Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

update to alpaka pre-0.4.0 version #2958

Conversation

psychocoderHPC
Copy link
Member

@psychocoderHPC psychocoderHPC commented May 9, 2019

updated version of #2807 to verify issues before we release a new alpaka version

update Cupla from :ComputationalRadiationPhysics/cupla@73fb269 to ComputationalRadiationPhysics/cupla@a96b293

GIT_AUTHOR_NAME="Third Party" GIT_AUTHOR_EMAIL="picongpu@hzdr.de" git subtree pull --prefix thirdParty/cupla git@github.com:ComputationalRadiationPhysics/cupla.git dev --squash

update alpaka from ComputationalRadiationPhysics/alpaka@b5c66c2 to ComputationalRadiationPhysics/alpaka@6b42f89

GIT_AUTHOR_NAME="Third Party" GIT_AUTHOR_EMAIL="picongpu@hzdr.de" \
  git subtree pull --prefix thirdParty/alpaka \                                                  
  git@github.com:ComputationalRadiationPhysics/alpaka.git develop --squash

@psychocoderHPC
Copy link
Member Author

@ax3l Do you see why the CI is not checking my PR. I have currently no access to the CI node and can not check it by myself :-(

@psychocoderHPC
Copy link
Member Author

We need to check the cmake version of our CI because travis is also showing that we are using an to old cmake https://travis-ci.org/ComputationalRadiationPhysics/picongpu/jobs/530345749#L812

@psychocoderHPC
Copy link
Member Author

psychocoderHPC commented May 10, 2019

Note: The problem why the CI is not showing any error is that the error log is 12MiB and this is too large.
In general local tests show that we still have the issue with cuda8 and this PR what we already observed in #2807.
We need to switch to CUDA 9.X for this PR. Some examples still run with CUDA8 so we should not forbid the usage but should remove support for this version.

@psychocoderHPC psychocoderHPC added the component: third party third party libraries that are shipped and/or linked label May 10, 2019
@ax3l
Copy link
Member

