Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
H
hpvm-release
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Model registry
Operate
Environments
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
llvm
hpvm-release
Commits
9bc2ca73
Commit
9bc2ca73
authored
5 years ago
by
Hashim Sharif
Browse files
Options
Downloads
Patches
Plain Diff
Grouped Conv support in tensorHalfConvolution
parent
a9f49265
No related branches found
Branches containing commit
No related tags found
Tags containing commit
No related merge requests found
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
llvm/projects/hpvm-tensor-rt/tensor_runtime/include/half_precision_api.h
+11
-2
11 additions, 2 deletions
...pvm-tensor-rt/tensor_runtime/include/half_precision_api.h
with
11 additions
and
2 deletions
llvm/projects/hpvm-tensor-rt/tensor_runtime/include/half_precision_api.h
+
11
−
2
View file @
9bc2ca73
...
...
@@ -149,7 +149,7 @@ void* tensorHalfGemmGPU(void* lhs_ptr, void* rhs_ptr){
void
*
tensorHalfConvolution
(
void
*
input_ptr
,
void
*
filter_ptr
,
int
vertical_pad
,
int
horizontal_pad
,
int
vertical_stride
,
int
horizontal_stride
,
int
conv_mode
,
int
co
mpute_precision
){
int
conv_mode
,
int
co
nv_groups
){
INFO
(
"*** TensorHConvolution
\n
"
);
profileEvent
(
"tensorHalfConv"
);
...
...
@@ -200,8 +200,17 @@ void* tensorHalfConvolution(void* input_ptr, void* filter_ptr,
/******* END OF INPUT DATA CONVERSIONS*/
checkCUDNN
(
cudnnCreateConvolutionDescriptor
(
&
convDesc
));
//FIXME: Current hack to preserve backward compatibilty
if
(
conv_groups
==
0
){
conv_groups
=
1
;
}
// NOTE: Adding support for grouped convolution
checkCUDNN
(
cudnnSetConvolutionGroupCount
(
convDesc
,
conv_groups
));
// FIXIT: Think if upscaling values need to be configurable?
// IMP-FIXIT: CUDNN Cross correlation is only used in the Lenet context
// IMP-FIXIT: Either make mode configurable OR see if CUDNN_CONVOLUTION MODE should be used?
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment