[petsc-users] Install PETSc with option `--with-shared-libraries=1` failed on MacOS
Pierre Jolivet
pierre at joliv.et
Mon Mar 18 12:12:35 CDT 2024
> On 18 Mar 2024, at 5:13 PM, Satish Balay via petsc-users <petsc-users at mcs.anl.gov> wrote:
>
> Ah - the compiler did flag code bugs.
>
>> (current version is 0.3.26 but we can’t update because there is a huge performance regression which makes the pipeline timeout)
>
> maybe we should retry - updating to the latest snapshot and see if this issue persists.
Well, that’s easy to see it is _still_ broken: https://urldefense.us/v3/__https://gitlab.com/petsc/petsc/-/jobs/6419779589__;!!G_uCfscf7eWS!cZG6l8dQlL2q2LEgYiQw4bVE64zferDGxmonm_Z2I-6VXhae4u8oQiPv0BSGhXpi3y27-tKR-5wxh9MDGerWMQ$
The infamous gcc segfault that can’t let us run the pipeline, but that builds fine when it’s you that connect to the machine (I bothered you about this a couple of months ago in case you don’t remember, see https://urldefense.us/v3/__https://gitlab.com/petsc/petsc/-/merge_requests/7143__;!!G_uCfscf7eWS!cZG6l8dQlL2q2LEgYiQw4bVE64zferDGxmonm_Z2I-6VXhae4u8oQiPv0BSGhXpi3y27-tKR-5wxh9NMU62L0Q$ ).
Thanks,
Pierre
>
> Satish
>
> On Mon, 18 Mar 2024, Zongze Yang wrote:
>
>> The issue of openblas was resolved by this pr https://urldefense.us/v3/__https://github.com/OpenMathLib/OpenBLAS/pull/4565__;!!G_uCfscf7eWS!b09n5clcTFuLceLY_9KfqtSsgmmCIBLFbqciRVCKvnvFw9zTaNF8ssK0MiQlBOXUJe7H88nl-7ExdfhB-cMXLQ2d$
>>
>> Best wishes,
>> Zongze
>>
>>> On 18 Mar 2024, at 00:50, Zongze Yang <yangzongze at gmail.com> wrote:
>>>
>>> It can be resolved by adding CFLAGS=-Wno-int-conversion. Perhaps the default behaviour of the new version compiler has been changed?
>>>
>>> Best wishes,
>>> Zongze
>>>> On 18 Mar 2024, at 00:23, Satish Balay <balay at mcs.anl.gov> wrote:
>>>>
>>>> Hm - I just tried a build with balay/xcode15-mpich - and that goes through fine for me. So don't know what the difference here is.
>>>>
>>>> One difference is - I have a slightly older xcode. However your compiler appears to behave as using -Werror. Perhaps CFLAGS=-Wno-int-conversion will help here?
>>>>
>>>> Satish
>>>>
>>>> ----
>>>> Executing: gcc --version
>>>> stdout:
>>>> Apple clang version 15.0.0 (clang-1500.3.9.4)
>>>>
>>>> Executing: /Users/zzyang/workspace/repos/petsc/arch-darwin-c-debug/bin/mpicc -show
>>>> stdout: gcc -fPIC -fno-stack-check -Qunused-arguments -g -O0 -Wno-implicit-function-declaration -fno-common -I/Users/zzyang/workspace/repos/petsc/arch-darwin-c-debug/include -L/Users/zzyang/workspace/repos/petsc/arch-darwin-c-debug/lib -lmpi -lpmpi
>>>>
>>>> /Users/zzyang/workspace/repos/petsc/arch-darwin-c-debug/bin/mpicc -O2 -DMAX_STACK_ALLOC=2048 -Wall -DF_INTERFACE_GFORT -fPIC -DNO_WARMUP -DMAX_CPU_NUMBER=12 -DMAX_PARALLEL_NUMBER=1 -DBUILD_SINGLE=1 -DBUILD_DOUBLE=1 -DBUILD_COMPLEX=1 -DBUILD_COMPLEX16=1 -DVERSION=\"0.3.21\" -march=armv8-a -UASMNAME -UASMFNAME -UNAME -UCNAME -UCHAR_NAME -UCHAR_CNAME -DASMNAME=_lapack_wrappers -DASMFNAME=_lapack_wrappers_ -DNAME=lapack_wrappers_ -DCNAME=lapack_wrappers -DCHAR_NAME=\"lapack_wrappers_\" -DCHAR_CNAME=\"lapack_wrappers\" -DNO_AFFINITY -I.. -c src/lapack_wrappers.c -o src/lapack_wrappers.o
