Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

Integrate MKLDNN Conv1d and support 3d layout #13530

Merged
merged 14 commits into from
Jan 2, 2019
Merged

Integrate MKLDNN Conv1d and support 3d layout #13530

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
14 commits
Select commit Hold shift + click to select a range
06901b8
add 3d layout support for MKLDNN Conv and Activation
xinyu-intel Dec 4, 2018
49296fc
fix lint
xinyu-intel Dec 4, 2018
cc6ebf3
resolve conflict
xinyu-intel Dec 5, 2018
2da1e2a
code refactor
xinyu-intel Dec 7, 2018
8f70c99
Merge remote-tracking branch 'upstream/master' into conv1d
xinyu-intel Dec 7, 2018
869ad9d
add testcase for group1 conv and skip quantization for conv1d
xinyu-intel Dec 10, 2018
5e1a5f3
Merge remote-tracking branch 'upstream/master' into conv1d
xinyu-intel Dec 10, 2018
a3f843b
fix lint
xinyu-intel Dec 10, 2018
25a0e4f
avoid conv1d quantization
xinyu-intel Dec 11, 2018
e20807f
Merge remote-tracking branch 'upstream/master' into conv1d
xinyu-intel Dec 15, 2018
b627592
Merge remote-tracking branch 'upstream/master' into conv1d
xinyu-intel Dec 21, 2018
4d51c8f
code refactor and add activation ut
xinyu-intel Dec 25, 2018
95bef09
del todo
xinyu-intel Jan 2, 2019
a13d87a
Merge remote-tracking branch 'upstream/master' into conv1d
xinyu-intel Jan 2, 2019
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
10 changes: 2 additions & 8 deletions src/ndarray/ndarray.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -454,24 +454,18 @@ void NDArray::Chunk::SetMKLMem(const TShape &shape, int dtype) {

mkldnn::memory::dims dims;
// These are shapes supprted by MKLDNN.
if (shape.ndim() == 1 || shape.ndim() == 2 || shape.ndim() == 4
|| shape.ndim() == 5) {
if (shape.ndim() >= 1 && shape.ndim() <= 5) {
dims.resize(shape.ndim());
for (size_t i = 0; i < dims.size(); i++)
dims[i] = shape[i];
} else if (shape.ndim() == 3) {
// If there are 3 dimensions, we'll force it to 4 dimensions.
dims.resize(shape.ndim() + 1);
dims[0] = 1;
for (size_t i = 0; i < shape.ndim(); i++)
dims[i + 1] = shape[i];
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is there a performance difference between 3D and 4D implementation?

} else {
LOG(FATAL) << "MKLDNN doesn't support " << shape.ndim() << " dimensions";
}
mkldnn::memory::format layout = mkldnn::memory::format::format_undef;
switch (dims.size()) {
case 1: layout = mkldnn::memory::format::x; break;
case 2: layout = mkldnn::memory::format::nc; break;
case 3: layout = mkldnn::memory::format::ncw; break;
case 4: layout = mkldnn::memory::format::nchw; break;
// This isn't the right layout when the data has 5 dimensions in MXNet.
// MXNet interprets 5 dimensions as ncdhw, but MKLDNN doesn't have
Expand Down
5 changes: 3 additions & 2 deletions src/operator/nn/activation.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -97,9 +97,10 @@ static void ActivationComputeExCPU(const nnvm::NodeAttrs& attrs,
const std::vector<NDArray>& inputs,
const std::vector<OpReqType>& req,
const std::vector<NDArray>& outputs) {
const ActivationParam& param = nnvm::get<ActivationParam>(attrs.parsed);
CHECK_EQ(inputs.size(), 1U);
CHECK_EQ(outputs.size(), 1U);
if (SupportMKLDNN(inputs[0])) {
if (SupportMKLDNNAct(param, inputs[0])) {
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
MKLDNN_OPCHECK_INIT(false, outputs.size(), inputs, outputs);
MKLDNNActivationForward(attrs, ctx, inputs[0], req[0], outputs[0]);
MKLDNN_OPCHECK_RUN(ActivationCompute<cpu>, attrs, ctx, inputs, req, outputs);
Expand All @@ -115,7 +116,7 @@ void ActivationGradComputeExCPU(const nnvm::NodeAttrs& attrs,
const std::vector<NDArray>& outputs) {
const ActivationParam& param = nnvm::get<ActivationParam>(attrs.parsed);
CHECK_EQ(inputs.size(), activation::GradNumInputs(param.act_type));
if (SupportMKLDNN(inputs[0])) {
if (SupportMKLDNNAct(param, inputs[0])) {
MKLDNN_OPCHECK_INIT(true, outputs.size(), inputs, outputs);
// XXX: for y = relu(x), y is passed as "in_data" to Backward()
const bool relu = param.act_type == activation::kReLU;
Expand Down
11 changes: 11 additions & 0 deletions src/operator/nn/mkldnn/mkldnn_act.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,17 @@ bool SupportMKLDNNAct(const ActivationParam& param) {
|| param.act_type == activation::kTanh;
}

