Skip to content

optimize phi::DenseTensor in paddle/fluid/imperative/ [fluid_ops] #77237

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

Merged
luotao1 merged 2 commits into from
Jan 12, 2026
Merged

optimize phi::DenseTensor in paddle/fluid/imperative/ [fluid_ops] #77237

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
69ab027
optimize phi::DenseTensor in paddle/fluid/imperative/
co63oc Jan 7, 2026
e02909b
fix
co63oc Jan 7, 2026
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
18 changes: 8 additions & 10 deletions paddle/fluid/imperative/all_reduce.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,8 +37,8 @@
namespace paddle::imperative {

static const phi::Place &GetVarPlace(const framework::Variable &src) {
if (src.IsType<phi::DenseTensor>()) {
return src.Get<phi::DenseTensor>().place();
if (src.IsType<DenseTensor>()) {
return src.Get<DenseTensor>().place();
#if NCCL_VERSION_CODE >= 2212
} else if (src.IsType<phi::SelectedRows>()) {
return src.Get<phi::SelectedRows>().value().place();
Expand All @@ -52,8 +52,8 @@ static const phi::Place &GetVarPlace(const framework::Variable &src) {
}
}

static void AllReduce(const phi::DenseTensor &src,
phi::DenseTensor *dst,
static void AllReduce(const DenseTensor &src,
DenseTensor *dst,
const gpuStream_t stream,
const platform::NCCLComm *comm) {
const auto &place = src.place();
Expand Down Expand Up @@ -224,14 +224,12 @@ void AllReduce(const framework::Variable &src,
platform::NCCLCommContext::Instance().Get(ring_id, place);
gpuStream_t stream = (use_calc_stream ? dev_ctx->stream() : comm->stream());

if (src.IsType<phi::DenseTensor>()) {
if (!dst->IsType<phi::DenseTensor>()) {
if (src.IsType<DenseTensor>()) {
if (!dst->IsType<DenseTensor>()) {
dst->Clear();
}
AllReduce(src.Get<phi::DenseTensor>(),
dst->GetMutable<phi::DenseTensor>(),
stream,
comm);
AllReduce(
src.Get<DenseTensor>(), dst->GetMutable<DenseTensor>(), stream, comm);
#if NCCL_VERSION_CODE >= 2212
} else if (src.IsType<phi::SelectedRows>()) {
if (&src != dst) {
Expand Down
25 changes: 11 additions & 14 deletions paddle/fluid/imperative/basic_engine.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -100,9 +100,9 @@ void BasicEngine::Init(
true,
common::errors::NotFound("Tensor %s has no gradient", var->Name()));

auto& fwd_var = var->Var().Get<phi::DenseTensor>();
auto& fwd_var = var->Var().Get<DenseTensor>();
auto* grad_var =
var->GradVarBase()->MutableVar()->GetMutable<phi::DenseTensor>();
var->GradVarBase()->MutableVar()->GetMutable<DenseTensor>();
VLOG(6) << "init loss grad:" << var->GradVarBase()->Name()
<< " as stop_gradient false";
var->GradVarBase()->InnerSetOverriddenStopGradient(false);
Expand All @@ -112,7 +112,7 @@ void BasicEngine::Init(
grad_var->mutable_data(fwd_var.place(), fwd_var.type());
phi::funcs::set_constant(*dev_ctx, grad_var, 1.0f);
} else {
paddle::framework::TensorCopy(grad_tensor->Var().Get<phi::DenseTensor>(),
paddle::framework::TensorCopy(grad_tensor->Var().Get<DenseTensor>(),
fwd_var.place(),
*dev_ctx,
grad_var);
Expand Down Expand Up @@ -149,10 +149,9 @@ void BasicEngine::CheckBackwardInputs(const OpBase& op) {
}

auto* inner_var = var->MutableVar();
phi::DenseTensor* tensor = nullptr;
if (!inner_var->IsInitialized() ||
inner_var->IsType<phi::DenseTensor>()) {
tensor = inner_var->GetMutable<phi::DenseTensor>();
DenseTensor* tensor = nullptr;
if (!inner_var->IsInitialized() || inner_var->IsType<DenseTensor>()) {
tensor = inner_var->GetMutable<DenseTensor>();
}

