Skip to content
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

Node hash new #3514

Merged
merged 4 commits into from
Apr 21, 2022
Merged

Node hash new #3514

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
1c48c7b
Move hash to the XlaNode
JackCaoG Apr 20, 2022
738ccba
pin
JackCaoG Apr 20, 2022
1144e42
reuse node_hash when computing dag_hash
JackCaoG Apr 20, 2022
e0ae574
Delete .torch_pin
JackCaoG Apr 21, 2022
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
22 changes: 8 additions & 14 deletions torch_xla/csrc/ir.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -114,15 +114,10 @@ const xla::Shape& Value::xla_node_shape() const {

Node::Node(torch::lazy::OpKind op, OpList operands, xla::Shape shape,
size_t num_outputs, torch::lazy::hash_t hash_seed)
: torch::lazy::Node(
op, num_outputs,
/*node_hash=*/torch::lazy::HashCombine(op.hash(), hash_seed),
/*dag_hash_fn=*/
[&](bool /*bakeInSizes*/) -> torch::lazy::hash_t {
return GetOperandHashes(
operands, torch::lazy::HashCombine(op.hash(), hash_seed));
}),
xla_shape_(std::move(shape)) {
: torch::lazy::Node(op, num_outputs),
xla_shape_(std::move(shape)),
node_hash_(torch::lazy::HashCombine(op.hash(), hash_seed)),
dag_hash_(GetOperandHashes(operands, node_hash_)) {
for (auto& operand : operands) {
AddOperand(operand.node, operand.index);
}
Expand All @@ -139,11 +134,10 @@ Node::Node(torch::lazy::OpKind op, OpList operands,

Node::Node(torch::lazy::OpKind op, xla::Shape shape, size_t num_outputs,
torch::lazy::hash_t hash_seed)
: torch::lazy::Node(op, num_outputs, /*node_hash_fn=*/
[&](bool /*bakeInSizes*/) -> torch::lazy::hash_t {
return GetOpHash(op, shape, hash_seed);
}),
xla_shape_(std::move(shape)) {}
: torch::lazy::Node(op, num_outputs),
xla_shape_(std::move(shape)),
node_hash_(GetOpHash(op, xla_shape_, hash_seed)),
dag_hash_(node_hash_) {}

Node::~Node() {
for (size_t i = 0; i < operands_as_outputs_.size(); ++i) {
Expand Down
8 changes: 8 additions & 0 deletions torch_xla/csrc/ir.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -125,6 +125,12 @@ class Node : public torch::lazy::Node {
XlaOpVector ReturnOps(absl::Span<const xla::XlaOp> ops,
LoweringContext* loctx) const;

torch::lazy::hash_t node_hash() const { return node_hash_; }

torch::lazy::hash_t hash() const override { return dag_hash_; }

torch::lazy::hash_t shapeHash() const override { return dag_hash_; }

private:
// Adds node's index output number as operand.
void AddOperand(torch::lazy::NodePtr node, size_t index = 0);
Expand All @@ -144,6 +150,8 @@ class Node : public torch::lazy::Node {
xla::Shape xla_shape_;
// We use a set for uses, as we want deterministic use sequencing.
std::set<Use> uses_;
torch::lazy::hash_t node_hash_;
torch::lazy::hash_t dag_hash_;
};

// RAII data structure to be used a stack variable to enter a new IR scope. IR
Expand Down
2 changes: 1 addition & 1 deletion torch_xla/csrc/op_by_op_executor.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@ torch::lazy::hash_t ComputeNodeKey(
}
const ir::Node* casted = dynamic_cast<const ir::Node*>(node);
key = torch::lazy::HashCombine(key, torch::lazy::Hash(casted->xla_shape()));
return torch::lazy::HashCombine(key, node->node_hash());
return torch::lazy::HashCombine(key, casted->node_hash());
}

xla::XlaComputation BuildNodeComputation(
Expand Down