FastSmallVector: initialize array member

[akva2]

Quells downstream warning

In file included from /home/akva/kode/opm/opm-common/opm/material/densead/EvaluationSpecializations.hpp:34,
                 from /home/akva/kode/opm/opm-common/opm/material/densead/Evaluation.hpp:631,
                 from /home/akva/kode/opm/opm-simulators/opm/simulators/aquifers/AquiferAnalytical.hpp:32,
                 from /home/akva/kode/opm/opm-simulators/opm/simulators/aquifers/AquiferCarterTracy.hpp:30,
                 from /home/akva/kode/opm/opm-simulators/opm/simulators/aquifers/BlackoilAquiferModel.hpp:32,
                 from /home/akva/kode/opm/opm-simulators/opm/simulators/flow/SimulatorFullyImplicitBlackoil.hpp:43,
                 from /home/akva/kode/opm/opm-simulators/flow/flow_oilwater_polymer.cpp:25:
In member function ‘Opm::DenseAd::Evaluation<ValueT, -1, staticSize>& Opm::DenseAd::Evaluation<ValueT, -1, staticSize>::operator-=(const RhsValueType&) [with RhsValueType = double; ValueT = double; unsigned int staticSize = 7]’,
    inlined from ‘Opm::DenseAd::Evaluation<ValueT, -1, staticSize> Opm::DenseAd::Evaluation<ValueT, -1, staticSize>::operator-(const RhsValueType&) const [with RhsValueType = double; ValueT = double; unsigned int staticSize = 7]’ at /home/akva/kode/opm/opm-common/opm/material/densead/DynamicEvaluation.hpp:382:16,
    inlined from ‘Evaluation Opm::Tabulated1DFunction<Scalar>::eval(const Evaluation&, Opm::SegmentIndex) const [with Evaluation = Opm::DenseAd::Evaluation<double, -1, 7>; Scalar = double]’ at /home/akva/kode/opm/opm-common/opm/material/common/Tabulated1DFunction.hpp:278:34,
    inlined from ‘Evaluation Opm::Tabulated1DFunction<Scalar>::eval(const Evaluation&, bool) const [with Evaluation = Opm::DenseAd::Evaluation<double, -1, 7>; Scalar = double]’ at /home/akva/kode/opm/opm-common/opm/material/common/Tabulated1DFunction.hpp:265:30,
    inlined from ‘Opm::BlackOilPolymerModule<Opm::Properties::TTag::FlowOilWaterPolymerProblem, true>::computeShearFactor<Opm::DenseAd::Evaluation<double, -1, 7> >(const Opm::DenseAd::Evaluation<double, -1, 7>&, unsigned int, const Opm::DenseAd::Evaluation<double, -1, 7>&)::<lambda(const Opm::DenseAd::Evaluation<double, -1, 7>&)>’ at /home/akva/kode/opm/opm-simulators/opm/models/blackoil/blackoilpolymermodules.hh:499:53,
    inlined from ‘static Evaluation Opm::BlackOilPolymerModule<TypeTag, enablePolymerV>::computeShearFactor(const Evaluation&, unsigned int, const Evaluation&) [with Evaluation = Opm::DenseAd::Evaluation<double, -1, 7>; TypeTag = Opm::Properties::TTag::FlowOilWaterPolymerProblem; bool enablePolymerV = true]’ at /home/akva/kode/opm/opm-simulators/opm/models/blackoil/blackoilpolymermodules.hh:512:18:
/home/akva/kode/opm/opm-common/opm/material/densead/DynamicEvaluation.hpp:274:28: warning: ‘<unnamed>.Opm::DenseAd::Evaluation<double, -1, 7>::data_.Opm::FastSmallVector<double, 7>::smallBuf_.std::array<double, 7>::_M_elems[0]’ may be used uninitialized [-Wmaybe-uninitialized]
  274 |         data_[valuepos_()] -= other;
      |         ~~~~~~~~~~~~~~~~~~~^~~~~~~~

[akva2]

jenkins build this please

[bska]

We may want to benchmark this. From a C++ point of view this is obviously fine, but the FastSmallVector is supposed to be fast (and small). This forced initialisation for built-in scalar types such as double may be a bit too costly.

[akva2]

I agree but I think you'd be hard pressed to pick it up in anything but micro benches and I don't think you'll see it in the macro benches. But let's run the benches at least.

[akva2]

benchmark please

[ytelses]

Benchmark result overview:

Test	Configuration	Relative
opm-git	OPM Benchmark: drogon - Threads: 1	0.997
opm-git	OPM Benchmark: drogon - Threads: 8	1.211
opm-git	OPM Benchmark: punqs3 - Threads: 1	0.981
opm-git	OPM Benchmark: punqs3 - Threads: 8	1.009
opm-git	OPM Benchmark: smeaheia - Threads: 1	0.994
opm-git	OPM Benchmark: smeaheia - Threads: 8	1.079
opm-git	OPM Benchmark: spe10_model_1 - Threads: 1	1.002
opm-git	OPM Benchmark: spe10_model_1 - Threads: 8	0.99
opm-git	OPM Benchmark: flow_mpi_extra - Threads: 1	1
opm-git	OPM Benchmark: flow_mpi_extra - Threads: 8	1.129
opm-git	OPM Benchmark: flow_mpi_norne - Threads: 1	0.996
opm-git	OPM Benchmark: flow_mpi_norne - Threads: 8	1.076
opm-git	OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 1 - FOPT (Total Oil Production At End Of Run)	1
opm-git	OPM Benchmark: flow_mpi_norne_4c_msw - Threads: 8 - FOPT (Total Oil Production At End Of Run)	1

• Speed-up = Total time master / Total time pull request. Above 1.0 is an improvement. *

View result details @ https://www.ytelses.com/opm/?page=result&id=2693