ax3l commented May 10, 2019 via email

@psychocoderHPC psychocoderHPC force-pushed the topic-alpakaAndCuplaUpdate branch 2 times, most recently from daa39d6 to ff2aa27 Compare May 17, 2019 12:28
Third Party and others added 3 commits May 17, 2019 17:17
git-subtree-dir: thirdParty/alpaka
git-subtree-split: 6b42f89204a6800a926b36f07d8fd2908892378e
a96b293 Merge pull request ComputationalRadiationPhysics#106 from fwyzard/cupla_accelerator_namespace
1af2b2b Merge pull request ComputationalRadiationPhysics#110 from tdd11235813/cmake-macro
663521c Changes function to macro 'CUPLA_ADD_EXECUTABLE'.
a440a7c Fix namespace definition if two accelerators are selected
64f02bb Move all accelerator-specific code to an inline namespece
bbef699 Merge pull request ComputationalRadiationPhysics#105 from fwyzard/tbb_IsThreadSeqAcc
b463a25 Mark the TBB backend as "thread sequential"
b45d4e9 Merge pull request ComputationalRadiationPhysics#102 from psychocoderHPC/topic-updateToLatestAlapkaDev
7e8b832 createTaskExec -> createTaskKernel
84b6dc1 update to latest alapka develop branch
037f7b2 Merge pull request ComputationalRadiationPhysics#100 from DerWaldschrat/dev
7add5ed fix -Wreorder in cuplaPitchedPtr (#4)
9fa745b Merge pull request ComputationalRadiationPhysics#99 from ax3l/topic-cmake312rootHints
041d454 Merge pull request ComputationalRadiationPhysics#98 from DerWaldschrat/dev-fix
ecc5839 CMake: Honor _ROOT Env Hints
24d159a Added conditional check for backends.
5822555 Add cudaEventBlockingSync to event creation
906cd41 Merge pull request ComputationalRadiationPhysics#97 from psychocoderHPC/topic-HIPBackend
bc67259 alpaka backend: HIP
87fe58d Merge pull request ComputationalRadiationPhysics#96 from psychocoderHPC/fix-kernelSharedMemCalc
68877b5 fix dynamic shared mem definition
9452a70 Merge pull request ComputationalRadiationPhysics#95 from psychocoderHPC/topic-updateDocs
8b26918 update docs
e822bfa Merge pull request ComputationalRadiationPhysics#89 from psychocoderHPC/topic-switchToAlpakaDevelop0.4.X
50cedbc alpaka renamings + cmake version incerase
5ecd227 update to alapka dev upcoming version 0.4.X
66a9240 Merge pull request ComputationalRadiationPhysics#94 from psychocoderHPC/topic-ciUpdateToCUDA8
3ba6104 CI: update CUDA from 7.5 to 8.0
5af56ce Merge pull request ComputationalRadiationPhysics#93 from psychocoderHPC/topic-updateCIToClang4
68a6555 Travis: CUDA Compiler Ignore
279bfd0 Travis: Xenial
9d583b7 CI: switch from clan 3.5 to 4.0
11eb874 Merge pull request ComputationalRadiationPhysics#92 from ax3l/topic-ciCMake3110
da617da Travis: CMake 3.11.0
da63524 Merge pull request ComputationalRadiationPhysics#91 from ComputationalRadiationPhysics/dev
66a2da7 Merge pull request ComputationalRadiationPhysics#88 from psychocoderHPC/topic-versionIncreaseTo0.1.1
e029584 version increase to 0.1.1
ffdb268 Merge pull request ComputationalRadiationPhysics#90 from psychocoderHPC/fix-tbbSelection
06030af fix TBB backend activation
aa8a921 Merge pull request ComputationalRadiationPhysics#84 from psychocoderHPC/topic-cuplaGetErrorString
55c99b7 add implementation for `cuplaGetErrorString`
c765052 Merge pull request ComputationalRadiationPhysics#86 from psychocoderHPC/topic-cuplaPeekAtLastError
49612bb Merge pull request ComputationalRadiationPhysics#87 from psychocoderHPC/topic-helpUpdatecuplaGetLastError
c763de4 Merge pull request ComputationalRadiationPhysics#85 from psychocoderHPC/topic-cuplaStreamQuerySupport
5d19923 update documentation of `cuplaGetLastError`
3806b8f implement `cuplaPeekAtLastError()`
2e79816 add `cuplaStreamQuery()`
7c73dc8 Merge pull request ComputationalRadiationPhysics#83 from psychocoderHPC/fix-cuplaGetLastError
b02299a fix `cuplaGetLastError`
1b7d7ea Merge pull request ComputationalRadiationPhysics#81 from tdd11235813/master
b0adfd9 Updates alpaka to 0.3.5.

git-subtree-dir: thirdParty/cupla
git-subtree-split: a96b293dcb14662a57901b55dfedaa1cd615adc9
@psychocoderHPC
Copy link
Member Author

psychocoderHPC commented May 20, 2019

Since the CI is not showing the full error I run the test manually on the CI node:

