[tizen] Version of modules in external directory

[chunseoklee]

Let's check tizen's version of modules in external directory.

We have the following modules :

~/ONE/externals $ ls *.stamp
ABSEIL.stamp            GTEST.stamp              TENSORFLOW-1.14.stamp
ARMCOMPUTE.stamp        HDF5.stamp               TENSORFLOW-2.3.0-EIGEN.stamp
BOOST.stamp             NEON2SSE.stamp           TENSORFLOW-2.3.0-GEMMLOWP.stamp
CAFFE.stamp             ONNX-1.6.0.stamp         TENSORFLOW-2.3.0-RUY.stamp
CPUINFO.stamp           OOURAFFT.stamp           TENSORFLOW-2.3.0.stamp
EIGEN.stamp             OPENCL_HEADERS.stamp     TENSORFLOW-2.6.0-EIGEN.stamp
FARMHASH.stamp          PROTOBUF.stamp           TENSORFLOW-2.6.0-GEMMLOWP.stamp
FLATBUFFERS-1.10.stamp  PSIMD.stamp              TENSORFLOW-2.6.0-RUY.stamp
FLATBUFFERS-1.12.stamp  PTHREADPOOL.stamp        TENSORFLOW-2.6.0.stamp
FP16.stamp              PYTORCH.stamp            TENSORFLOW_GPU.stamp
FXDIV.stamp             RUY.stamp                XNNPACK.stamp
GEMMLOWP.stamp          TENSORFLOW-1.13.1.stamp  

[chunseoklee]

module	one	tizen
ABSEIL	df3ea785	20210324.2	one uses no tag
ARMCOMPUTE
BOOST	1.58	1.71	Use 1.58 on source build
CAFFE
CPUINFO
EIGEN
FARMHASH
FLATBUFFERS-1.10
FLATBUFFERS-1.12
FP16
FXDIV
GEMMLOWP
GTEST
HDF5	1.8.16	1.10.1
NEON2SSE
ONNX-1.6.0
OOURAFFT
OPENCL_HEADERS
PROTOBUF
PSIMD
PTHREADPOOL
PYTORCH
RUY
TENSORFLOW_GPU
TENSORFLOW-1.13.1
TENSORFLOW-1.14
TENSORFLOW-2.3.0
TENSORFLOW-2.3.0-EIGEN
TENSORFLOW-2.3.0-GEMMLOWP
TENSORFLOW-2.3.0-RUY
TENSORFLOW-2.6.0
TENSORFLOW-2.6.0-EIGEN
TENSORFLOW-2.6.0-GEMMLOWP
TENSORFLOW-2.6.0-RUY
XNNPACK

[chunseoklee]

gtest patch : https://github.com/Samsung/ONE/issues/8567

[hseok-oh]

@mhs4670go Is there any reason to decide using HDF version 1.8.16 ? (#3101)

[seanshpark]

ping @mhs4670go CC @jinevening

[jinevening]

Q) Is it necessary to match ONE's HDF5 version with Tizen? ONE builds its own hdf5 library and uses it.

It seems that HDF5 file created by 1.8 is compatible with 1.10. https://support.hdfgroup.org/HDF5/faq/bkfwd-compat.html "An HDF5 Library of any given release is designed to read all existing HDF5 files from that or any prior release."

[mhs4670go]

@mhs4670go Is there any reason to decide using HDF version 1.8.16 ? (#3101)

There is no specific reason.

[seanshpark]

One thing.. if we change HDF5 version, I need to revise HDF5Source.patch for ARM32 build .. -_-;;;