bool SupportMKLDNNAct(const ActivationParam& param, const NDArray &input) {
if ((input.shape().ndim() < 1) ||
(input.shape().ndim() > 4) ||
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
(input.dtype() != mshadow::kFloat32))
return false;
return param.act_type == activation::kReLU
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
|| param.act_type == activation::kSigmoid
|| param.act_type == activation::kSoftReLU
|| param.act_type == activation::kTanh;
}

static inline mkldnn::algorithm GetMKLDNNActAlgo(const ActivationParam& param) {
switch (param.act_type) {
case activation::kReLU:
Expand Down
28 changes: 19 additions & 9 deletions src/operator/nn/mkldnn/mkldnn_base-inl.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -175,10 +175,11 @@ struct ConvolutionParam;
struct DeconvolutionParam;
struct SoftmaxParam;
bool SupportMKLDNNAct(const ActivationParam& param);
bool SupportMKLDNNAct(const ActivationParam& param, const NDArray &input);
bool SupportMKLDNNConv(const ConvolutionParam& params, const NDArray &input);
bool SupportMKLDNNDeconv(const DeconvolutionParam& params, const NDArray &input);
bool SupportMKLDNNSoftmax(const SoftmaxParam& param);
}
} // namespace op

static int GetTypeSize(int dtype) {
int size = -1;
Expand Down Expand Up @@ -253,14 +254,23 @@ inline static mkldnn::memory::desc GetWeightDesc(const NDArray &arr,
if (num_groups == 1) {
return GetMemDesc(arr);
} else {
CHECK_EQ(arr.shape().ndim(), 4U);
mkldnn::memory::dims tz = mkldnn::memory::dims{ num_groups,
static_cast<int>(arr.shape()[0] / num_groups),
static_cast<int>(arr.shape()[1]),
static_cast<int>(arr.shape()[2]),
static_cast<int>(arr.shape()[3])};
return mkldnn::memory::desc{tz, get_mkldnn_type(arr.dtype()),
mkldnn::memory::format::any};
CHECK((arr.shape().ndim() == 3) || (arr.shape().ndim() == 4));
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
if (arr.shape().ndim() == 3) {
mkldnn::memory::dims tz = mkldnn::memory::dims{ num_groups,
static_cast<int>(arr.shape()[0] / num_groups),
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
static_cast<int>(arr.shape()[1]),
static_cast<int>(arr.shape()[2])};
return mkldnn::memory::desc{tz, get_mkldnn_type(arr.dtype()),
mkldnn::memory::format::any};
} else {
mkldnn::memory::dims tz = mkldnn::memory::dims{ num_groups,
static_cast<int>(arr.shape()[0] / num_groups),
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
static_cast<int>(arr.shape()[1]),
static_cast<int>(arr.shape()[2]),
static_cast<int>(arr.shape()[3])};
return mkldnn::memory::desc{tz, get_mkldnn_type(arr.dtype()),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This line is the common part of dim3 and dim4, right?

mkldnn::memory::format::any};
}
}
}

Expand Down
84 changes: 63 additions & 21 deletions src/operator/nn/mkldnn/mkldnn_base.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -238,39 +238,44 @@ const mkldnn::memory *GetWeights(const NDArray &arr,
return mem;

