Skip to content

Commit

Permalink
[refactor] [ci] Enable identifier naming in clang-tidy (#3960)
Browse files Browse the repository at this point in the history 
* Tidy identifier naming

* Clang tidy auto fix

* Auto Format

Co-authored-by: Taichi Gardener <taichigardener@gmail.com>
  • Loading branch information
@qiao-bo @taichi-gardener
qiao-bo and taichi-gardener authored Jan 10, 2022
1 parent 8cfd80c commit a93d985
Show file tree
Hide file tree
Showing 6 changed files with 180 additions and 178 deletions.
2 changes: 1 addition & 1 deletion .github/workflows/scripts/check_clang_tidy.sh
View file
Original file line number Diff line number Diff line change
Expand Up @@ -9,4 +9,4 @@ rm -rf build && mkdir build && cd build
cmake $CI_SETUP_CMAKE_ARGS ..

cd ..
python3 ./scripts/run_clang_tidy.py $PWD/taichi -clang-tidy-binary clang-tidy-10 -checks=-*,performance-inefficient-string-concatenation -header-filter=$PWD/taichi -p $PWD/build -j2
python3 ./scripts/run_clang_tidy.py $PWD/taichi -clang-tidy-binary clang-tidy-10 -checks=-*,performance-inefficient-string-concatenation,readability-identifier-naming -header-filter=$PWD/taichi -p $PWD/build -j2
4 changes: 2 additions & 2 deletions taichi/backends/opengl/codegen_opengl.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -119,7 +119,7 @@ class KernelGen : public IRVisitor {
bool is_top_level_{true};
CompiledTaichiKernel compiled_program_;
UsedFeature used; // TODO: is this actually per-offload?
int arr_bind_idx = static_cast<int>(GLBufId::Arr);
int arr_bind_idx_ = static_cast<int>(GLBufId::Arr);

// per-offload variables:
LineAppender line_appender_;
Expand Down Expand Up @@ -850,7 +850,7 @@ class KernelGen : public IRVisitor {
const auto dt = opengl_data_type_name(stmt->element_type());
if (stmt->is_ptr) {
if (!used.arr_arg_to_bind_idx.count(stmt->arg_id)) {
used.arr_arg_to_bind_idx[stmt->arg_id] = arr_bind_idx++;
used.arr_arg_to_bind_idx[stmt->arg_id] = arr_bind_idx_++;
}
} else {
used.buf_args = true;
Expand Down
96 changes: 48 additions & 48 deletions taichi/codegen/spirv/spirv_ir_builder.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -178,25 +178,25 @@ PhiValue IRBuilder::make_phi(const SType &out_type, uint32_t num_incoming) {
Value IRBuilder::int_immediate_number(const SType &dtype,
int64_t value,
bool cache) {
return get_const_(dtype, reinterpret_cast<uint64_t *>(&value), cache);
return get_const(dtype, reinterpret_cast<uint64_t *>(&value), cache);
}

Value IRBuilder::uint_immediate_number(const SType &dtype,
uint64_t value,
bool cache) {
return get_const_(dtype, &value, cache);
return get_const(dtype, &value, cache);
}

Value IRBuilder::float_immediate_number(const SType &dtype,
double value,
bool cache) {
if (data_type_bits(dtype.dt) == 64) {
return get_const_(dtype, reinterpret_cast<uint64_t *>(&value), cache);
return get_const(dtype, reinterpret_cast<uint64_t *>(&value), cache);
} else if (data_type_bits(dtype.dt) == 32) {
float fvalue = static_cast<float>(value);
uint32_t *ptr = reinterpret_cast<uint32_t *>(&fvalue);
uint64_t data = ptr[0];
return get_const_(dtype, &data, cache);
return get_const(dtype, &data, cache);
} else {
TI_ERROR("Type {} not supported.", dtype.dt->to_string());
}
Expand Down Expand Up @@ -406,46 +406,46 @@ void IRBuilder::set_work_group_size(const std::array<int, 3> group_size) {
Value size_z =
uint_immediate_number(t_uint32_, static_cast<uint64_t>(group_size[2]));

if (gl_work_group_size.id == 0) {
gl_work_group_size.id = id_counter_++;
if (gl_work_group_size_.id == 0) {
gl_work_group_size_.id = id_counter_++;
}
ib_.begin(spv::OpConstantComposite)
.add_seq(t_v3_uint_, gl_work_group_size, size_x, size_y, size_z)
.add_seq(t_v3_uint_, gl_work_group_size_, size_x, size_y, size_z)
.commit(&global_);
this->decorate(spv::OpDecorate, gl_work_group_size, spv::DecorationBuiltIn,
this->decorate(spv::OpDecorate, gl_work_group_size_, spv::DecorationBuiltIn,
spv::BuiltInWorkgroupSize);
}