>>>> src/lapack_wrappers.c:570:81: error: incompatible integer to pointer conversion passing 'blasint' (aka 'int') to parameter of type 'const blasint *' (aka 'const int *'); take the address with & [-Wint-conversion]
>>>> RELAPACK_sgemmt(uplo, transA, transB, n, k, alpha, A, ldA, B, ldB, beta, C, info);
>>>> ^~~~
>>>> &
>>>>
>>>> vs:
>>>> Executing: gcc --version
>>>> stdout:
>>>> Apple clang version 15.0.0 (clang-1500.1.0.2.5)
>>>>
>>>> Executing: /Users/balay/petsc/arch-darwin-c-debug/bin/mpicc -show
>>>> stdout: gcc -fPIC -fno-stack-check -Qunused-arguments -g -O0 -Wno-implicit-function-declaration -fno-common -I/Users/balay/petsc/arch-darwin-c-debug/include -L/Users/balay/petsc/arch-darwin-c-debug/lib -lmpi -lpmpi
>>>>
>>>>
>>>> /Users/balay/petsc/arch-darwin-c-debug/bin/mpicc -O2 -DMAX_STACK_ALLOC=2048 -Wall -DF_INTERFACE_GFORT -fPIC -DNO_WARMUP -DMAX_CPU_NUMBER=24 -DMAX_PARALLEL_NUMBER=1 -DBUILD_SINGLE=1 -DBUILD_DOUBLE=1 -DBUILD_COMPLEX=1 -DBUILD_COMPLEX16=1 -DVERSION=\"0.3.21\" -march=armv8-a -UASMNAME -UASMFNAME -UNAME -UCNAME -UCHAR_NAME -UCHAR_CNAME -DASMNAME=_lapack_wrappers -DASMFNAME=_lapack_wrappers_ -DNAME=lapack_wrappers_ -DCNAME=lapack_wrappers -DCHAR_NAME=\"lapack_wrappers_\" -DCHAR_CNAME=\"lapack_wrappers\" -DNO_AFFINITY -I.. -c src/lapack_wrappers.c -o src/lapack_wrappers.o
>>>> src/lapack_wrappers.c:570:81: warning: incompatible integer to pointer conversion passing 'blasint' (aka 'int') to parameter of type 'const blasint *' (aka 'const int *'); take the address with & [-Wint-conversion]
>>>> RELAPACK_sgemmt(uplo, transA, transB, n, k, alpha, A, ldA, B, ldB, beta, C, info);
>>>> ^~~~
>>>> &
>>>>
>>>>
>>>>
>>>>
>>>> On Sun, 17 Mar 2024, Pierre Jolivet wrote:
>>>>
>>>>> Ah, my bad, I misread linux-opt-arm as a macOS runner, no wonder the option is not helping…
>>>>> Take Barry’s advice.
>>>>> Furthermore, it looks like OpenBLAS people are steering in the opposite direction as us, by forcing the use of ld-classic https://urldefense.us/v3/__https://github.com/OpenMathLib/OpenBLAS/commit/103d6f4e42fbe532ae4ea48e8d90d7d792bc93d2__;!!G_uCfscf7eWS!bY2l3X9Eb5PRzNQYrfPFXhgcUodHCiDinhQYga0PeQn1IQzJYD376fk-pZfktGAkpTvBmzy7BFDc9SrazFoooQ$ , so that’s another good argument in favor of -framework Accelerate.
>>>>>
>>>>> Thanks,
>>>>> Pierre
>>>>>
>>>>> PS: anyone benchmarked those https://urldefense.us/v3/__https://developer.apple.com/documentation/accelerate/sparse_solvers__;!!G_uCfscf7eWS!bY2l3X9Eb5PRzNQYrfPFXhgcUodHCiDinhQYga0PeQn1IQzJYD376fk-pZfktGAkpTvBmzy7BFDc9SrpnDvT5g$ ? I didn’t even know they existed.
>>>>>
>>>>>> On 17 Mar 2024, at 3:06 PM, Zongze Yang <yangzongze at gmail.com <mailto:yangzongze at gmail.com>> wrote:
>>>>>>
>>>>>> This Message Is From an External Sender
>>>>>> This message came from outside your organization.
>>>>>> Understood. Thank you for your advice.
>>>>>>
>>>>>> Best wishes,
>>>>>> Zongze
>>>>>>
>>>>>>> On 17 Mar 2024, at 22:04, Barry Smith <bsmith at petsc.dev <mailto:bsmith at petsc.dev> <mailto:bsmith at petsc.dev>> wrote:
>>>>>>>
>>>>>>>
>>>>>>> I would just avoid the --download-openblas option. The BLAS/LAPACK provided by Apple should perform fine, perhaps even better than OpenBLAS on your system.
