ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-14 02:02:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

Revert "[SWDEV-435347] disable instances failed with mainlien compiler (#1077)" (#1101)

Browse Source 
This reverts commit df67a32c08db7a0eddc3b3d197d9f2d7c7e590d8.

[ROCm/composable_kernel commit: a35e466c86]
This commit is contained in:
Bartłomiej Kocot
2024-01-02 11:36:45 +01:00
committed by GitHub
parent 6e776f21d9
commit fda0e3e6ee
2 changed files with 56 additions and 60 deletions
Download Patch File Download Diff File Expand all files Collapse all files

43
library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_scaleadd_ab.hpp
View File

@@ -23,20 +23,19 @@ using ScaleAdd = ck::tensor_operation::element_wise::ScaleAdd;
#ifdef CK_ENABLE_BF16
// grouped conv3d forward multi AB scaleadd, NDHWGC/GKZYXC/NDHWGK
// TODO: Workaround for https://ontrack-internal.amd.com/browse/SWDEV-435347
// void add_device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_bf16_instances(
// std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<3,
// NDHWGC,
// GKZYXC,
// ck::Tuple<>,
// NDHWGK,
// ck::Tuple<BF16, BF16>,
// ck::Tuple<BF16, BF16>,
// ck::Tuple<>,
// BF16,
// ScaleAdd,
// ScaleAdd,
// PassThrough>>>& instances);
void add_device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_bf16_instances(
std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<3,
NDHWGC,
GKZYXC,
ck::Tuple<>,
NDHWGK,
ck::Tuple<BF16, BF16>,
ck::Tuple<BF16, BF16>,
ck::Tuple<>,
BF16,
ScaleAdd,
ScaleAdd,
PassThrough>>>& instances);
#endif
#ifdef CK_ENABLE_FP16
@@ -152,15 +151,13 @@ struct DeviceOperationInstanceFactory<ck::tensor_operation::device::DeviceGroupe
}
#endif
#ifdef CK_ENABLE_BF16
// TODO: Workaround for https://ontrack-internal.amd.com/browse/SWDEV-435347
// if constexpr(is_same_v<InDataType, ck::Tuple<ck::bhalf_t, ck::bhalf_t>> &&
// is_same_v<WeiDataType, ck::Tuple<ck::bhalf_t, ck::bhalf_t>> &&
// is_same_v<OutDataType, ck::bhalf_t> && is_same_v<ComputeType,
// ck::bhalf_t>)
// {
// add_device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_bf16_instances(
// op_ptrs);
// }
if constexpr(is_same_v<InDataType, ck::Tuple<ck::bhalf_t, ck::bhalf_t>> &&
is_same_v<WeiDataType, ck::Tuple<ck::bhalf_t, ck::bhalf_t>> &&
is_same_v<OutDataType, ck::bhalf_t> && is_same_v<ComputeType, ck::bhalf_t>)
{
add_device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_bf16_instances(
op_ptrs);
}
#endif
#ifdef CK_ENABLE_INT8
if constexpr(is_same_v<InDataType, ck::Tuple<int8_t, int8_t>> &&

73
library/src/tensor_operation_instance/gpu/grouped_conv3d_fwd_scaleadd_ab/xdl/device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_bf16_instance.cpp
View File

@@ -9,43 +9,42 @@ namespace tensor_operation {
namespace device {
namespace instance {
// TODO: Workaround for https://ontrack-internal.amd.com/browse/SWDEV-435347
// void add_device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_bf16_instances(
// std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<3,
// NDHWGC,
// GKZYXC,
// ck::Tuple<>,
// NDHWGK,
// ck::Tuple<BF16, BF16>,
// ck::Tuple<BF16, BF16>,
// ck::Tuple<>,
// BF16,
// ScaleAdd,
// ScaleAdd,
// PassThrough>>>& instances)
// {
// add_device_operation_instances(
// instances,
// device_grouped_conv_fwd_xdl_scaleadd_ab_bf16_instances<3,
// NDHWGC,
// GKZYXC,
// NDHWGK,
// ConvFwdDefault>{});
// add_device_operation_instances(
// instances,
// device_grouped_conv_fwd_xdl_scaleadd_ab_bf16_instances<3,
// NDHWGC,
// GKZYXC,
// NDHWGK,
// ConvFwd1x1P0>{});
// add_device_operation_instances(
// instances,
// device_grouped_conv_fwd_xdl_scaleadd_ab_bf16_instances<3,
// NDHWGC,
// GKZYXC,
// NDHWGK,
// ConvFwd1x1S1P0>{});
// }
void add_device_grouped_conv3d_fwd_xdl_scaleadd_ab_ndhwgc_gkzyxc_ndhwgk_bf16_instances(
std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<3,
NDHWGC,
GKZYXC,
ck::Tuple<>,
NDHWGK,
ck::Tuple<BF16, BF16>,
ck::Tuple<BF16, BF16>,
ck::Tuple<>,
BF16,
ScaleAdd,
ScaleAdd,
PassThrough>>>& instances)
{
add_device_operation_instances(
instances,
device_grouped_conv_fwd_xdl_scaleadd_ab_bf16_instances<3,
NDHWGC,
GKZYXC,
NDHWGK,
ConvFwdDefault>{});
add_device_operation_instances(
instances,
device_grouped_conv_fwd_xdl_scaleadd_ab_bf16_instances<3,
NDHWGC,
GKZYXC,
NDHWGK,
ConvFwd1x1P0>{});
add_device_operation_instances(
instances,
device_grouped_conv_fwd_xdl_scaleadd_ab_bf16_instances<3,
NDHWGC,
GKZYXC,
NDHWGK,
ConvFwd1x1S1P0>{});
}
} // namespace instance
} // namespace device