Value IRBuilder::get_num_work_groups(uint32_t dim_index) {
if (gl_num_work_groups.id == 0) {
if (gl_num_work_groups_.id == 0) {
SType ptr_type = this->get_pointer_type(t_v3_uint_, spv::StorageClassInput);
gl_num_work_groups = new_value(ptr_type, ValueKind::kVectorPtr);
gl_num_work_groups_ = new_value(ptr_type, ValueKind::kVectorPtr);
ib_.begin(spv::OpVariable)
.add_seq(ptr_type, gl_num_work_groups, spv::StorageClassInput)
.add_seq(ptr_type, gl_num_work_groups_, spv::StorageClassInput)
.commit(&global_);
this->decorate(spv::OpDecorate, gl_num_work_groups, spv::DecorationBuiltIn,
this->decorate(spv::OpDecorate, gl_num_work_groups_, spv::DecorationBuiltIn,
spv::BuiltInNumWorkgroups);
}
SType pint_type = this->get_pointer_type(t_uint32_, spv::StorageClassInput);
Value ptr = this->make_value(
spv::OpAccessChain, pint_type, gl_num_work_groups,
spv::OpAccessChain, pint_type, gl_num_work_groups_,
uint_immediate_number(t_uint32_, static_cast<uint64_t>(dim_index)));

return this->make_value(spv::OpLoad, t_uint32_, ptr);
}
Value IRBuilder::get_global_invocation_id(uint32_t dim_index) {
if (gl_global_invocation_id.id == 0) {
if (gl_global_invocation_id_.id == 0) {
SType ptr_type = this->get_pointer_type(t_v3_uint_, spv::StorageClassInput);
gl_global_invocation_id = new_value(ptr_type, ValueKind::kVectorPtr);
gl_global_invocation_id_ = new_value(ptr_type, ValueKind::kVectorPtr);
ib_.begin(spv::OpVariable)
.add_seq(ptr_type, gl_global_invocation_id, spv::StorageClassInput)
.add_seq(ptr_type, gl_global_invocation_id_, spv::StorageClassInput)
.commit(&global_);
this->decorate(spv::OpDecorate, gl_global_invocation_id,
this->decorate(spv::OpDecorate, gl_global_invocation_id_,
spv::DecorationBuiltIn, spv::BuiltInGlobalInvocationId);
}
SType pint_type = this->get_pointer_type(t_uint32_, spv::StorageClassInput);
Value ptr = this->make_value(
spv::OpAccessChain, pint_type, gl_global_invocation_id,
spv::OpAccessChain, pint_type, gl_global_invocation_id_,
uint_immediate_number(t_uint32_, static_cast<uint64_t>(dim_index)));