>>>>>>>
>>>>>>>
>>>>>>>> On Mar 17, 2024, at 9:58 AM, Zongze Yang <yangzongze at gmail.com <mailto:yangzongze at gmail.com> <mailto:yangzongze at gmail.com>> wrote:
>>>>>>>>
>>>>>>>> This Message Is From an External Sender
>>>>>>>> This message came from outside your organization.
>>>>>>>> Adding the flag `--download-openblas-make-options=TARGET=GENERIC` did not resolve the issue. The same error persisted.
>>>>>>>>
>>>>>>>> Best wishes,
>>>>>>>> Zongze
>>>>>>>>
>>>>>>>>> On 17 Mar 2024, at 20:58, Pierre Jolivet <pierre at joliv.et <mailto:pierre at joliv.et> <mailto:pierre at joliv.et>> wrote:
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>> On 17 Mar 2024, at 1:04 PM, Zongze Yang <yangzongze at gmail.com <mailto:yangzongze at gmail.com> <mailto:yangzongze at gmail.com>> wrote:
>>>>>>>>>>
>>>>>>>>>> Thank you for providing the instructions. I try the first option.
>>>>>>>>>>
>>>>>>>>>> Now, the error of the configuration is related to OpenBLAS.
>>>>>>>>>> Add `--CFLAGS=-Wno-int-conversion` to configure command resolve this. Should this be reported to OpenBLAS? Or need to fix the configure in petsc?
>>>>>>>>>
>>>>>>>>> I see our linux-opt-arm runner is using the additional flag '--download-openblas-make-options=TARGET=GENERIC', could you maybe try to add that as well?
>>>>>>>>> I don’t think there is much to fix on our end, OpenBLAS has been very broken lately on arm (current version is 0.3.26 but we can’t update because there is a huge performance regression which makes the pipeline timeout).
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>> Pierre
>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> The configure.log is attached. The errors are show below:
>>>>>>>>>> ```
>>>>>>>>>> src/lapack_wrappers.c:570:81: error: incompatible integer to pointer conversion passing 'blasint' (aka 'int') to parameter of type 'const blasint *' (aka 'const int *'); take the address with & [-Wint-conversion]
>>>>>>>>>> RELAPACK_sgemmt(uplo, transA, transB, n, k, alpha, A, ldA, B, ldB, beta, C, info);
>>>>>>>>>> ^~~~
>>>>>>>>>> &
>>>>>>>>>> src/../inc/relapack.h:74:216: note: passing argument to parameter here
>>>>>>>>>> void RELAPACK_sgemmt(const char *, const char *, const char *, const blasint *, const blasint *, const float *, const float *, const blasint *, const float *, const blasint *, const float *, float *, const blasint *);
>>>>>>>>>> ^
>>>>>>>>>> src/lapack_wrappers.c:583:81: error: incompatible integer to pointer conversion passing 'blasint' (aka 'int') to parameter of type 'const blasint *' (aka 'const int *'); take the address with & [-Wint-conversion]
>>>>>>>>>> RELAPACK_dgemmt(uplo, transA, transB, n, k, alpha, A, ldA, B, ldB, beta, C, info);
>>>>>>>>>> ^~~~
>>>>>>>>>> &
>>>>>>>>>> src/../inc/relapack.h:75:221: note: passing argument to parameter here
>>>>>>>>>> void RELAPACK_dgemmt(const char *, const char *, const char *, const blasint *, const blasint *, const double *, const double *, const blasint *, const double *, const blasint *, const double *, double *, const blasint *);
>>>>>>>>>> ^
>>>>>>>>>> src/lapack_wrappers.c:596:81: error: incompatible integer to pointer conversion passing 'blasint' (aka 'int') to parameter of type 'const blasint *' (aka 'const int *'); take the address with & [-Wint-conversion]
>>>>>>>>>> RELAPACK_cgemmt(uplo, transA, transB, n, k, alpha, A, ldA, B, ldB, beta, C, info);
>>>>>>>>>> ^~~~
>>>>>>>>>> &
>>>>>>>>>> src/../inc/relapack.h:76:216: note: passing argument to parameter here
>>>>>>>>>> void RELAPACK_cgemmt(const char *, const char *, const char *, const blasint *, const blasint *, const float *, const float *, const blasint *, const float *, const blasint *, const float *, float *, const blasint *);
>>>>>>>>>> ^
>>>>>>>>>> src/lapack_wrappers.c:609:81: error: incompatible integer to pointer conversion passing 'blasint' (aka 'int') to parameter of type 'const blasint *' (aka 'const int *'); take the address with & [-Wint-conversion]
>>>>>>>>>> RELAPACK_zgemmt(uplo, transA, transB, n, k, alpha, A, ldA, B, ldB, beta, C, info);
>>>>>>>>>> ^~~~
>>>>>>>>>> &
>>>>>>>>>> src/../inc/relapack.h:77:221: note: passing argument to parameter here
>>>>>>>>>> void RELAPACK_zgemmt(const char *, const char *, const char *, const blasint *, const blasint *, const double *, const double *, const blasint *, const double *, const blasint *, const double *, double *, const blasint *);
>>>>>>>>>> ^
>>>>>>>>>> 4 errors generated.
