Skip to content

Commit 30fd906

Browse files
wj-laskowskiwj-laskowski
committed
removing 1D instances
1 parent 4af57fd commit 30fd906

6 files changed

Lines changed: 3 additions & 176 deletions

File tree

‎library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_fwd/device_grouped_conv_fwd_wmma_cshufflev3_scaleadd_ab_instance.hpp‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -48,7 +48,7 @@ using device_grouped_conv_fwd_wmma_cshufflev3_scaleadd_ab_bf16_instances =
4848
//########################################| | | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Wave| Wave| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerWmma| _NWaveNPerWmma| | |
4949
//########################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
5050
// generic instance
51-
DeviceGroupedConvFwdMultipleABD_Wmma_CShuffle_V3<NDimSpatial, ALayout, BLayout, ck::Tuple<>, ELayout, ck::Tuple<BF16, BF16>, ck::Tuple<BF16, BF16>, F32, BF16, ck::Tuple<>, BF16, ScaleAdd, ScaleAdd, PassThrough, ConvSpec, GemmMNKPadding, 256, 128, 256, 32, 8, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 0, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 0, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Interwave, BlockGemmPipelineVersion::v1>
51+
DeviceGroupedConvFwdMultipleABD_Wmma_CShuffle_V3<NDimSpatial, ALayout, BLayout, ck::Tuple<>, ELayout, ck::Tuple<BF16, BF16>, ck::Tuple<BF16, BF16>, F32, BF16, ck::Tuple<>, BF16, ScaleAdd, ScaleAdd, PassThrough, ConvSpec, GemmMNKPadding, 256, 128, 256, 32, 8, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 0, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 0, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v1>
5252
#ifndef ONE_INSTANCE_PER_LIST
5353
,
5454
DeviceGroupedConvFwdMultipleABD_Wmma_CShuffle_V3<NDimSpatial, ALayout, BLayout, ck::Tuple<>, ELayout, ck::Tuple<BF16, BF16>, ck::Tuple<BF16, BF16>, F32, BF16, ck::Tuple<>, BF16, ScaleAdd, ScaleAdd, PassThrough, ConvSpec, GemmMNKPadding, 256, 128, 128, 64, 8, 8, 16, 16, 4, 2, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 8>, 1, BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v1>,
@@ -100,7 +100,7 @@ using device_grouped_conv_fwd_wmma_cshufflev3_scaleadd_ab_f16_instances =
100100
//########################################| | | | | | | | | | | | Operation| Operation| Operation| | | | | | | | | | | Wave| Wave| Lengths_K0_M_K1| ArrangeOrder| | | PerVector| PerVector_K1| | Lengths_K0_N_K1| ArrangeOrder| | | PerVector| PerVector_K1| | PerShuffle| PerShuffle| _NBlock_NWaveNPerWmma| _NWaveNPerWmma| | |
101101
//########################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
102102
// generic instance
103-
DeviceGroupedConvFwdMultipleABD_Wmma_CShuffle_V3<NDimSpatial, ALayout, BLayout, ck::Tuple<>, ELayout, ck::Tuple<F16, F16>, ck::Tuple<F16, F16>, F32, F16, ck::Tuple<>, F16, ScaleAdd, ScaleAdd, PassThrough, ConvSpec, GemmMNKPadding, 256, 128, 256, 32, 8, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 0, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 0, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Interwave, BlockGemmPipelineVersion::v1>
103+
DeviceGroupedConvFwdMultipleABD_Wmma_CShuffle_V3<NDimSpatial, ALayout, BLayout, ck::Tuple<>, ELayout, ck::Tuple<F16, F16>, ck::Tuple<F16, F16>, F32, F16, ck::Tuple<>, F16, ScaleAdd, ScaleAdd, PassThrough, ConvSpec, GemmMNKPadding, 256, 128, 256, 32, 8, 8, 16, 16, 4, 4, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 0, S<4, 64, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 0, 1, 1, S<1, 32, 1, 8>, 8, BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v1>
104104
#ifndef ONE_INSTANCE_PER_LIST
105105
,
106106
DeviceGroupedConvFwdMultipleABD_Wmma_CShuffle_V3<NDimSpatial, ALayout, BLayout, ck::Tuple<>, ELayout, ck::Tuple<F16, F16>, ck::Tuple<F16, F16>, F32, F16, ck::Tuple<>, F16, ScaleAdd, ScaleAdd, PassThrough, ConvSpec, GemmMNKPadding, 256, 128, 128, 64, 8, 8, 16, 16, 4, 2, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, S<8, 32, 1>, S<1, 0, 2>, S<1, 0, 2>, 2, 8, 8, 1, 1, 1, S<1, 32, 1, 8>, 1, BlockGemmPipelineScheduler::Intrawave, BlockGemmPipelineVersion::v1>,

‎library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward.hpp‎

Lines changed: 0 additions & 26 deletions
Original file line numberDiff line numberDiff line change
@@ -777,32 +777,6 @@ struct DeviceOperationInstanceFactory<ck::tensor_operation::device::DeviceGroupe
777777
#endif // CK_USE_WMMA_OLD
778778