In file included from /tmp/tmpxft_00006220_00000000-5_main.cudafe1.stub.c:5:0,
                 from tmpxft_00006220_00000000-5_main.cudafe1.stub.c:1:
/tmp/tmpxft_00006220_00000000-5_main.cudafe1.stub.c: In function ‘void __device_stub__ZN6alpaka6kernel4cuda6detail10cudaKernelISt17integral_constantImLm3EEjN5cupla16cupla_cuda_async11CuplaKernelIN8picongpu21KernelEnergyParticlesILj256EEEEEJN5pmacc12ParticlesBoxINSD_5FrameINSD_15ParticlesBufferINSD_19ParticleDescriptionINSD_11compileTime6StringIJLc101ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0EEEENSD_4math2CT6VectorIN4mpl_10integral_cIiLi8EEESQ_NSP_IiLi4EEEEEN5boost3mpl6v_itemINS9_24placeholder_definition309weightingENSV_INS9_24placeholder_definition288momentumENSV_INS9_24placeholder_definition258positionINS9_24placeholder_definition2712position_picENSD_24placeholder_definition2213pmacc_isAliasEEENSU_7vector0INSO_2naEEELi0EEELi0EEELi0EEENSV_INS9_24placeholder_definition5111chargeRatioINS9_25placeholder_definition13120ChargeRatioElectronsES15_EENSV_INS9_24placeholder_definition509massRatioINS9_25placeholder_definition13018MassRatioElectronsES15_EENSV_INS9_24placeholder_definition467currentINS9_13currentSolver9EsirkepovINS9_9particles6shapes3CICELj3EEES15_EENSV_INS9_24placeholder_definition4513interpolationINS9_28FieldToParticleInterpolationIS1T_NS9_30AssignedTrilinearInterpolationEEES15_EENSV_INS9_24placeholder_definition385shapeIS1T_S15_EENSV_INS9_24placeholder_definition3914particlePusherINS1R_6pusher5BorisES15_EES19_Li0EEELi0EEELi0EEELi0EEELi0EEELi0EEENSD_17HandleGuardRegionINSD_9particles8policies17ExchangeParticlesENS1R_8boundary29CallPluginsAndDeleteParticlesEEES19_S19_EESS_N8mallocMC9AllocatorINS2O_16CreationPolicies7ScatterINS9_16DeviceHeapConfigENS2Q_11ScatterConf27DefaultScatterHashingParamsEEENS2O_20DistributionPolicies4NoopENS2O_11OOMPolicies10ReturnNullENS2O_19ReservePoolPolicies16SimpleCudaMallocENS2O_17AlignmentPolicies6ShrinkINS32_12ShrinkConfig19DefaultShrinkConfigEEEEELj3EE29OperatorCreatePairStaticArrayILj256EEENSH_ISK_SS_NSV_INSD_24placeholder_definition249multiMaskENSV_INSD_24placeholder_definition2312localCellIdxES1C_Li0EEELi0EEES2F_S2M_S19_NSV_INSD_12NextFramePtrINSO_3argILi1EEEEENSV_INSD_16PreviousFramePtrIS3J_EES19_Li0EEELi0EEEEEEENS2O_19AllocatorHandleImplIS37_EELj3EEENSD_7DataBoxINSD_10PitchedBoxIdLj1EEEEENSD_11AreaMappingILj3ENSD_18MappingDescriptionILj3ESS_EEEENSD_7functor9InterfaceINS1R_6filter3AllELj1EbEEEEEvNS_3vec3VecIT_T0_EET1_DpT2_(const _ZN6alpaka3vec3VecISt17integral_constantImLm3EEjEE&, const _ZN5cupla16cupla_cuda_async11CuplaKernelIN8picongpu21KernelEnergyParticlesILj256EEEEE&, _ZN5pmacc13ParticlesBaseINS_19ParticleDescriptionINS_11compileTime6StringIJLc101ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0ELc0EEEENS_4math2CT6VectorIN4mpl_10integral_cIiLi8EEESA_NS9_IiLi4EEEEEN5boost3mpl6v_itemIN8picongpu24placeholder_definition309weightingENSF_INSG_24placeholder_definition288momentumENSF_INSG_24placeholder_definition258positionINSG_24placeholder_definition2712position_picENS_24placeholder_definition2213pmacc_isAliasEEENSE_7vector0INS8_2naEEELi0EEELi0EEELi0EEENSF_INSG_24placeholder_definition5111chargeRatioINSG_25placeholder_definition13120ChargeRatioElectronsESQ_EENSF_INSG_24placeholder_definition509massRatioINSG_25placeholder_definition13018MassRatioElectronsESQ_EENSF_INSG_24placeholder_definition467currentINSG_13currentSolver9EsirkepovINSG_9particles6shapes3CICELj3EEESQ_EENSF_INSG_24placeholder_definition4513interpolationINSG_28FieldToParticleInterpolationIS1E_NSG_30AssignedTrilinearInterpolationEEESQ_EENSF_INSG_24placeholder_definition385shapeIS1E_SQ_EENSF_INSG_24placeholder_definition3914particlePusherINS1C_6pusher5BorisESQ_EESU_Li0EEELi0EEELi0EEELi0EEELi0EEELi0EEENS_17HandleGuardRegionINS_9particles8policies17ExchangeParticlesENS1C_8boundary29CallPluginsAndDeleteParticlesEEESU_SU_EENS_18MappingDescriptionILj3ESC_EEN8mallocMC9AllocatorINS2B_16CreationPolicies7ScatterINSG_16DeviceHeapConfigENS2D_11ScatterConf27DefaultScatterHashingParamsEEENS2B_20DistributionPolicies4NoopENS2B_11OOMPolicies10ReturnNullENS2B_19ReservePoolPolicies16SimpleCudaMallocENS2B_17AlignmentPolicies6ShrinkINS2P_12ShrinkConfig19DefaultShrinkConfigEEEEEE16ParticlesBoxTypeE&, _ZN5pmacc7DataBoxINS_10PitchedBoxIdLj1EEEEE&, _ZN5pmacc11AreaMappingILj3ENS_18MappingDescriptionILj3ENS_4math2CT6VectorIN4mpl_10integral_cIiLi8EEES7_NS6_IiLi4EEEEEEEEE&, _ZNSt3_MuISt12_PlaceholderILi1EELb0ELb1EE6resultIFS2_S1_St5tupleIJN5pmacc7functor9InterfaceIN8picongpu9particles6filter3AllELj1EbEEEEEE11__base_typeE&)’:
/tmp/tmpxft_00006220_00000000-5_main.cudafe1.stub.c:995:4921: error: invalid application of ‘sizeof’ to incomplete type ‘_ZNSt3_MuISt12_PlaceholderILi1EELb0ELb1EE6resultIFS2_S1_St5tupleIJN5pmacc7functor9InterfaceIN8picongpu9particles6filter3AllELj1EbEEEEEE11__base_typeE {aka _ZN5pmacc7functor9InterfaceIN8picongpu9particles6filter3AllELj1EbEE}’

