#2858: Compiling CarpetX: issues with PDESolvers
Reporter: Alejandra Gonzalez
Status: submitted
Milestone:
Version:
Type: bug
Priority: major
Component: EinsteinToolkit thorn
Hello, while trying to compile ETK with CarpetX I get the following error:
```
COMPILING CarpetX/PDESolvers/src/pdesolvers.cxx
/gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/arrangements/CarpetX/Arith/src/spvect.hxx(169): error: calling a __host__ function from a __host__ __device__ function is not allowed
const auto &lastidx = elts.back().first;
^
detected during instantiation of "void Arith::spvect<I, T, V>::emplace_back(const I &, const T &) [with I=std::tuple<int, int>, T=CCTK_REAL8, V=Arith::vector1]" at line 183 of /gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/configs/sim_cx/build/PDESolvers/pdesolvers.cxx
/gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/arrangements/CarpetX/Arith/src/spvect.hxx(172): error: calling a __host__ function from a __host__ __device__ function is not allowed
elts.back().second += val;
^
detected during instantiation of "void Arith::spvect<I, T, V>::emplace_back(const I &, const T &) [with I=std::tuple<int, int>, T=CCTK_REAL8, V=Arith::vector1]" at line 183 of /gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/configs/sim_cx/build/PDESolvers/pdesolvers.cxx
```
Important to know is that I’m compiling with CUDA and added the necessary modifications to my cfg file \(attached\).
These are also the modules I’m using while compiling:
- oneapi/2023.2.0 which includes: intel/2023.2.0, impi/2021.10.0, mkl/2023.2.0, and ucx/1.15.0
- cuda/12.2
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2858/compiling-carpetx…
#2857: Compiling CarpetX with OpenPMD
Reporter: Alejandra Gonzalez
Status: open
Milestone: ET_2024_11
Version:
Type: bug
Priority: major
Component: EinsteinToolkit thorn
Comment (by Roland Haas):
New ticket, please, so that this ticket can be used to keep track of a solution that will mean including the required files in ExternalLibraries-openPMD same as we do for NSIMD.
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2857/compiling-carpetx…
#2857: Compiling CarpetX with OpenPMD
Reporter: Alejandra Gonzalez
Status: open
Milestone: ET_2024_11
Version:
Type: bug
Priority: major
Component: EinsteinToolkit thorn
Comment (by Alejandra Gonzalez):
Actually building with --reconfig seemed to do the trick!
But now I get another error unrelated to this ticket, should I open a new ticket or just post it here? \(It has to do with some functions under CarpetX/PDESolvers\)
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2857/compiling-carpetx…
#2857: Compiling CarpetX with OpenPMD
Reporter: Alejandra Gonzalez
Status: submitted
Milestone: ET_2024_11
Version:
Type: bug
Priority: major
Component: EinsteinToolkit thorn
Comment (by Roland Haas):
Note that when updating the `OptionList` you often have to build \(once\) with `--reconfig` to make sure setting are correctly or even \(dependency tracking gets fuzzy with options\) have to do a
```
touch arrangements/ExternalLibraries/openPMD/*.ccl arrangements/ExternalLibraries/openPMD/src/*
```
to trigger options being correctly transferred to `make.OPENPMD_API.defn`.
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2857/compiling-carpetx…
#2857: Compiling CarpetX with OpenPMD
Reporter: Alejandra Gonzalez
Status: submitted
Milestone: ET_2024_11
Version:
Type: bug
Priority: major
Component: EinsteinToolkit thorn
Comment (by Roland Haas):
Hmm, could you provide the content of the files
* `configs/sim/bindings/Configuration/Capabilities/make.OPENPMD_API.defn`
* `configs/sim/bindings/Configuration/Capabilities/make.CARPETX.defn`
?
Those contain the values used by the build system. The first one to see what OpenPMD recorded and the second one to make sure CarpetX is actually using OpenPMD’s settings.
Also if you could do a:
```
./simfactory/bin/sim build -j1
```
which will fail and then another build run:
```
VERBOSE=yes ./simfactory/bin/sim build -j1 &>make.log
```
then attach the file `make.log` to this ticket \(the “Attach file” option is hidden behind the “More v” button near the top right of the screen\), that would help.