mkldnn::memory::data_type type = get_mkldnn_type(arr.dtype());
mkldnn::memory::dims tz = mkldnn::memory::dims{0};
mkldnn::memory::format format = mkldnn::memory::format::format_undef;
auto engine = CpuEngine::Get()->get_engine();
if (arr.shape().ndim() == 2) {
mkldnn::memory::dims tz = mkldnn::memory::dims{
tz = mkldnn::memory::dims{
static_cast<int>(arr.shape()[0]), static_cast<int>(arr.shape()[1])};
mkldnn::memory::desc md =
mkldnn::memory::desc{tz, type, mkldnn::memory::format::oi};
mkldnn::memory::primitive_desc pd =
mkldnn::memory::primitive_desc{md, engine};
mem = arr.GetMKLDNNData(pd);
} else if (arr.shape().ndim() == 4 && num_groups == 1) {
mkldnn::memory::dims tz = mkldnn::memory::dims{
static_cast<int>(arr.shape()[0]), static_cast<int>(arr.shape()[1]),
static_cast<int>(arr.shape()[2]), static_cast<int>(arr.shape()[3])};
mkldnn::memory::desc md =
mkldnn::memory::desc{tz, type, mkldnn::memory::format::oihw};
mkldnn::memory::primitive_desc pd =
mkldnn::memory::primitive_desc{md, engine};
mem = arr.GetMKLDNNData(pd);
format = mkldnn::memory::format::oi;
} else if (arr.shape().ndim() == 3) {
tz = num_groups > 1 ? mkldnn::memory::dims{ num_groups,
static_cast<int>(arr.shape()[0] / num_groups),
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
static_cast<int>(arr.shape()[1]),
static_cast<int>(arr.shape()[2])} :
mkldnn::memory::dims{
static_cast<int>(arr.shape()[0]),
static_cast<int>(arr.shape()[1]),
static_cast<int>(arr.shape()[2])};
format = num_groups > 1 ? mkldnn::memory::format::goiw : mkldnn::memory::format::oiw;
} else if (arr.shape().ndim() == 4) {
mkldnn::memory::dims tz = mkldnn::memory::dims{ num_groups,
tz = num_groups > 1 ? mkldnn::memory::dims{ num_groups,
static_cast<int>(arr.shape()[0] / num_groups),
static_cast<int>(arr.shape()[1]),
static_cast<int>(arr.shape()[2]),
static_cast<int>(arr.shape()[3])} :
mkldnn::memory::dims{
static_cast<int>(arr.shape()[0]),
static_cast<int>(arr.shape()[1]),
static_cast<int>(arr.shape()[2]),
static_cast<int>(arr.shape()[3])};
mkldnn::memory::desc md =
mkldnn::memory::desc{tz, type, mkldnn::memory::format::goihw};
mkldnn::memory::primitive_desc pd =
mkldnn::memory::primitive_desc{md, engine};
mem = arr.GetMKLDNNData(pd);
format = num_groups > 1 ? mkldnn::memory::format::goihw : mkldnn::memory::format::oihw;
} else {
LOG(FATAL) << "The weight array has an unsupported number of dimensions";
return nullptr;
}
mkldnn::memory::desc md =
mkldnn::memory::desc{tz, type, format};
mkldnn::memory::primitive_desc pd =
mkldnn::memory::primitive_desc{md, engine};
mem = arr.GetMKLDNNData(pd);
if (mem == nullptr)
mem = arr.GetMKLDNNDataReorder(target_pd);
if (mem->get_primitive_desc() == target_pd) return mem;
Expand All @@ -284,6 +289,7 @@ mkldnn_memory_format_t GetDefaultFormat(int num_dims) {
switch (num_dims) {
case 1: return mkldnn_x;
case 2: return mkldnn_nc;
case 3: return mkldnn_ncw;
case 4: return mkldnn_nchw;
case 5: return mkldnn_goihw;
default:
Expand All @@ -300,6 +306,30 @@ mkldnn_memory_format_t GetDefaultFormat(const mkldnn::memory::desc &desc) {
return mkldnn_oi;
else
return desc.data.format;
} else if (desc.data.ndims == 3) {
switch (desc.data.format) {
case mkldnn_ncw:
case mkldnn_nwc:
case mkldnn_nCw8c:
case mkldnn_nCw16c:
return mkldnn_ncw;
case mkldnn_oiw:
case mkldnn_wio:
case mkldnn_Owi8o:
case mkldnn_OIw8i8o:
case mkldnn_OIw8o8i:
case mkldnn_OIw16i16o:
case mkldnn_OIw16o16i:
case mkldnn_Oiw16o:
case mkldnn_Owi16o:
case mkldnn_OIw8i16o2i:
case mkldnn_OIw8o16i2o:
case mkldnn_IOw16o16i:
return mkldnn_oiw;
default:
LOG(FATAL) << "Unknown MKLDNN format for 3 dimensions: " << desc.data.format;
return mkldnn_format_undef;
}
} else if (desc.data.ndims == 4) {
switch (desc.data.format) {
case mkldnn_nchw:
Expand Down Expand Up @@ -328,6 +358,18 @@ mkldnn_memory_format_t GetDefaultFormat(const mkldnn::memory::desc &desc) {
case mkldnn_Ohwi16o:
case mkldnn_OhIw16o4i:
return mkldnn_oihw;
case mkldnn_goiw:
case mkldnn_gOwi8o:
case mkldnn_gOIw8o8i:
case mkldnn_gOIw8i8o:
case mkldnn_gOIw16i16o:
case mkldnn_gOIw16o16i:
case mkldnn_gOiw16o:
case mkldnn_gOwi16o:
case mkldnn_gOIw8i16o2i:
case mkldnn_gOIw8o16i2o:
case mkldnn_gIOw16o16i:
return mkldnn_goiw;
default:
LOG(FATAL) << "Unknown MKLDNN format for 4 dimensions: " << desc.data.format;
return mkldnn_format_undef;
Expand Down
115 changes: 77 additions & 38 deletions src/operator/nn/mkldnn/mkldnn_convolution.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,9 +37,12 @@ namespace op {
DMLC_REGISTER_PARAMETER(MKLDNNConvParam);

bool SupportMKLDNNConv(const ConvolutionParam& params, const NDArray &input) {
if (params.kernel.ndim() != 2)
if ((params.kernel.ndim() != 1) &&
(params.kernel.ndim() != 2))
return false;
return SupportMKLDNNQuantize(input.dtype()) && input.shape().ndim() == 4;
return SupportMKLDNNQuantize(input.dtype()) &&
((input.shape().ndim() == 3) ||
(input.shape().ndim() == 4));
}

mkldnn::convolution_forward::primitive_desc GetConvFwdImpl(
Expand All @@ -51,15 +54,23 @@ mkldnn::convolution_forward::primitive_desc GetConvFwdImpl(
auto weight_md = GetWeightDesc(weights, param.conv_param.num_group);
auto out_md = GetMemDesc(output);
auto engine = CpuEngine::Get()->get_engine();
CHECK_GE(param.conv_param.stride.ndim(), 2U);
CHECK_GE(param.conv_param.pad.ndim(), 2U);
CHECK_GE(param.conv_param.dilate.ndim(), 2U);
mkldnn::memory::dims strides{0, 0};
strides[0] = param.conv_param.stride[0];
strides[1] = param.conv_param.stride[1];
mkldnn::memory::dims padding{0, 0};
padding[0] = param.conv_param.pad[0];
padding[1] = param.conv_param.pad[1];
mkldnn::memory::dims strides(param.conv_param.kernel.ndim());
mkldnn::memory::dims padding(param.conv_param.kernel.ndim());
if (param.conv_param.kernel.ndim() == 1) {
CHECK_GE(param.conv_param.stride.ndim(), 1U);
CHECK_GE(param.conv_param.pad.ndim(), 1U);
CHECK_GE(param.conv_param.dilate.ndim(), 1U);
strides[0] = param.conv_param.stride[0];
padding[0] = param.conv_param.pad[0];
} else if (param.conv_param.kernel.ndim() == 2) {
CHECK_GE(param.conv_param.stride.ndim(), 2U);
CHECK_GE(param.conv_param.pad.ndim(), 2U);
CHECK_GE(param.conv_param.dilate.ndim(), 2U);
strides[0] = param.conv_param.stride[0];
strides[1] = param.conv_param.stride[1];
padding[0] = param.conv_param.pad[0];
padding[1] = param.conv_param.pad[1];
}
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
mkldnn::primitive_attr attr;
mkldnn::post_ops ops;
if (param.mkldnn_param.with_relu) {
Expand Down Expand Up @@ -113,9 +124,13 @@ mkldnn::convolution_forward::primitive_desc GetConvFwdImpl(
}
return conv_pd;
} else {
mkldnn::memory::dims dilates{0, 0};
dilates[0] = param.conv_param.dilate[0] - 1;
dilates[1] = param.conv_param.dilate[1] - 1;
mkldnn::memory::dims dilates(param.conv_param.kernel.ndim());
if (param.conv_param.dilate.ndim() == 1) {
dilates[0] = param.conv_param.dilate[0] - 1;
} else if (param.conv_param.dilate.ndim() == 2) {
dilates[0] = param.conv_param.dilate[0] - 1;
dilates[1] = param.conv_param.dilate[1] - 1;
}
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
if (bias == nullptr) {
mkldnn::convolution_forward::desc desc(prop, mkldnn::algorithm::convolution_direct,
data_md, weight_md, out_md, strides, dilates, padding, padding,
Expand Down Expand Up @@ -151,15 +166,23 @@ static mkldnn::convolution_backward_data::primitive_desc GetConvBwdData(
auto weight_md = GetWeightDesc(weights, param.num_group);
auto out_md = GetMemDesc(output);
auto engine = CpuEngine::Get()->get_engine();
CHECK_GE(param.stride.ndim(), 2U);
CHECK_GE(param.pad.ndim(), 2U);
CHECK_GE(param.dilate.ndim(), 2U);
mkldnn::memory::dims strides{0, 0};
strides[0] = param.stride[0];
strides[1] = param.stride[1];
mkldnn::memory::dims padding{0, 0};
padding[0] = param.pad[0];
padding[1] = param.pad[1];
mkldnn::memory::dims strides(param.kernel.ndim());
mkldnn::memory::dims padding(param.kernel.ndim());
if (param.kernel.ndim() == 1) {
CHECK_GE(param.stride.ndim(), 1U);
CHECK_GE(param.pad.ndim(), 1U);
CHECK_GE(param.dilate.ndim(), 1U);
strides[0] = param.stride[0];
padding[0] = param.pad[0];
} else if (param.kernel.ndim() == 2) {
CHECK_GE(param.stride.ndim(), 2U);
CHECK_GE(param.pad.ndim(), 2U);
CHECK_GE(param.dilate.ndim(), 2U);
strides[0] = param.stride[0];
strides[1] = param.stride[1];
padding[0] = param.pad[0];
padding[1] = param.pad[1];
}
TaoLv marked this conversation as resolved.
Show resolved Hide resolved