779779
#ifdef CK_USE_WMMA
780-
// 1D
781-
// layout GNWC/GKXC/GNWK
782-
if constexpr(NumDimSpatial == 1 && is_same_v<InLayout, GNWC> &&
783-
is_same_v<WeiLayout, GKXC> && is_same_v<OutLayout, GNWK>)
784-
{
785-
#ifdef CK_ENABLE_FP16
786-
if constexpr(is_same_v<InDataType, half_t> && is_same_v<WeiDataType, half_t> &&
787-
is_same_v<OutDataType, half_t> && is_same_v<AComputeType, half_t> &&
788-
is_same_v<BComputeType, half_t>)
789-
{
790-
add_device_grouped_conv1d_fwd_wmma_cshufflev3_gnwc_gkxc_gnwk_f16_instances(op_ptrs);
791-
}
792-
#endif
793-
#ifdef CK_ENABLE_BF16
794-
if constexpr(is_same_v<InDataType, ck::bhalf_t> &&
795-
is_same_v<WeiDataType, ck::bhalf_t> &&
796-
is_same_v<OutDataType, ck::bhalf_t> &&
797-
is_same_v<AComputeType, ck::bhalf_t> &&
798-
is_same_v<BComputeType, ck::bhalf_t>)
799-
{
800-
add_device_grouped_conv1d_fwd_wmma_cshufflev3_gnwc_gkxc_gnwk_bf16_instances(
801-
op_ptrs);
802-
}
803-
#endif
804-
}
805-
806780
// layout NHWGC/GKYXC/NHWGK
807781
if constexpr(NumDimSpatial == 2 && is_same_v<InLayout, NHWGC> &&
808782
is_same_v<WeiLayout, GKYXC> && is_same_v<OutLayout, NHWGK>)

‎library/include/ck/library/tensor_operation_instance/gpu/grouped_convolution_forward_wmma_cshufflev3.inc‎

Lines changed: 0 additions & 34 deletions
Original file line numberDiff line numberDiff line change
@@ -8,40 +8,6 @@ namespace tensor_operation {
88
namespace device {
99
namespace instance {
1010

11-
// grouped conv1d forward, GNWC/GKXC/GNWK
12-
#ifdef CK_ENABLE_BF16
13-
void add_device_grouped_conv1d_fwd_wmma_cshufflev3_gnwc_gkxc_gnwk_bf16_instances(
14-
std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<1,
15-
GNWC,
16-
GKXC,
17-
Empty_Tuple,
18-
GNWK,
19-
BF16,
20-
BF16,
21-
Empty_Tuple,
22-
BF16,
23-
PassThrough,
24-
PassThrough,
25-
PassThrough>>>& instances);
26-
#endif
27-
28-
#ifdef CK_ENABLE_FP16
29-
void add_device_grouped_conv1d_fwd_wmma_cshufflev3_gnwc_gkxc_gnwk_f16_instances(
30-
std::vector<std::unique_ptr<DeviceGroupedConvFwdMultipleABD<1,
31-
GNWC,
32-
GKXC,
33-
Empty_Tuple,
34-
GNWK,
35-
F16,
36-
F16,
37-
Empty_Tuple,
38-
F16,
39-
PassThrough,
40-
PassThrough,
41-
PassThrough>>>& instances);
42-
#endif
43-
44-
4511
// grouped conv2d forward, NHWGC/GKYXC/NHWGK
4612
#ifdef CK_ENABLE_BF16
4713
void add_device_grouped_conv2d_fwd_wmma_cshufflev3_nhwgc_gkyxc_nhwgk_bf16_instances(

‎library/src/tensor_operation_instance/gpu/grouped_conv1d_fwd/CMakeLists.txt‎

Lines changed: 1 addition & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1,13 +1,10 @@
11
# Copyright (c) Advanced Micro Devices, Inc., or its affiliates.
22
# SPDX-License-Identifier: MIT
33

4-
# ONLY XDL_AND_WMMA_KERNELS
4+
# ONLY XDL_KERNELS
55
add_instance_library(device_grouped_conv1d_fwd_instance
66
xdl/device_grouped_conv1d_fwd_xdl_gnwc_gkxc_gnwk_bf16_instance.cpp
77
xdl/device_grouped_conv1d_fwd_xdl_gnwc_gkxc_gnwk_f16_instance.cpp
88
xdl/device_grouped_conv1d_fwd_xdl_gnwc_gkxc_gnwk_f32_instance.cpp
99
xdl/device_grouped_conv1d_fwd_xdl_gnwc_gkxc_gnwk_int8_instance.cpp
10-
11-
wmma/device_grouped_conv1d_fwd_wmma_cshufflev3_gnwc_gkxc_gnwk_bf16_instance.cpp
12-
wmma/device_grouped_conv1d_fwd_wmma_cshufflev3_gnwc_gkxc_gnwk_f16_instance.cpp
1310
)

‎library/src/tensor_operation_instance/gpu/grouped_conv1d_fwd/wmma/device_grouped_conv1d_fwd_wmma_cshufflev3_gnwc_gkxc_gnwk_bf16_instance.cpp‎

Lines changed: 0 additions & 55 deletions
This file was deleted.

‎library/src/tensor_operation_instance/gpu/grouped_conv1d_fwd/wmma/device_grouped_conv1d_fwd_wmma_cshufflev3_gnwc_gkxc_gnwk_f16_instance.cpp‎

Lines changed: 0 additions & 55 deletions
This file was deleted.

0 commit comments

Comments
 (0)