human readable with c++file:

incomplete type: std::_Mu<std::_Placeholder<1>, false, true>::result<std::_Mu<std::_Placeholder<1>, false, true> (std::_Placeholder<1>, std::tuple<pmacc::functor::Interface<picongpu::particles::filter::All, 1u, bool> >)>::__base_type 
aka
pmacc::functor::Interface<picongpu::particles::filter::All, 1u, bool>

@psychocoderHPC
Copy link
Member Author

psychocoderHPC commented May 20, 2019

Test failed with CUDA 9.0 and 9.1. :-(
CUDA 9.2 works fine.

@psychocoderHPC
Copy link
Member Author

Tested CUDA 10.1: compile without issues

 ...........,_
 ........../.(|
 ..........\..\ 
 ........___\..\,. ~~~~~~
 .......(__)_)...\ 
 ......(__)__)|...|
 ......(__)__)|.__|
 .......(__)__)___/~~~~~~

@psychocoderHPC psychocoderHPC changed the title [WIP] update to alpaka pre-0.4.0 version update to alpaka pre-0.4.0 version Jun 7, 2019
@psychocoderHPC psychocoderHPC requested a review from ax3l June 7, 2019 09:37
@psychocoderHPC
Copy link
Member Author

@ax3l As discussed please merge this PR as it is even if we not update directly to the latest version of alpaka and cupla. Do to the renamings in the libraries there was some hand work needed to solve merge conflicts (I checked my changes against a clean checkout of the corresponding library)

@ax3l
Copy link
Member

ax3l commented Jun 7, 2019

CI and project updated to only support CUDA 9.2 or newer in #2975

@ax3l ax3l merged commit de75529 into ComputationalRadiationPhysics:dev Jun 7, 2019
@ax3l ax3l mentioned this pull request Jan 8, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
component: third party third party libraries that are shipped and/or linked
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants