From: @r1chardf1d0
Reviewed-by: @gaoxiong1,@ckey_dou
Signed-off-by: @ckey_dou

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

From: @yepei6
Reviewed-by: 
Signed-off-by:

used the flag "opt_level" to control GraphKernel,
0 means disabled while non-zero value means enabled.
the default value is controlled by context "enable_graph_kernel",
but if it's also set in "graph_kernel_flags", then the flag will prevail.

supported the whitelist and blacklist operators for GraphKernelExpander.
"enable_expand_ops", "enable_expand_ops_only", "disable_expand_ops".

add hostpush part modify

optimize previous code

provide aclhandle access method

modify CMakeList format

add device_id parameter into TransferNode

update acltdt api

add hostpush part modify

optimize previous code

provide aclhandle access method

modify CMakeList format

add device_id parameter into TransferNode

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

Merge pull request !2807 from riemann_penn/apply_indexed_slices

Merge pull request !2421 from yankai10/merge_master_new_fearture

- It provides a unified style to express graph and kernel for user.
- It provides a unified IR to represent graph and kernel for developer.
- It breaks the boundary between graph and kernel.
- It provides more opportunities to do compile optimization.

Merge pull request !1897 from jinyaohui/print

Merge pull request !2065 from ghzl/dump_ge_switch_add_dumpMode