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

Jump to bottom

Fix some typos (noedes, mulitple, etc.) #61788

Merged
merged 1 commit into from
Feb 22, 2024
Merged

Fix some typos (noedes, mulitple, etc.) #61788

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 1 deletion paddle/cinn/optim/ir_simplify.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,7 +43,7 @@ using utils::Replace;
namespace {

//! Simplify some sub-expression in the `expr`. Due to the simplify strategy
//! just fit several kinds of IR noedes, we partition the original expression to
//! just fit several kinds of IR nodes, we partition the original expression to
//! several sub-expression those supported by simplify, and process each of
//! them.
void PartialSimplify(
Expand Down
4 changes: 2 additions & 2 deletions paddle/cinn/optim/lower_intrin.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -69,10 +69,10 @@ void LowerIntrin(Expr *e, Target target) {
void Visit(const ir::Call *op, Expr *expr) override {
auto *node = expr->As<ir::Call>();
CHECK(node);
LowerCpuintrinsicOp(node, expr);
LowerCpuIntrinsicOp(node, expr);
}

void LowerCpuintrinsicOp(ir::Call *op, Expr *expr) {
void LowerCpuIntrinsicOp(ir::Call *op, Expr *expr) {
auto *node = expr->As<ir::Call>();
if (kIntrinsicCalls.count(node->name)) {
CHECK(!node->name.empty());
Expand Down
10 changes: 5 additions & 5 deletions paddle/cinn/optim/map_extern_call.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,17 +57,17 @@ void MapExternCall(Expr *e, Target target) {
CHECK(node);
OptimizeConstantPow(node);
if (target.arch == Target::Arch::NVGPU) {
DealWithNvGpuintrinsics(node, expr);
DealWithNvGpuIntrinsics(node, expr);
} else {
DealWithCpuintrinsics(node, expr);
DealWithCpuIntrinsics(node, expr);
}
}

void DealWithCpuintrinsics(ir::Call *node, Expr *expr) {
void DealWithCpuIntrinsics(ir::Call *node, Expr *expr) {
if (kExternFp32CallsCPU.count(node->name)) {
CHECK_GE(node->read_args.size(), 1UL);
CHECK(node->read_args.front().type().is_float())
<< "CPU extern call instrinsices only support float now! Please "
<< "CPU extern call intrinsics only support float now! Please "
"check.";
if (node->read_args.front().type().is_float(32)) {
auto out_type = node->type();
Expand All @@ -76,7 +76,7 @@ void MapExternCall(Expr *e, Target target) {
}
}

void DealWithNvGpuintrinsics(ir::Call *node, Expr *expr) {
void DealWithNvGpuIntrinsics(ir::Call *node, Expr *expr) {
auto arg_size = node->read_args.size();
if (arg_size == 0UL) {
// some node like __syncthreads hasn't arguments
Expand Down
14 changes: 7 additions & 7 deletions paddle/cinn/optim/update_buffer_axis_pass.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,20 +30,20 @@ namespace optim {

bool ExprMathEqual(const Expr& expr1, const Expr& expr2) {
ir::Expr cmp_expr = common::AutoSimplify(ir::Sub::Make(expr1, expr2));
// This is ugry code since AutoSimplify is not powerful enough. Modify it
// This is ugly code since AutoSimplify is not powerful enough. Modify it
// after we make auto simplify better
ir::Expr simplied = common::AutoSimplify(cmp_expr);
ir::Expr simplified = common::AutoSimplify(cmp_expr);
int count = 0;
while (simplied != cmp_expr) {
cmp_expr = simplied;
simplied = common::AutoSimplify(cmp_expr);
while (simplified != cmp_expr) {
cmp_expr = simplified;
simplified = common::AutoSimplify(cmp_expr);
++count;
// Control dead loop
if (count >= 5) {
break;
}
}
return simplied.is_constant() && simplied.get_constant() == 0;
return simplified.is_constant() && simplified.get_constant() == 0;
}

void FormalizeSingleIndex(const ir::Tensor& tensor,
Expand Down Expand Up @@ -126,7 +126,7 @@ class AnalyzeBufferAxis : public ir::IRMutator<> {
if (!buffer_name_access_same_index_expr.count(buffer_name)) {
for (int i = 0; i < indices.size(); ++i) {
if (tensor->buffer->memory_type == ir::MemoryType::GPUShared) {
// In GPUShared case, the thread vars cannot be simplied
// In GPUShared case, the thread vars cannot be simplified
std::set<ir::Expr> var_nodes =
ir::ir_utils::CollectIRNodesWithoutTensor(
indices[i], [&](const Expr* x) {
Expand Down
18 changes: 9 additions & 9 deletions paddle/cinn/optim/vectorize_loops.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -137,7 +137,7 @@ class TensorVectorizeTeller : public ir::IRMutator<const Expr *> {
return false;
}

// the iter val can't appear in mulitple indices
// the iter val can't appear in multiple indices
for (int i = 0; i < indices.size() - 1; ++i) {
auto repeat_found =
ir::ir_utils::CollectIRNodes(indices[i], find_matched_var_fn);
Expand Down Expand Up @@ -184,9 +184,9 @@ class TensorVectorizeTeller : public ir::IRMutator<const Expr *> {
};

// find tensors accessed sequentially in a for-loop to be vectorized,
// and substitue the corresponding cuda built-in vector for them
// and substitute the corresponding cuda built-in vector for them
class CudaVectorizer : public IRMutator<Expr *> {
const Var iter_var_; // the loop var of the vecotrized loop
const Var iter_var_; // the loop var of the vectorized loop
const int factor_; // the factor for vectorize

std::set<std::string> write_teller_;
Expand Down Expand Up @@ -258,7 +258,7 @@ class CudaVectorizer : public IRMutator<Expr *> {
}

auto vectorized_var = tensor2vectorized_vars_.at(tensor->name);
// substitue a new tensor with the vector name and dtype
// substitute a new tensor with the vector name and dtype
auto t = vectorized_var->type().is_cpp_handle()
? node->tensor->type().PointerOf()
: node->tensor->type();
Expand Down Expand Up @@ -286,7 +286,7 @@ class CudaVectorizer : public IRMutator<Expr *> {
GET_CUDA_VECTOR_TYPE_NAME(type.is_bfloat16(), "bfloat16");
#undef GET_CUDA_VECTOR_TYPE_NAME

// others are not implementd yet
// others are not implemented yet
CINN_NOT_IMPLEMENTED
return "";
}
Expand Down Expand Up @@ -727,15 +727,15 @@ struct VectorizeLoops_ : public IRMutator<Expr *> {

void Visit(const For *forloop, Expr *expr) {
auto *node = expr->As<For>();
auto loopvar_name = forloop->loop_var->name;
auto loop_var_name = forloop->loop_var->name;
if (forloop->extent.As<IntImm>()) {
var_intervals.emplace(
loopvar_name,
loop_var_name,
cinn::common::CasInterval{static_cast<int64_t>(0),
forloop->extent.as_int64() - 1});
} else {
var_intervals.emplace(
loopvar_name,
loop_var_name,
cinn::common::CasInterval{Expr(0), forloop->extent - 1});
}
// the extent the forloops marked as Vectorized should be int constant
Expand Down Expand Up @@ -842,7 +842,7 @@ struct VectorizeLoops_ : public IRMutator<Expr *> {
} else {
IRMutator::Visit(forloop, expr);
}
var_intervals.erase(loopvar_name);
var_intervals.erase(loop_var_name);
}

//! unroll the forloop if its' extent is min type by solving the condition
Expand Down