Skip to content

fix(ck): PrefetchStages on host for wave32 instances #13373

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

Merged
bartekxk merged 2 commits into from
Oct 14, 2026
Merged

fix(ck): PrefetchStages on host for wave32 instances #13373

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
77b765b
Fix pipeline prefetch stages used on host
EnricoDeg Oct 9, 2026
f802cd8
Add test
EnricoDeg Oct 9, 2026
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
2 changes: 1 addition & 1 deletion ...de/ck/tensor_operation/gpu/device/impl/device_grouped_conv_bwd_weight_xdl_cshuffle_v3.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1577,7 +1577,7 @@ struct DeviceGroupedConvBwdWeight_Xdl_CShuffleV3
const auto num_k_loop = gemm_arg.AK0 / (K0PerBlock / K1);
if constexpr(BlkGemmPipelineVer != BlockGemmPipelineVersion::v1)
{
if(num_k_loop <= GridwiseGemm64::BlockwiseGemmPipe::PrefetchStages)
if(num_k_loop <= GridwiseGemm32::BlockwiseGemmPipe::PrefetchStages)
Comment thread
EnricoDeg marked this conversation as resolved.
Resolved
{
return false;
}
Expand Down
72 changes: 63 additions & 9 deletions ...kernel/test/grouped_convnd_bwd_weight/test_grouped_convnd_bwd_weight_v3_interface_xdl.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,8 +54,8 @@ class TestGroupedConvndBwdWeight : public ::testing::Test
ck::utils::conv::ConvParam conv_param;
std::vector<ck::index_t> split_ks{-1, 2};

template <ck::index_t NDimSpatial>
bool Run()
template <typename DeviceInstance, ck::index_t NDimSpatial>
bool RunInstance()
{

const auto in_g_n_c_wis_desc =
Expand Down Expand Up @@ -94,7 +94,7 @@ class TestGroupedConvndBwdWeight : public ::testing::Test
range_copy(conv_param.input_left_pads_, begin(input_left_pads));
range_copy(conv_param.input_right_pads_, begin(input_right_pads));

auto conv = GroupedConvBwdWeightDeviceInstance{};
auto conv = DeviceInstance{};

bool is_supported = true;

Expand All @@ -121,6 +121,37 @@ class TestGroupedConvndBwdWeight : public ::testing::Test
}
return is_supported;
}

template <ck::index_t NDimSpatial>
bool Run()
{
return RunInstance<GroupedConvBwdWeightDeviceInstance, NDimSpatial>();
}
};

template <typename Tuple, ConvolutionBackwardWeightSpecialization ConvSpec>
class TestGroupedConvndBwdWeightInstanceV2 : public TestGroupedConvndBwdWeight<Tuple, ConvSpec>
{
using Base = TestGroupedConvndBwdWeight<Tuple, ConvSpec>;
// clang-format off
using GroupedConvBwdWeightDeviceInstance =
ck::tensor_operation::device::DeviceGroupedConvBwdWeight_Xdl_CShuffleV3
<Base::NDimSpatial, typename Base::InLayout, typename Base::WeiLayout, typename Base::OutLayout,
F16, F16, F16, F32, PassThrough, PassThrough, PassThrough, ConvSpec,
64, 32, 32, 32, 8, 32, 32, 1, 1,
S<4, 8, 1>, S<2, 0, 1>, S<1, 0, 2>, 1, 2, 2, false,
S<4, 16, 1>, S<2, 0, 1>, S<1, 0, 2>, 1, 2, 2, false,
1, 1, S<1, 8, 1, 8>, 2,
ck::BlockGemmPipelineScheduler::Intrawave,
ck::BlockGemmPipelineVersion::v2>;
// clang-format on

protected:
template <ck::index_t NDimSpatial>
bool Run()
{
return this->template RunInstance<GroupedConvBwdWeightDeviceInstance, NDimSpatial>();
}
};

using GNHWC = ck::tensor_layout::convolution::GNHWC;
Expand All @@ -135,21 +166,28 @@ using KernelTypes =
::testing::Types<std::tuple<GNHWK, GKYXC, GNHWC>, std::tuple<NHWGK, GKYXC, NHWGC>>;

template <typename Tuple>
class TestGroupedConvndBwdWeightV3Default
class TestGroupedConvndBwdWeightV3DefaultPipeV1
: public TestGroupedConvndBwdWeight<Tuple, ConvBwdWeightDefault>
{
};

template <typename Tuple>
class TestGroupedConvndBwdWeightV3Filter1x1
class TestGroupedConvndBwdWeightV3Filter1x1PipeV1
: public TestGroupedConvndBwdWeight<Tuple, Filter1x1Stride1Pad0>
{
};

TYPED_TEST_SUITE(TestGroupedConvndBwdWeightV3Default, KernelTypes);
TYPED_TEST_SUITE(TestGroupedConvndBwdWeightV3Filter1x1, KernelTypes);
template <typename Tuple>
class TestGroupedConvndBwdWeightV3DefaultPipeV2
: public TestGroupedConvndBwdWeightInstanceV2<Tuple, ConvBwdWeightDefault>
{
};

TYPED_TEST_SUITE(TestGroupedConvndBwdWeightV3DefaultPipeV1, KernelTypes);
TYPED_TEST_SUITE(TestGroupedConvndBwdWeightV3DefaultPipeV2, KernelTypes);
TYPED_TEST_SUITE(TestGroupedConvndBwdWeightV3Filter1x1PipeV1, KernelTypes);

TYPED_TEST(TestGroupedConvndBwdWeightV3Filter1x1, SpecializationCheck)
TYPED_TEST(TestGroupedConvndBwdWeightV3Filter1x1PipeV1, SpecializationCheck)
{
// Check filter 3,3 instead of 1,1
this->conv_param = {2, 2, 4, 192, 192, {3, 3}, {28, 28}, {1, 1}, {1, 1}, {0, 0}, {0, 0}};
Expand All @@ -172,7 +210,7 @@ TYPED_TEST(TestGroupedConvndBwdWeightV3Filter1x1, SpecializationCheck)
EXPECT_FALSE(is_supported);
}

TYPED_TEST(TestGroupedConvndBwdWeightV3Default, VectorLoadCheck)
TYPED_TEST(TestGroupedConvndBwdWeightV3DefaultPipeV1, VectorLoadCheck)
{
// vector load for A
this->conv_param = {2, 2, 128, 129, 256, {1, 1}, {7, 7}, {2, 2}, {1, 1}, {0, 0}, {0, 0}};
Expand All @@ -183,3 +221,19 @@ TYPED_TEST(TestGroupedConvndBwdWeightV3Default, VectorLoadCheck)
is_supported = this->template Run<2>();
EXPECT_FALSE(is_supported);
}

TYPED_TEST(TestGroupedConvndBwdWeightV3DefaultPipeV2, PrefetchStagesCheck)
{
this->conv_param = {2, 1, 8, 120, 120, {3, 3}, {24, 24}, {1, 1}, {1, 1}, {1, 1}, {1, 1}};
this->split_ks = {64};

bool is_supported = this->template Run<2>();
if(ck::is_gfx11_supported() || ck::is_gfx12_supported())
{
EXPECT_FALSE(is_supported);
}
else
{
EXPECT_TRUE(is_supported);
}
}
Loading