Dear Florian and David,
I may know the solution. I also encountered the same error yesterday but solved it by myself.
I guess you have already installed boost by using Homebrew. However, the latest boost might cause the error on the compilation of moft and metatwist.
The workaround is to unlink boost temporarily before the installation of AmberTools21 (`brew unlink boost`), then retry cmake. After installation of AmberTools21, I relinked boost.
Here is my procedure.
```
brew install wget cmake gcc open-mpi
ln -sf /opt/homebrew/bin/gfortran-10 /opt/homebrew/bin/gfortran # for M1 Mac
brew unlink boost # if present
cd amber20_src/build
AMBER_PREFIX=$(dirname $(dirname `pwd`))
$AMBER_PREFIX/amber20_src \
-DCMAKE_INSTALL_PREFIX=$AMBER_PREFIX/amber20 \
-DCOMPILER=CLANG -DBLA_VENDOR=Apple \
-DMPI=TRUE -DBUILD_GUI=FALSE -DCUDA=FALSE -DINSTALL_TESTS=TRUE \
-DDOWNLOAD_MINICONDA=TRUE -DMINICONDA_USE_PY3=TRUE 2>&1 | tee cmake.log
make -j8 && make install
brew link boost
```
On 2021/05/02 1:51, David A Case wrote:
> On Sat, May 01, 2021, Florian Nachon wrote:
>>
>> I tried to install Ambertools21 using cmake, but it fails at the linking
>> stage, much like Ambertools20 was doing (see below).
>
> So, I gather this not a regression, but that the same thing happens with
> AmberTools20; is that correct?
>
> Can you post your cmake.log file? The intent is that any external boost
> libraries (e.g. from homebrew) should be ignored. But the log file will
> provide more details.
>
> As a simple workaround, you can add -DDISABLE_TOOLS="moft" to your run_cmake
> script. I suspect that few AmberTools users actually use metatwist, but of
> course you may be one of them....
>
> We have some people testing on the new arm64 cpus with OSX, and know that
> it's possible to get everything working there. But, as with Intel Macs, the
> number of possible combinations of OSX version, XCode version, installation
> of external libraries, and versions of gfortran, is somewhat overwhelming.
> It you are willing to try some things and report results, that might help
> other users.
>
> ...thx....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun May 02 2021 - 00:30:02 PDT
