* In llvm, we have a combine pass to transform some operators to another
operator , sush as mull + add --> mla/madd.
so I make a simple demo to test it in tvm(tbe), and find it is not done in
**tir passes** (common pass for tvm and tbe), does TVM have similar pass to
combine the operators?
> PS：I don't want it is been done in graph as such operator may be the part of
> some operator , which are not independent operator.
* test:
```
(py310) /data/zhongyunde/source/test/dsl # cat testMadd.py
# import tvm
import te.lang.cce
from tbe import tvm
from tbe import dsl
from tbe.common.platform import set_current_compile_soc_info
set_current_compile_soc_info("Ascend310P3")
shape = (280000,280000)
dtype = "float16"
data = tvm.placeholder(shape, name="data", dtype=dtype)
with tvm.target.cce():
const_val = tvm.const(-1, "float32")
mul_val = te.lang.cce.vmuls(data, const_val)
res = te.lang.cce.vadds(mul_val, const_val)
sch = dsl.auto_schedule(res)
print (res)
print (sch)
config = { "name" : "abs_28_28_float16", "tensor_list" : [data,res]}
dsl.build(sch, config)
```
---
[Visit
Topic](https://discuss.tvm.apache.org/t/which-pass-is-expected-to-fused-or-combined-operator-in-tvm/18398/1)
to respond.
You are receiving this because you enabled mailing list mode.
To unsubscribe from these emails, [click
here](https://discuss.tvm.apache.org/email/unsubscribe/3cf4bcdf0c69b98ccbee5c8a0cff53604325dda2b3024a52a5c96daf1ed4267f).