if (tensor && !tensor->IsInitialized()) {
Expand Down Expand Up @@ -340,8 +339,8 @@ static std::shared_ptr<NameVarMap<VariableWrapper>> CallGradientHooks(

static bool IsInputCanInplace(const std::shared_ptr<VariableWrapper>& var) {
auto* inner_var = var->MutableVar();
if (inner_var->IsInitialized() && inner_var->IsType<phi::DenseTensor>()) {
auto tensor = inner_var->GetMutable<phi::DenseTensor>();
if (inner_var->IsInitialized() && inner_var->IsType<DenseTensor>()) {
auto tensor = inner_var->GetMutable<DenseTensor>();
if (tensor->IsInitialized()) {
return true;
}
Expand All @@ -358,7 +357,7 @@ static void PerformBackwardInplace(const std::string& op_type,
if (infer_inplace) {
auto in_to_outs = infer_inplace(true);
for (auto& pair : in_to_outs) {
phi::DenseTensor *in_tensor = nullptr, *out_tensor = nullptr;
DenseTensor *in_tensor = nullptr, *out_tensor = nullptr;
for (auto& p : ins) {
if (p.first == pair.first) {
// has at least one var
Expand All @@ -368,8 +367,7 @@ static void PerformBackwardInplace(const std::string& op_type,
// the refcount of var to be inplaced should be 1
if (in_var.use_count() == 1) {
if (IsInputCanInplace(in_var)) {
in_tensor =
in_var->MutableVar()->GetMutable<phi::DenseTensor>();
in_tensor = in_var->MutableVar()->GetMutable<DenseTensor>();
}
}
}
Expand All @@ -383,8 +381,7 @@ static void PerformBackwardInplace(const std::string& op_type,
if (!p.second.empty() && p.second[0]) {
auto& out_var = p.second[0];
if (out_var->Type() == framework::proto::VarType::DENSE_TENSOR) {
out_tensor =
out_var->MutableVar()->GetMutable<phi::DenseTensor>();
out_tensor = out_var->MutableVar()->GetMutable<DenseTensor>();
}
}
}
Expand Down
16 changes: 7 additions & 9 deletions paddle/fluid/imperative/bkcl_context.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -33,8 +33,8 @@
namespace paddle {
namespace imperative {

static void AllReduce(const phi::DenseTensor &src,
phi::DenseTensor *dst,
static void AllReduce(const DenseTensor &src,
DenseTensor *dst,
const XPUStream stream,
const platform::BKCLComm *comm) {
const auto &place = src.place();
Expand Down Expand Up @@ -162,14 +162,12 @@ void BKCLParallelContext::AllReduceByStream(const framework::Variable &src,
XPUStream stream =
use_calc_stream ? dev_ctx->x_context()->xpu_stream : comm->stream();

if (src.IsType<phi::DenseTensor>()) {
if (!dst->IsType<phi::DenseTensor>()) {
if (src.IsType<DenseTensor>()) {
if (!dst->IsType<DenseTensor>()) {
dst->Clear();
}
AllReduce(src.Get<phi::DenseTensor>(),
dst->GetMutable<phi::DenseTensor>(),
stream,
comm);
AllReduce(
src.Get<DenseTensor>(), dst->GetMutable<DenseTensor>(), stream, comm);
} else {
PADDLE_THROW(common::errors::InvalidArgument(
"XPU unsupported variable type %s for imperative allreduce, only "
Expand All @@ -180,7 +178,7 @@ void BKCLParallelContext::AllReduceByStream(const framework::Variable &src,

