Skip to content

Add dtype parameter to CumsumKernel #1774

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 3 commits into from
Closed

Add dtype parameter to CumsumKernel #1774

Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
094e43e
fix cumsum
aquagull Jun 20, 2025
8ee2c84
fix attributeMap
aquagull Jun 20, 2025
bb7b937
fix
aquagull Jun 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 11 additions & 5 deletions backends/npu/kernels/cum_kernel.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -72,13 +72,16 @@ void AclopCumsumKernel(const Context& dev_ctx,
bool flatten,
bool exclusive,
bool reverse,
phi::DataType dtype,
phi::DenseTensor* out) {
dev_ctx.template Alloc<T>(out);

auto axis = axis_scalar.to<int>();

NPUAttributeMap attr_input = {
{"axis", axis}, {"exclusive", exclusive}, {"reverse", reverse}};
NPUAttributeMap attr_input = {{"axis", axis},
{"exclusive", exclusive},
{"reverse", reverse},
{"dtype", static_cast<int>(dtype)}};

if (flatten) {
PADDLE_ENFORCE_EQ(
Expand All @@ -105,11 +108,12 @@ void CumsumKernel(const Context& dev_ctx,
bool flatten,
bool exclusive,
bool reverse,
phi::DataType dtype,
phi::DenseTensor* out) {
DO_COMPATIBILITY(
aclnnCumsumV2,
(custom_kernel::AclopCumsumKernel<T, Context>(
dev_ctx, x, axis_scalar, flatten, exclusive, reverse, out)));
dev_ctx, x, axis_scalar, flatten, exclusive, reverse, dtype, out)));
dev_ctx.template Alloc<T>(out);
auto axis = axis_scalar.to<int64_t>();

Expand All @@ -125,9 +129,11 @@ void CumsumKernel(const Context& dev_ctx,
Tensor new_x(x);
new_x.Resize(phi::make_ddim({x.numel()}));

EXEC_NPU_CMD(aclnnCumsumV2, dev_ctx, new_x, axis, exclusive, reverse, *out);
EXEC_NPU_CMD(
aclnnCumsumV2, dev_ctx, new_x, axis, exclusive, reverse, dtype, *out);
} else {
EXEC_NPU_CMD(aclnnCumsumV2, dev_ctx, x, axis, exclusive, reverse, *out);
EXEC_NPU_CMD(
aclnnCumsumV2, dev_ctx, x, axis, exclusive, reverse, dtype, *out);
}
}

Expand Down