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2857/compiling-carpetx…
#2857: Compiling CarpetX with OpenPMD
Reporter: Alejandra Gonzalez
Status: submitted
Milestone: ET_2024_11
Version:
Type: bug
Priority: major
Component: EinsteinToolkit thorn
Dear ETK developers,
While attempting to compile CarpetX with the most recent version of ETK,
everything seemed to go smoothly except when it
tries to build the external library openPMD. Apparently the library
tries to fetch a github repository during compilation, e.g.
nlohmann\_json, which is not possible to do on the cluster I'm trying
to do it since it has no connection to the outside.
Naturally, I downloaded the repository locally, sent it to my cluster
manually and set the path on my cluster optionlist file
`simfactory/mdb/optionlists/marenostrum5.cfg`:
openPMD\_json\_src = path/to/my/local/repo
However during compilation it still attempts to clone the repo and I
still fail to find the correct place to set this path. For this reason I compiled
OpenPMD separately on my system and added the corresponding paths:
```
OPENPMD_DIR = /home/uib/uib416720/ETK2024/CarpetX/openPMD-api_build
OPENPMD_LIB_DIRS = /home/uib/uib416720/ETK2024/CarpetX/openPMD-api_build/lib
OPENPMD_INC_DIRS = /home/uib/uib416720/ETK2024/CarpetX/openPMD-api_build/openPMD-api/include
openPMD_SOURCE_DIR = /home/uib/uib416720/ETK2024/CarpetX/openPMD-api_build/openPMD-api
```
to simfactory/mdb/optionlists/marenostrum5.cfg which I see during compilation that simfactory recognizes them:
```
Setting OPENPMD_DIR to '/home/uib/uib416720/ETK2024/CarpetX/openPMD-api_build'
Setting OPENPMD_LIB_DIRS to '/home/uib/uib416720/ETK2024/CarpetX/openPMD-api_build/lib'
Setting OPENPMD_INC_DIRS to '/home/uib/uib416720/ETK2024/CarpetX/openPMD-api_build/openPMD-api/include'
Setting openPMD_SOURCE_DIR to '/home/uib/uib416720/ETK2024/CarpetX/openPMD-api_build/openPMD-api'
End of options from '/home/uib/uib416720/ETK2024/CarpetX/Cactus/configs/sim_cx/OptionList'.
```
Unfortunately I still get an error:
```
COMPILING CarpetX/CarpetX/src/io_openpmd.cxx
/gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/configs/sim_cx/build/CarpetX/io_openpmd.cxx:17:10: fatal error: openPMD/openPMD.hpp: No such file or directory
17 | #include <openPMD/openPMD.hpp>
| ^~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[3]: *** [/gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/configs/sim_cx/config-data/make.config.rules:318: io_openpmd.cxx.o] Error 1
make[2]: *** [/gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/lib/make/make.thornlib:113: make.checked] Error 2
make[1]: *** [/gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/lib/make/make.configuration:179: /gpfs/home/uib/uib416720/ETK2024/CarpetX/Cactus/configs/sim_cx/lib/libthorn_CarpetX.a] Error 2
```
Even though the paths I passed are correct and the hpp file does exist there. I did a small test script where I added `#include <openPMD/openPMD.hpp>` and worked without problems, which makes me guess it’s an issue within simfactory or CarpetX itself.
I would really appreciate some guidance or suggestions about what I could do. The system is MareNostrum 5, which is relatively new.
Thanks in advance.
Sincerely,
Alejandra
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2857/compiling-carpetx…
#2845: support (again) filesystem in std::experimental in Silo output
Reporter: Roland Haas
Status: open
Milestone:
Version:
Type: enhancement
Priority: major
Component: EinsteinToolkit thorn
Comment (by Roland Haas):
Unless objected I will apply this after 2025-03-07
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2845/support-again-fil…
#2172: Test "Binary black hole GW150914" example
Reporter: Roland Haas
Status: open
Milestone: ET_2024_11
Version: development version
Type: task
Priority: major
Component: EinsteinToolkit website
Comment (by Roland Haas):
@{557058:f7fd5133-6eee-4385-a5e5-3e03342a0b24} could you update the web-page and the downloads section, please?
--
Ticket URL: https://bitbucket.org/einsteintoolkit/tickets/issues/2172/test-binary-black…