Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
gaoqiong
composable_kernel
Commits
b920e225
"vscode:/vscode.git/clone" did not exist on "d3ce6f4b1eeb9ce50698a61701f5b52457ef41e2"
Commit
b920e225
authored
Oct 06, 2023
by
Rostyslav Geyyer
Browse files
Fix conv bwd data instance build
parent
150e1b33
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
12 additions
and
7 deletions
+12
-7
library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_data/device_grouped_conv_bwd_data_xdl_instance.hpp
...nv_bwd_data/device_grouped_conv_bwd_data_xdl_instance.hpp
+12
-7
No files found.
library/include/ck/library/tensor_operation_instance/gpu/grouped_conv_bwd_data/device_grouped_conv_bwd_data_xdl_instance.hpp
View file @
b920e225
...
@@ -18,8 +18,12 @@ namespace instance {
...
@@ -18,8 +18,12 @@ namespace instance {
using
BF16
=
ck
::
bhalf_t
;
using
BF16
=
ck
::
bhalf_t
;
using
F16
=
ck
::
half_t
;
using
F16
=
ck
::
half_t
;
using
F32
=
float
;
using
F32
=
float
;
using
BF8
=
ck
::
bf8_t
;
#ifdef CK_ENABLE_BF8
using
F8
=
ck
::
f8_t
;
using
BF8
=
ck
::
bf8_t
;
#endif
#ifdef CK_ENABLE_FP8
using
F8
=
ck
::
f8_t
;
#endif
using
Empty_Tuple
=
ck
::
Tuple
<>
;
using
Empty_Tuple
=
ck
::
Tuple
<>
;
...
@@ -152,13 +156,13 @@ template <index_t NDimSpatial,
...
@@ -152,13 +156,13 @@ template <index_t NDimSpatial,
typename
DsLayout
,
typename
DsLayout
,
typename
ELayout
,
typename
ELayout
,
ConvolutionBackwardDataSpecialization
ConvSpec
>
ConvolutionBackwardDataSpecialization
ConvSpec
>
using
device_grouped_conv_bwd_data_xdl_input_fp16_comp_bf8f8_instances
=
using
device_grouped_conv_bwd_data_xdl_input_fp16_comp_bf8f8_instances
=
std
::
tuple
<
std
::
tuple
<
// clang-format off
// clang-format off
// ##############################################| NDim| ALayout| BLayout| DsLayout| ELayout| AData| BData| AccData| CShuffle| DsData| EData| AElementwise| BElementwise| CDEElementwise| ConvolutionBackward| DoPad| DoPad| NumGemmK| Block| MPer| NPer| KPer| AK1| BK1| MPer| NPer| MXdl| NXdl| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffleMXdl| CShuffleNXdl| CDEBlockTransfer| CDEBlockTransfer|
// ##############################################| NDim| ALayout| BLayout| DsLayout| ELayout| AData| BData| AccData| CShuffle| DsData| EData| AElementwise| BElementwise| CDEElementwise| ConvolutionBackward| DoPad| DoPad| NumGemmK| Block| MPer| NPer| KPer| AK1| BK1| MPer| NPer| MXdl| NXdl| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockTransfer| ABlockLds| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockTransfer| BBlockLds| CShuffleMXdl| CShuffleNXdl| CDEBlockTransfer| CDEBlockTransfer|
// ##############################################| Spatial| | | | | Type| Type| Type| DataType| Type| Type| Operation| Operation| Operation| DataSpecialization| GemmM| GemmN| PrefetchStage| Size| Block| Block| Block| | | XDL| XDL| PerWave| PerWave| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| ExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| ExtraN| PerWave| PerWave| _MBlock_MPerBlock| ScalarPerVector|
// ##############################################| Spatial| | | | | Type| Type| Type| DataType| Type| Type| Operation| Operation| Operation| DataSpecialization| GemmM| GemmN| PrefetchStage| Size| Block| Block| Block| | | XDL| XDL| PerWave| PerWave| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| ExtraM| ThreadCluster| ThreadCluster| SrcAccessOrder| SrcVectorDim| SrcScalar| DstScalar| ExtraN| PerWave| PerWave| _MBlock_MPerBlock| ScalarPerVector|
// ##############################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | Lengths_AK0_M_AK1| ArrangeOrder| | | PerVector| PerVector_AK1| | Lengths_BK0_N_BK1| ArrangeOrder| | | PerVector| PerVector_BK1| | PerShuffle| PerShuffle| _NBlock_NPerBlock| _NPerBlock|
// ##############################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | Lengths_AK0_M_AK1| ArrangeOrder| | | PerVector| PerVector_AK1| | Lengths_BK0_N_BK1| ArrangeOrder| | | PerVector| PerVector_BK1| | PerShuffle| PerShuffle| _NBlock_NPerBlock| _NPerBlock|
// ##############################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
// ##############################################| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | |
#if defined CK_ENABLE_FP8 && defined CK_ENABLE_BF8
// generic instance
// generic instance
DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
<
NDimSpatial
,
ALayout
,
BLayout
,
DsLayout
,
ELayout
,
F16
,
F16
,
F32
,
F32
,
Empty_Tuple
,
F16
,
PassThrough
,
PassThrough
,
PassThrough
,
ConvSpec
,
true
,
true
,
1
,
64
,
64
,
64
,
32
,
8
,
8
,
32
,
32
,
2
,
2
,
S
<
4
,
16
,
1
>
,
S
<
1
,
0
,
2
>
,
S
<
1
,
0
,
2
>
,
2
,
1
,
4
,
1
,
S
<
4
,
16
,
1
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
1
,
4
,
1
,
1
,
1
,
S
<
1
,
16
,
1
,
4
>
,
1
,
LoopScheduler
::
Default
,
BF8
,
F8
>
,
DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
<
NDimSpatial
,
ALayout
,
BLayout
,
DsLayout
,
ELayout
,
F16
,
F16
,
F32
,
F32
,
Empty_Tuple
,
F16
,
PassThrough
,
PassThrough
,
PassThrough
,
ConvSpec
,
true
,
true
,
1
,
64
,
64
,
64
,
32
,
8
,
8
,
32
,
32
,
2
,
2
,
S
<
4
,
16
,
1
>
,
S
<
1
,
0
,
2
>
,
S
<
1
,
0
,
2
>
,
2
,
1
,
4
,
1
,
S
<
4
,
16
,
1
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
1
,
4
,
1
,
1
,
1
,
S
<
1
,
16
,
1
,
4
>
,
1
,
LoopScheduler
::
Default
,
BF8
,
F8
>
,
// instances for small conv.K and conv.C
// instances for small conv.K and conv.C
...
@@ -179,8 +183,9 @@ using device_grouped_conv_bwd_data_xdl_input_fp16_comp_bf8f8_instances =
...
@@ -179,8 +183,9 @@ using device_grouped_conv_bwd_data_xdl_input_fp16_comp_bf8f8_instances =
DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
<
NDimSpatial
,
ALayout
,
BLayout
,
DsLayout
,
ELayout
,
F16
,
F16
,
F32
,
F32
,
Empty_Tuple
,
F16
,
PassThrough
,
PassThrough
,
PassThrough
,
ConvSpec
,
true
,
true
,
1
,
128
,
32
,
128
,
32
,
8
,
