From: @tronzhang
Reviewed-by: @gaoxiong1,@ckey_dou
Signed-off-by: @ckey_dou

fix compiling

Signed-off-by: zhoufeng <zhoufeng54@huawei.com>

when the attr is a list that mixed with Int32Imm and Int64Imm (unreasonable, but it occured),
it will crash whether we use GetValue<std::vector<int>> or GetValue<std::vector<int64_t>>.
so we need to traverse the list and pick the numbers manually.

From: @lingyunli63
Reviewed-by: @gaoxiong1,@ckey_dou
Signed-off-by: @ckey_dou

From: @dayschan
Reviewed-by: @gaoxiong1,@gaoxiong1,@anyrenwei
Signed-off-by: @anyrenwei

Decoupled from the front-end interfaces.
1. Removed the call to "Renormalize".
   Completed the infer-format in model_builder.
   Only used the device shape and device format to
   infer an abstract shape without considering padding.
2. Removed the call to python's Primitive interfaces.
   The "Renormalize" relies on the PrimitivePy, so they can be
   removed together. After that, the functions "ConstAttrToInput",
   "DeleteAttrInInput" and related can be removed.
3. Reuse the AkgKernelJsonGenerator in GraphKernelExpander.
   1) set the attribute "extract_opinfo_from_anf" to true, so that
   the generator can handle the basic operator with anf info.
   2) added a function "extract_expand_info" in expander.py
   to convert the json into a more friendly format. The attrs
   was converted to a dict instead of a list.
4. Scalars only support DefaultFormat.
   Removed the argument "format" from graph_builder.value
5. Moved the expander op list from graph_kernel_helper.cc to graph_kernel_expander.cc

add tile expander

add BroadcastTo in model

fix BroadcastTo op calling error and infer shape

rewrite tile expander

not split broadcast_to

add SqrtGrad expander

1. Extract op info from AnfNode when generating composite operators.
2. Ignore the op register info when selecting kernel.
3. Update akg submodule.
4. Remove akg gpu op registers that added for GraphKernel.
5. Add an interface `IsNodeInGraphKernel` into `AnfRuntimeAlgorithm`, next to the `IsGraphKernel`.
6. Sort included fils for the clang-format rule is changed.

TODO: Remove the json generator in GraphKernelExpander, it's ok to call this interface.

1. Update akg submodule
2. Refactor akg_kernel_build, akg_ascend_kernel_build, akg_gpu_kernel_build
3. Add akg_kernel_json_decoder to support converting kernel_json to AnfNode.
4. Add GraphKernel Cost Model. (mindspore/_extends/graph_kernel)
5. Add some GraphKernel passes to GpuSession, move these passes to backend/optimizer/graph_kernel.
6. Add global id for ir files.
7. Fix bug in ConstInputToAttr.

replace frontend/operator/ops.h in backend with base/core_ops.h as
backend should not use any frontend-only primitive

Signed-off-by: zhoufeng <zhoufeng54@huawei.com>

Merge pull request !3938 from liubuyu/master