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_ROCM
Commits
1ccb8112
"vscode:/vscode.git/clone" did not exist on "bb42498c33769239b48c7775eb4a0ba3d44c0907"
Commit
1ccb8112
authored
Nov 04, 2024
by
Andriy Roshchenko
Browse files
Prevent instantiation of operators that are not supported by FP8 data types
parent
97a5cca9
Changes
1
Show whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
3 additions
and
2 deletions
+3
-2
example/44_elementwise_permute/elementwise_scale_permute_amax_2D_fp16_fp8.cpp
...se_permute/elementwise_scale_permute_amax_2D_fp16_fp8.cpp
+3
-2
No files found.
example/44_elementwise_permute/elementwise_scale_permute_amax_2D_fp16_fp8.cpp
View file @
1ccb8112
...
@@ -68,7 +68,7 @@ using DeviceElementwisePermuteInstance = ck::tensor_operation::device::DeviceEle
...
@@ -68,7 +68,7 @@ using DeviceElementwisePermuteInstance = ck::tensor_operation::device::DeviceEle
using
DeviceReduceInstance
=
using
DeviceReduceInstance
=
ck
::
tensor_operation
::
device
::
DeviceReduceMultiBlock
<
OutputDataType
,
ck
::
tensor_operation
::
device
::
DeviceReduceMultiBlock
<
OutputDataType
,
Output
DataType
,
Scale
DataType
,
OutputDataType
,
OutputDataType
,
NumDim
,
NumDim
,
NumDim
,
NumDim
,
...
@@ -108,7 +108,8 @@ void reference_scale_permute_amax(Tensor<InputDataType>& input,
...
@@ -108,7 +108,8 @@ void reference_scale_permute_amax(Tensor<InputDataType>& input,
host_output_scaled_casted_transposed
(
m
,
k
)
=
y1
;
host_output_scaled_casted_transposed
(
m
,
k
)
=
y1
;
const
OutputDataType
y_fabs
=
const
OutputDataType
y_fabs
=
ck
::
type_convert
<
OutputDataType
>
(
ck
::
math
::
abs
(
ck
::
type_convert
<
float
>
(
y0
)));
ck
::
type_convert
<
OutputDataType
>
(
ck
::
math
::
abs
(
ck
::
type_convert
<
float
>
(
y0
)));
host_output_amax
(
0
)
=
ck
::
math
::
max
(
y_fabs
,
host_output_amax
(
0
));
host_output_amax
(
0
)
=
ck
::
type_convert
<
OutputDataType
>
(
ck
::
math
::
max
(
ck
::
type_convert
<
float
>
(
y_fabs
),
ck
::
type_convert
<
float
>
(
host_output_amax
(
0
))));
}
}
}
}
}
}
...
...
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