Skip to content

keep a cpu path for embeddingbag & fall back to cpu if not all input tensors are dpcpp tensors #78

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
zhuhaozhe wants to merge 2 commits into from
Closed

keep a cpu path for embeddingbag & fall back to cpu if not all input tensors are dpcpp tensors #78

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
636a5b1
keep-embedding_bag-cpu-path
zhuhaozhe Jun 17, 2020
55c6ba0
check-if-all-input-tensors-are-dpcpp-tensors
zhuhaozhe Jun 17, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 6 additions & 1 deletion intel_pytorch_extension_py/ops/embeddingbag.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,8 +15,13 @@ def embeddingbag(weights, inputs, offsets, scale_grad_by_freq, mode, sparse, per
else:
assert(0, "unimplement embeddingbag path in extension")
'''
torch_embedding_bag = torch.embedding_bag

def embeddingbag(weights, indices, offsets, scale_grad_by_freq, mode, sparse, per_sample_weights, include_last_offset):
ret = EmbeddingBagFunction.apply(weights, indices, offsets, scale_grad_by_freq, mode, sparse, per_sample_weights, include_last_offset)
if indices.device==torch.device("dpcpp"):
ret = EmbeddingBagFunction.apply(weights, indices, offsets, scale_grad_by_freq, mode, sparse, per_sample_weights, include_last_offset)
else:
ret = torch_embedding_bag(weights, indices, offsets, scale_grad_by_freq, mode, sparse, per_sample_weights, include_last_offset)
return ret


Expand Down
13 changes: 12 additions & 1 deletion torch_ipex/csrc/cpu/dbl/DNNLChecker.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,7 +11,8 @@ namespace chk {
bool dnnl_support_the_tensors(const std::vector<at::Tensor> &tensor_vec) {
return dnnl_tensor_has_data(tensor_vec) &&
dnnl_support_the_dimension_of(tensor_vec) &&
dnnl_support_the_data_type_of(tensor_vec);
dnnl_support_the_data_type_of(tensor_vec) &&
dnnl_support_the_device_type_of(tensor_vec);
}

bool dnnl_inplace_support_the_tensors(const std::vector<at::Tensor> &tensor_vec) {
Expand Down Expand Up @@ -44,6 +45,16 @@ bool dnnl_support_the_data_type_of(const std::vector<at::Tensor> &tensor_vec) {
return true;
}

bool dnnl_support_the_device_type_of(const std::vector<at::Tensor> &tensor_vec) {
for (auto it = tensor_vec.begin(); it != tensor_vec.end(); ++it) {
if (!it.device().is_dpcpp()) {
return false;
}
}

return true;
}

bool dnnl_support_the_dimension_of(const std::vector<at::Tensor> &tensor_vec) {
for (auto it = tensor_vec.begin(); it != tensor_vec.end(); ++it) {
if (it->dim() <= 0) {
Expand Down
8 changes: 8 additions & 0 deletions torch_ipex/csrc/cpu/dbl/DNNLChecker.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,6 +52,14 @@ bool dnnl_support_the_memory_layout_of(const std::vector<at::Tensor> &tensor_vec
*/
bool dnnl_support_the_data_type_of(const std::vector<at::Tensor> &tensor_vec);

/**
* Check if the device type of the input tenosrs can be supported by DNNL
*
* @param tensor_vec input tensors
*
*/
bool dnnl_support_the_device_type_of(const std::vector<at::Tensor> &tensor_vec);

/**
* Check if the dimension of the input tenosrs can be supported by DNNL. The dimension
* of the input tensor should be > 0.
Expand Down