Laurawly commented on a change in pull request #6839:
URL: https://github.com/apache/tvm/pull/6839#discussion_r539693783
##########
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:
Yes, that method was actually my first implementation of nms before PR
#5339. And in that PR, what I improved was only getting the correct valid_count
from nms_ir, and marking -1 to invalid outputs. And after nms_ir, I do a sort
(in descending order), to get the sorted outputs as well as sorted output
indicies. The second implementation is currently in main and much faster than
the previous one.
----------------------------------------------------------------
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]