8
,
32
,
32
,
1
,
2
,
S
<
4
,
32
,
1
>
,
S
<
1
,
0
,
2
>
,
S
<
1
,
0
,
2
>
,
2
,
4
,
4
,
1
,
S
<
4
,
32
,
1
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
4
,
4
,
1
,
1
,
1
,
S
<
1
,
16
,
1
,
8
>
,
4
,
LoopScheduler
::
Default
,
BF8
,
F8
>
,
DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
<
NDimSpatial
,
ALayout
,
BLayout
,
DsLayout
,
ELayout
,
F16
,
F16
,
F32
,
F32
,
Empty_Tuple
,
F16
,
PassThrough
,
PassThrough
,
PassThrough
,
ConvSpec
,
true
,
true
,
1
,
128
,
32
,
128
,
32
,
8
,
8
,
32
,
32
,
1
,
2
,
S
<
4
,
32
,
1
>
,
S
<
1
,
0
,
2
>
,
S
<
1
,
0
,
2
>
,
2
,
4
,
4
,
1
,
S
<
4
,
32
,
1
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
4
,
4
,
1
,
1
,
1
,
S
<
1
,
16
,
1
,
8
>
,
4
,
LoopScheduler
::
Default
,
BF8
,
F8
>
,
DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
<
NDimSpatial
,
ALayout
,
BLayout
,
DsLayout
,
ELayout
,
F16
,
F16
,
F32
,
F32
,
Empty_Tuple
,
F16
,
PassThrough
,
PassThrough
,
PassThrough
,
ConvSpec
,
true
,
true
,
1
,
64
,
64
,
32
,
32
,
8
,
8
,
32
,
32
,
2
,
1
,
S
<
4
,
16
,
1
>
,
S
<
1
,
0
,
2
>
,
S
<
1
,
0
,
2
>
,
2
,
4
,
4
,
1
,
S
<
4
,
8
,
1
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
4
,
4
,
1
,
1
,
1
,
S
<
1
,
16
,
1
,
4
>
,
4
,
LoopScheduler
::
Default
,
BF8
,
F8
>
,
DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
<
NDimSpatial
,
ALayout
,
BLayout
,
DsLayout
,
ELayout
,
F16
,
F16
,
F32
,
F32
,
Empty_Tuple
,
F16
,
PassThrough
,
PassThrough
,
PassThrough
,
ConvSpec
,
true
,
true
,
1
,
64
,
64
,
32
,
32
,
8
,
8
,
32
,
32
,
2
,
1
,
S
<
4
,
16
,
1
>
,
S
<
1
,
0
,
2
>
,
S
<
1
,
0
,
2
>
,
2
,
4
,
4
,
1
,
S
<
4
,
8
,
1
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
4
,
4
,
1
,
1
,
1
,
S
<
1
,
16
,
1
,
4
>
,
4
,
LoopScheduler
::
Default
,
BF8
,
F8
>
,
DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
<
NDimSpatial
,
ALayout
,
BLayout
,
DsLayout
,
ELayout
,
F16
,
F16
,
F32
,
F32
,
Empty_Tuple
,
F16
,
PassThrough
,
PassThrough
,
PassThrough
,
ConvSpec
,
true
,
true
,
1
,
64
,
32
,
64
,
32
,
8
,
8
,
32
,
32
,
1
,
2
,
S
<
4
,
16
,
1
>
,
S
<
1
,
0
,
2
>
,
S
<
1
,
0
,
2
>
,
2
,
8
,
8
,
1
,
S
<
4
,
8
,
1
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
8
,
8
,
1
,
1
,
1
,
S
<
1
,
16
,
1
,
4
>
,
8
,
LoopScheduler
::
Default
,
BF8
,
F8
>
DeviceGroupedConvBwdDataMultipleD_Xdl_CShuffle_v1
<
NDimSpatial
,
ALayout
,
BLayout
,
DsLayout
,
ELayout
,
F16
,
F16
,
F32
,
F32
,
Empty_Tuple
,
F16
,
PassThrough
,
PassThrough
,
PassThrough
,
ConvSpec
,
true
,
true
,
1
,
64
,
32
,
64
,
32
,
8
,
8
,
32
,
32
,
1
,
2
,
S
<
4
,
16
,
1
>
,
S
<
1
,
0
,
2
>
,
S
<
1
,
0
,
2
>
,
2
,
8
,
8
,
1
,
S
<
4
,
8
,
1
>
,
S
<
0
,
2
,
1
>
,
S
<
0
,
2
,
1
>
,
1
,
8
,
8
,
1
,
1
,
1
,
S
<
1
,
16
,
1
,
4
>
,
8
,
LoopScheduler
::
Default
,
BF8
,
F8
>
// clang-format on
#endif
>
;
// clang-format on
>
;
}
// namespace instance
}
// namespace instance
}
// namespace device
}
// namespace device
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment