Hzfengsy commented on code in PR #12170:
URL: https://github.com/apache/tvm/pull/12170#discussion_r929437407
##########
tests/python/unittest/test_tir_analysis_calculate_workspace.py:
##########
@@ -57,11 +57,11 @@ def primfunc_global_allocates(placeholder_144: T.handle,
placeholder_145: T.hand
def primfunc_local_allocates(placeholder_162: T.handle, placeholder_163:
T.handle, placeholder_164: T.handle, T_cast_76: T.handle) -> None:
# function attr dict
T.func_attr({"global_symbol":
"fused_nn_conv2d_add_cast_fixed_point_multiply_clip_cast_cast_9",
"tir.noalias": True})
- sid_21 = T.allocate_const([0,1,2,3,4,5,6,7], "int8", [8])
placeholder_165 = T.match_buffer(placeholder_162, [100352], dtype="int16",
elem_offset=0, align=128, offset_factor=1)
placeholder_166 = T.match_buffer(placeholder_163, [4608], dtype="int16",
elem_offset=0, align=128, offset_factor=1)
placeholder_167 = T.match_buffer(placeholder_164, [512], dtype="int32",
elem_offset=0, align=128, offset_factor=1)
T_cast_77 = T.match_buffer(T_cast_76, [100352], dtype="int16",
elem_offset=0, align=128, offset_factor=1)
+ sid_21 = T.allocate_const([0,1,2,3,4,5,6,7], "int8", [8])
Review Comment:
Why the order matters?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