>>>>>>>>>> ```
>>>>>>>>>>
>>>>>>>>>> Best wishes,
>>>>>>>>>> Zongze
>>>>>>>>>>
>>>>>>>>>> <configure.log.tar.gz>
>>>>>>>>>>
>>>>>>>>>>> On 17 Mar 2024, at 18:48, Pierre Jolivet <pierre at joliv.et <mailto:pierre at joliv.et> <mailto:pierre at joliv.et>> wrote:
>>>>>>>>>>>
>>>>>>>>>>> You need this MR https://urldefense.us/v3/__https://gitlab.com/petsc/petsc/-/merge_requests/7365__;!!G_uCfscf7eWS!bY2l3X9Eb5PRzNQYrfPFXhgcUodHCiDinhQYga0PeQn1IQzJYD376fk-pZfktGAkpTvBmzy7BFDc9SqG8HOUGQ$ <https://urldefense.us/v3/__https://gitlab.com/petsc/petsc/-/merge_requests/7365__;!!G_uCfscf7eWS!eCQRfbol7FDQiO0o78iDit2saij_ydIUtCfRQnsQAt-h_YcXr2Yi2BFnFnqHZp0FO3Lhpyr2RKdHZ-T-OF94HpwQ$>
>>>>>>>>>>> main has been broken for macOS since https://urldefense.us/v3/__https://gitlab.com/petsc/petsc/-/merge_requests/7341__;!!G_uCfscf7eWS!bY2l3X9Eb5PRzNQYrfPFXhgcUodHCiDinhQYga0PeQn1IQzJYD376fk-pZfktGAkpTvBmzy7BFDc9Soe8Kh_uQ$ <https://urldefense.us/v3/__https://gitlab.com/petsc/petsc/-/merge_requests/7341__;!!G_uCfscf7eWS!eCQRfbol7FDQiO0o78iDit2saij_ydIUtCfRQnsQAt-h_YcXr2Yi2BFnFnqHZp0FO3Lhpyr2RKdHZ-T-OIhlJwLx$>, so the alternative is to revert to the commit prior.
>>>>>>>>>>> It should work either way.
>>>>>>>>>>>
>>>>>>>>>>> Thanks,
>>>>>>>>>>> Pierre
>>>>>>>>>>>
>>>>>>>>>>>> On 17 Mar 2024, at 11:31 AM, Zongze Yang <yangzongze at gmail.com <mailto:yangzongze at gmail.com> <mailto:yangzongze at gmail.com>> wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> This Message Is From an External Sender
>>>>>>>>>>>> This message came from outside your organization.
>>>>>>>>>>>> Hi, PETSc Team,
>>>>>>>>>>>>
>>>>>>>>>>>> I am trying to install petsc with the following configuration
>>>>>>>>>>>> ```
>>>>>>>>>>>> ./configure \
>>>>>>>>>>>> --download-bison \
>>>>>>>>>>>> --download-mpich \
>>>>>>>>>>>> --download-mpich-configure-arguments=--disable-opencl \
>>>>>>>>>>>> --download-hwloc \
>>>>>>>>>>>> --download-hwloc-configure-arguments=--disable-opencl \
>>>>>>>>>>>> --download-openblas \
>>>>>>>>>>>> --download-openblas-make-options="'USE_THREAD=0 USE_LOCKING=1 USE_OPENMP=0'" \
>>>>>>>>>>>> --with-shared-libraries=1 \
>>>>>>>>>>>> --with-fortran-bindings=0 \
>>>>>>>>>>>> --with-zlib \
>>>>>>>>>>>> LDFLAGS=-Wl,-ld_classic
>>>>>>>>>>>> ```
>>>>>>>>>>>>
>>>>>>>>>>>> The log shows that
>>>>>>>>>>>> ```
>>>>>>>>>>>> Exhausted all shared linker guesses. Could not determine how to create a shared library!
>>>>>>>>>>>> ```
>>>>>>>>>>>>
>>>>>>>>>>>> I recently updated the system and Xcode, as well as homebrew.
>>>>>>>>>>>>
>>>>>>>>>>>> The configure.log is attached.
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks for your attention to this matter.
>>>>>>>>>>>>
>>>>>>>>>>>> Best wishes,
>>>>>>>>>>>> Zongze
>>>>>>>>>>>> <configure.log.tar.gz>
>>>>>
>>>> <configure.log.gz>
>>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20240318/cb0cd12b/attachment-0001.html>
More information about the petsc-users
mailing list