMBER Archive (2009)
Subject: Re: 回复： [AMBER] errors of PMEMD installation
From: Mengjuei Hsieh (mjhsieh_at_gmail.com)
Date: Fri Jul 24 2009 - 03:41:05 CDT
You indeed should use ia32 version of the compiler option instead of
em64t. First that's probably the reason that ifort couldn't find the
path of the header files. Second you won't be able to run em64t on
your system anyway.
i.e., please try "./configure linux_p4 ifort lam" to generate
corresponding config.h file.
If you continue to have the same problem after switching to everything
ia32 in the PMEMD configuration, try to replace every "ifort" in the
config.h with "env LAMHF77=ifort mpif77".
On Thu, Jul 23, 2009 at 6:27 PM, 廖青华<fantastic_0919_at_yahoo.com.cn> wrote:
> Dear Mengjuei,
> Firstly, thanks for your reply!
> The parallel version of amber10 has been installed successfully, and it runs correctly.
> When install PMEMD, what I do is just following the installation guide.
> first, run ./configure -help to get the information of the operating sytem and other things.
> second, run ./configure linux_em64t ifort lam to get the file config.h
> finally, run make install, but what I got were those errors I showed in the last mail.
> By the way, the hardware of our computer is 64, but the operating system of linux is 32, and
> also the intel compiler of ifort is 32. Did this result in this error?
> Thanks very much!
> Qinghua Liao
> 发件人： Mengjuei Hsieh <mjhsieh_at_gmail.com>
> 已发送： 2009/7/24(周五), 上午3:25:27
> Dear Qinghua,
> This is very interesting if you are sure that your lammpi was
> installed correctly. And no, the suggestion you mentioned previously
> is not immediately helpful, since it's for openmpi. Could you tell us
> a little more about how did you run the ./configure script?
> On Wed, Jul 22, 2009 at 7:28 AM, 廖青华<fantastic_0919_at_yahoo.com.cn> wrote:
>> Hi amber users,
>> I encountered the following errors when installed PMEMD in amber10.The MPI software
>> I use is Lam in amber10, the compiler is intel ifort, and the operating system is Red Hat 5.
>> [log skipped]
AMBER mailing list