void BKCLParallelContext::Broadcast(framework::Variable *src, int ring_id) {
VLOG(3) << "/// DEBUG /// start inter broadcast with ring_id: " << ring_id;
phi::DenseTensor *src_tensor = src->GetMutable<phi::DenseTensor>();
DenseTensor *src_tensor = src->GetMutable<DenseTensor>();
const auto &place = src_tensor->place();
platform::BKCLComm *comm =
platform::BKCLCommContext::Instance().Get(ring_id, place);
Expand Down
17 changes: 8 additions & 9 deletions paddle/fluid/imperative/dygraph_grad_maker.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -186,8 +186,8 @@ class GradOpBaseMakerBase {

if (!is_input) {
auto* tensor =
grad_var_base_tmp->MutableVar()->GetMutable<phi::DenseTensor>();
tensor->Resize(var_base_temp->Var().Get<phi::DenseTensor>().dims());
grad_var_base_tmp->MutableVar()->GetMutable<DenseTensor>();
tensor->Resize(var_base_temp->Var().Get<DenseTensor>().dims());
}
vec_temp.emplace_back(grad_var_base_tmp);
} else {
Expand Down Expand Up @@ -363,14 +363,13 @@ class TracedGradOp {
} else if (var_wrapper->InplaceVersionSnapshot() ==
var_wrapper->MutableVar()->CurrentInplaceVersion()) {
return var_wrapper;
} else if (var_wrapper->MutableVar()->IsType<phi::DenseTensor>() ||
} else if (var_wrapper->MutableVar()->IsType<DenseTensor>() ||
var_wrapper->MutableVar()->IsType<phi::SelectedRows>()) {
auto* tensor =
var_wrapper->MutableVar()->IsType<phi::DenseTensor>()
? var_wrapper->MutableVar()->GetMutable<phi::DenseTensor>()
: var_wrapper->MutableVar()
->GetMutable<phi::SelectedRows>()
->mutable_value();
auto* tensor = var_wrapper->MutableVar()->IsType<DenseTensor>()
? var_wrapper->MutableVar()->GetMutable<DenseTensor>()
: var_wrapper->MutableVar()
->GetMutable<phi::SelectedRows>()
->mutable_value();
if (!tensor->IsInitialized()) {
return var_wrapper;
}
Expand Down
10 changes: 5 additions & 5 deletions paddle/fluid/imperative/gloo_context.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,11 +79,11 @@ void GLOOParallelContext::AllReduceByStream(const framework::Variable &src,
int ring_id,
bool use_calc_stream) {
// AllReduce(src, dst, strategy_, ring_id, use_calc_stream);
if (src.IsType<phi::DenseTensor>()) {
if (!dst->IsType<phi::DenseTensor>()) {
if (src.IsType<DenseTensor>()) {
if (!dst->IsType<DenseTensor>()) {
dst->Clear();
}
AllReduce(src.Get<phi::DenseTensor>(), dst->GetMutable<phi::DenseTensor>());
AllReduce(src.Get<DenseTensor>(), dst->GetMutable<DenseTensor>());
} else if (src.IsType<phi::SelectedRows>()) {
if (&src != dst) {
if (!dst->IsType<phi::SelectedRows>()) {
Expand All @@ -106,8 +106,8 @@ void GLOOParallelContext::AllReduceByStream(const framework::Variable &src,
}
}

void GLOOParallelContext::AllReduce(const phi::DenseTensor &src_tensor,
phi::DenseTensor *dst_tensor) {
void GLOOParallelContext::AllReduce(const DenseTensor &src_tensor,
DenseTensor *dst_tensor) {
auto gloo_wrapper = framework::GlooWrapper::GetInstance();
dst_tensor->Resize(src_tensor.dims());
switch (framework::TransToProtoVarType(src_tensor.dtype())) {
Expand Down
2 changes: 1 addition & 1 deletion paddle/fluid/imperative/gloo_context.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -60,7 +60,7 @@ class GLOOParallelContext : public ParallelContext {
void SynchronizeCompute() override;

private:
void AllReduce(const phi::DenseTensor& src, phi::DenseTensor* dst);
void AllReduce(const DenseTensor& src, DenseTensor* dst);
void AllReduce(const phi::SelectedRows& src, phi::SelectedRows* dst);

private:
Expand Down
Loading
Loading