mbrookhart commented on a change in pull request #6839:
URL: https://github.com/apache/tvm/pull/6839#discussion_r538538904
##########
File path: python/tvm/topi/cuda/nms.py
##########
@@ -97,47 +97,44 @@ def get_valid_counts_ir(
valid_count = ib.buffer_ptr(valid_count)
out = ib.buffer_ptr(out)
out_indices = ib.buffer_ptr(out_indices)
- atomic_add_return = ib.allocate(
- valid_count.dtype, (1,), name="atomic_add_return", scope="local"
- )
- one_count = tvm.tir.const(1, dtype=valid_count.dtype)
one = tvm.tir.const(1, dtype=out.dtype)
- score_threshold = tvm.ir.make_node("FloatImm", dtype="float32",
value=score_threshold)
+ if isinstance(score_threshold, float):
+ score_threshold = tvm.ir.make_node("FloatImm", dtype="float32",
value=score_threshold)
id_index = tvm.ir.make_node("IntImm", dtype="int32", value=id_index)
score_index = tvm.ir.make_node("IntImm", dtype="int32", value=score_index)
max_threads =
int(tvm.target.Target.current(allow_none=False).max_num_threads)
- nthread_tx = max_threads
- nthread_bx = batch_size * num_anchors // max_threads + 1
- tx = te.thread_axis("threadIdx.x")
- bx = te.thread_axis("blockIdx.x")
- ib.scope_attr(tx, "thread_extent", nthread_tx)
- ib.scope_attr(bx, "thread_extent", nthread_bx)
- tid = bx * max_threads + tx
- idxd = tvm.tir.indexdiv
-
- # initialize valid_count
- with ib.if_scope(tid < batch_size):
- valid_count[tid] = 0
- with ib.if_scope(tid < batch_size * num_anchors):
- i = idxd(tid, num_anchors)
- with ib.if_scope(
- tvm.tir.all(
- data[tid * elem_length + score_index] > score_threshold,
- tvm.tir.any(id_index < 0, data[tid * elem_length + id_index]
>= 0),
- )
- ):
- atomic_add_return[0] = atomic_add(
- tvm.tir.call_intrin("handle", "tir.address_of",
valid_count[i]), one_count
- )
- with ib.for_range(0, elem_length) as k:
- out[tid * elem_length + k] = data[tid * elem_length + k]
- out_indices[tid + k] = tid + k
- with ib.else_scope():
- with ib.for_range(0, elem_length) as k:
- out[tid * elem_length + k] = -one
- out_indices[tid + k] = -one_count
-
+ with ib.new_scope():
+ nthread_tx = max_threads
+ nthread_bx = batch_size // max_threads + 1
+ tx = te.thread_axis("threadIdx.x")
+ bx = te.thread_axis("blockIdx.x")
+ ib.scope_attr(tx, "thread_extent", nthread_tx)
+ ib.scope_attr(bx, "thread_extent", nthread_bx)
+ tid = bx * max_threads + tx
+ with ib.if_scope(tid < batch_size):
+ valid_count[tid] = 0
+ i = tid
+ with ib.for_range(0, num_anchors) as j:
+ score = data[(i * num_anchors + j) * elem_length + score_index]
+ with ib.if_scope(
+ tvm.tir.all(
+ score > score_threshold,
+ tvm.tir.any(
+ id_index < 0, data[(i * num_anchors + j) *
elem_length + id_index] >= 0
+ ),
+ )
+ ):
+ with ib.for_range(0, elem_length) as k:
+ out[(i * num_anchors + valid_count[i]) * elem_length +
k] = data[
+ (i * num_anchors + j) * elem_length + k
+ ]
+ out_indices[i * num_anchors + valid_count[i]] = j
+ valid_count[i] += 1
Review comment:
The test is disabled on master:
https://github.com/apache/tvm/blob/7a0d10ce8d4df27734949b944992e5c516f620d0/tests/python/relay/test_op_level5.py#L317-L321
The current kernel doesn't follow the spec, specifically, it uses the wrong
index for the out_indices update:
```
out_indices[tid + k] = tid + k
```
instead of
```
out_indices[i * num_anchors + valid_count[i]] = j
```
Unfortunately, because valid_count depends on the iteration order of the
loop on this axis, I don't see a way to get the correct index without
serializing this axis.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
