[PMEM] It will abort when using PMEM allocator in EV

[shanzhou2186]

While using pmem allocator in the WDL model both on libpmem or memkind mode, it would cause "./tensorflow/core/framework/embedding/value_ptr.h:273] Unsupport FreqCounter in subclass of ValuePtrBase
Aborted (core dumped)"

Here are the call stack information.
#3 0x00001464e19d0f4e in tensorflow::ValuePtr::AddFreq (this=)
at ./tensorflow/core/framework/embedding/value_ptr.h:273
#4 0x00001464e19d6566 in tensorflow::NullableFilter<long long, float, tensorflow::EmbeddingVar<long long, float> >::LookupOrCreateWithFreq (this=0x145fb0105c90, key=, val=0x14609c00cac0, default_value_ptr=)
at ./tensorflow/core/framework/embedding/embedding_filter.h:526
#5 0x00001464e19c35cc in std::function<void (long long, float*, float*)>::operator()(long long, float*, float*) const (
__args#2=, __args#1=, __args#0=, this=0x146100083cb8)
at /usr/include/c++/7/bits/std_function.h:706
#6 tensorflow::KvResourceGatherOp<long long, float>::Compute(tensorflow::OpKernelContext*)::{lambda(long long, long long)#4}::operator()(long long, long long) const (limit=4, start=, __closure=0x146100083c80)
at tensorflow/core/kernels/kv_variable_ops.cc:413
#7 std::_Function_handler<void (long long, long long), tensorflow::KvResourceGatherOp<long long, float>::Compute(tensorflow::OpKernelContext*)::{lambda(long long, long long)#4}>::_M_invoke(std::_Any_data const&, long long&&, std::_Any_data const&) (
__functor=..., __args#0=, __args#1=) at /usr/include/c++/7/bits/std_function.h:316
#8 0x00001464d9948f1e in std::_Function_handler<void (long, long), tensorflow::thread::ThreadPool::ParallelFor(long long, long long, std::function<void (long long, long long)>)::{lambda(long, long)#1}>::_M_invoke(std::_Any_data const&, long&&, std::_Any_data const&) () from /home/zshan/deeprec-env/lib/python3.6/site-packages/tensorflow_core/python/../libtensorflow_framework.so.1
#9 0x00001464d994f48f in tensorflow::thread::ThreadPool::ParallelFor(long long, long long, std::function<void (long long, long long)>) () from /home/zshan/deeprec-env/lib/python3.6/site-packages/tensorflow_core/python/../libtensorflow_framework.so.1
#10 0x00001464d971fb52 in tensorflow::Shard(int, tensorflow::thread::ThreadPool*, long long, long long, std::function<void (long long, long long)>) ()
from /home/zshan/deeprec-env/lib/python3.6/site-packages/tensorflow_core/python/../libtensorflow_framework.so.1
#11 0x00001464e19dce74 in tensorflow::KvResourceGatherOp<long long, float>::Compute (this=0x560ce5050590, c=)
at tensorflow/core/kernels/kv_variable_ops.cc:427
#12 0x00001464d98766a6 in tensorflow::(anonymous namespace)::ExecutorStatetensorflow::PropagatorState::BatchProcess(std::vector<tensorflow::PropagatorState::TaggedNode, std::allocatortensorflow::PropagatorState::TaggedNode >, int, long) ()
--Type for more, q to quit, c to continue without paging--
from /home/zshan/deeprec-env/lib/python3.6/site-packages/tensorflow_core/python/../libtensorflow_framework.so.1
#13 0x00001464d9876a88 in tensorflow::(anonymous namespace)::ExecutorStatetensorflow::PropagatorState::Process(tensorflow::PropagatorState::TaggedNode, long) ()
from /home/zshan/deeprec-env/lib/python3.6/site-packages/tensorflow_core/python/../libtensorflow_framework.so.1
#14 0x00001464d9876b5f in std::_Function_handler<void (), tensorflow::(anonymous namespace)::ExecutorStatetensorflow::PropagatorState::RunTask<tensorflow::(anonymous namespace)::ExecutorStatetensorflow::PropagatorState::ScheduleReady(absl::InlinedVector<tensorflow::PropagatorState::TaggedNode, 8ul, std::allocatortensorflow::PropagatorState::TaggedNode >, tensorflow::PropagatorState::TaggedNodeReadyQueue)::{lambda()#1}>(tensorflow::(anonymous namespace)::ExecutorStatetensorflow::PropagatorState::ScheduleReady(absl::InlinedVector<tensorflow::PropagatorState::TaggedNode, 8ul, std::allocatortensorflow::PropagatorState::TaggedNode >, tensorflow::PropagatorState::TaggedNodeReadyQueue)::{lambda()#1}&&)::{lambda()#1}>::_M_invoke(std::_Any_data const&)
() from /home/zshan/deeprec-env/lib/python3.6/site-packages/tensorflow_core/python/../libtensorflow_framework.so.1
#15 0x00001464d994bb4f in std::_Function_handler<void (), Eigen::ThreadPoolTempltensorflow::thread::EigenEnvironment::ThreadPoolTempl(int, bool, tensorflow::thread::EigenEnvironment)::{lambda()#1}>::_M_invoke(std::_Any_data const&) ()
from /home/zshan/deeprec-env/lib/python3.6/site-packages/tensorflow_core/python/../libtensorflow_framework.so.1
#16 0x00001464d9948f78 in std::_Function_handler<void (), tensorflow::thread::EigenEnvironment::CreateThread(std::function<void ()>)::{lambda()#1}>::_M_invoke(std::_Any_data const&) ()
from /home/zshan/deeprec-env/lib/python3.6/site-packages/tensorflow_core/python/../libtensorflow_framework.so.1
#17 0x00001464d83a9ba3 in execute_native_thread_routine () from /lib64/libstdc++.so.6
#18 0x0000146577a1a17a in start_thread () from /lib64/libpthread.so.0
#19 0x0000146576fbfdc3 in clone () from /lib64/libc.so.6