// MKL-DNN introduced padded formats since 0.15 which require more memory
// for computation compared with the actual tensor size. Currently, MKL-DNN
Expand All @@ -177,9 +200,13 @@ static mkldnn::convolution_backward_data::primitive_desc GetConvBwdData(
}
return conv_pd;
} else {
mkldnn::memory::dims dilates{0, 0};
dilates[0] = param.dilate[0] - 1;
dilates[1] = param.dilate[1] - 1;
mkldnn::memory::dims dilates(param.kernel.ndim());
if (param.dilate.ndim() == 1) {
dilates[0] = param.dilate[0] - 1;
} else if (param.dilate.ndim() == 2) {
dilates[0] = param.dilate[0] - 1;
dilates[1] = param.dilate[1] - 1;
}
mkldnn::convolution_backward_data::desc desc(mkldnn::algorithm::convolution_direct,
data_md, weight_md, out_md, strides, dilates, padding, padding,
mkldnn::padding_kind::zero);
Expand All @@ -201,15 +228,23 @@ static mkldnn::convolution_backward_weights::primitive_desc GetConvBwdWeights(
auto weight_md = GetWeightDesc(weights, param.num_group);
auto out_md = GetMemDesc(output);
auto engine = CpuEngine::Get()->get_engine();
CHECK_GE(param.stride.ndim(), 2U);
CHECK_GE(param.pad.ndim(), 2U);
CHECK_GE(param.dilate.ndim(), 2U);
mkldnn::memory::dims strides{0, 0};
strides[0] = param.stride[0];
strides[1] = param.stride[1];
mkldnn::memory::dims padding{0, 0};
padding[0] = param.pad[0];
padding[1] = param.pad[1];
mkldnn::memory::dims strides(param.kernel.ndim());
mkldnn::memory::dims padding(param.kernel.ndim());
if (param.kernel.ndim() == 1) {
CHECK_GE(param.stride.ndim(), 1U);
CHECK_GE(param.pad.ndim(), 1U);
CHECK_GE(param.dilate.ndim(), 1U);
strides[0] = param.stride[0];
padding[0] = param.pad[0];
} else if (param.kernel.ndim() == 2) {
CHECK_GE(param.stride.ndim(), 2U);
CHECK_GE(param.pad.ndim(), 2U);
CHECK_GE(param.dilate.ndim(), 2U);
strides[0] = param.stride[0];
strides[1] = param.stride[1];
padding[0] = param.pad[0];
padding[1] = param.pad[1];
}

// MKL-DNN introduced padded formats since 0.15 which require more memory
// for computation compared with the actual tensor size. Currently, MKL-DNN
Expand Down Expand Up @@ -239,9 +274,13 @@ static mkldnn::convolution_backward_weights::primitive_desc GetConvBwdWeights(
}
return conv_pd;
} else {
mkldnn::memory::dims dilates{0, 0};
dilates[0] = param.dilate[0] - 1;
dilates[1] = param.dilate[1] - 1;
mkldnn::memory::dims dilates(param.kernel.ndim());
if (param.dilate.ndim() == 1) {
dilates[0] = param.dilate[0] - 1;
} else if (param.dilate.ndim() == 2) {
dilates[0] = param.dilate[0] - 1;
dilates[1] = param.dilate[1] - 1;
}
TaoLv marked this conversation as resolved.
Show resolved Hide resolved
if (bias == nullptr) {
mkldnn::convolution_backward_weights::desc desc(mkldnn::algorithm::convolution_direct,
data_md, weight_md, out_md, strides, dilates, padding, padding,
Expand Down