return this->make_value(spv::OpLoad, t_uint32_, ptr);
Expand Down Expand Up @@ -753,19 +753,19 @@ Value IRBuilder::rand_u32(Value global_tmp_) {
Value _19u = uint_immediate_number(t_uint32_, 19u);
Value _8u = uint_immediate_number(t_uint32_, 8u);
Value _1000000007u = uint_immediate_number(t_uint32_, 1000000007u);
Value tmp0 = load_variable(_rand_x_, t_uint32_);
Value tmp0 = load_variable(rand_x_, t_uint32_);
Value tmp1 = make_value(spv::OpShiftLeftLogical, t_uint32_, tmp0, _11u);
Value tmp_t = make_value(spv::OpBitwiseXor, t_uint32_, tmp0, tmp1); // t
store_variable(_rand_x_, load_variable(_rand_y_, t_uint32_));
store_variable(_rand_y_, load_variable(_rand_z_, t_uint32_));
Value tmp_w = load_variable(_rand_w_, t_uint32_); // reuse w
store_variable(_rand_z_, tmp_w);
store_variable(rand_x_, load_variable(rand_y_, t_uint32_));
store_variable(rand_y_, load_variable(rand_z_, t_uint32_));
Value tmp_w = load_variable(rand_w_, t_uint32_); // reuse w
store_variable(rand_z_, tmp_w);
Value tmp2 = make_value(spv::OpShiftRightLogical, t_uint32_, tmp_w, _19u);
Value tmp3 = make_value(spv::OpBitwiseXor, t_uint32_, tmp_w, tmp2);
Value tmp4 = make_value(spv::OpShiftRightLogical, t_uint32_, tmp_t, _8u);
Value tmp5 = make_value(spv::OpBitwiseXor, t_uint32_, tmp_t, tmp4);
Value new_w = make_value(spv::OpBitwiseXor, t_uint32_, tmp3, tmp5);
store_variable(_rand_w_, new_w);
store_variable(rand_w_, new_w);
Value val = make_value(spv::OpIMul, t_uint32_, new_w, _1000000007u);

return val;
Expand Down Expand Up @@ -794,9 +794,9 @@ Value IRBuilder::rand_i32(Value global_tmp_) {
return val;
}

Value IRBuilder::get_const_(const SType &dtype,
const uint64_t *pvalue,
bool cache) {
Value IRBuilder::get_const(const SType &dtype,
const uint64_t *pvalue,
bool cache) {
auto key = std::make_pair(dtype.id, pvalue[0]);
if (cache) {
auto it = const_tbl_.find(key);
Expand Down Expand Up @@ -862,30 +862,30 @@ SType IRBuilder::declare_primitive_type(DataType dt) {
void IRBuilder::init_random_function(Value global_tmp_) {
// variables declare
SType local_type = get_pointer_type(t_uint32_, spv::StorageClassPrivate);
_rand_x_ = new_value(local_type, ValueKind::kVariablePtr);
_rand_y_ = new_value(local_type, ValueKind::kVariablePtr);
_rand_z_ = new_value(local_type, ValueKind::kVariablePtr);
_rand_w_ = new_value(local_type, ValueKind::kVariablePtr);
global_values.push_back(_rand_x_);
global_values.push_back(_rand_y_);
global_values.push_back(_rand_z_);
global_values.push_back(_rand_w_);
rand_x_ = new_value(local_type, ValueKind::kVariablePtr);
rand_y_ = new_value(local_type, ValueKind::kVariablePtr);
rand_z_ = new_value(local_type, ValueKind::kVariablePtr);
rand_w_ = new_value(local_type, ValueKind::kVariablePtr);
global_values.push_back(rand_x_);
global_values.push_back(rand_y_);
global_values.push_back(rand_z_);
global_values.push_back(rand_w_);
ib_.begin(spv::OpVariable)
.add_seq(local_type, _rand_x_, spv::StorageClassPrivate)
.add_seq(local_type, rand_x_, spv::StorageClassPrivate)
.commit(&global_);
ib_.begin(spv::OpVariable)
.add_seq(local_type, _rand_y_, spv::StorageClassPrivate)
.add_seq(local_type, rand_y_, spv::StorageClassPrivate)
.commit(&global_);
ib_.begin(spv::OpVariable)
.add_seq(local_type, _rand_z_, spv::StorageClassPrivate)
.add_seq(local_type, rand_z_, spv::StorageClassPrivate)
.commit(&global_);
ib_.begin(spv::OpVariable)
.add_seq(local_type, _rand_w_, spv::StorageClassPrivate)
.add_seq(local_type, rand_w_, spv::StorageClassPrivate)
.commit(&global_);
debug(spv::OpName, _rand_x_, "_rand_x");
debug(spv::OpName, _rand_y_, "_rand_y");
debug(spv::OpName, _rand_z_, "_rand_z");
debug(spv::OpName, _rand_w_, "_rand_w");
debug(spv::OpName, rand_x_, "_rand_x");
debug(spv::OpName, rand_y_, "_rand_y");
debug(spv::OpName, rand_z_, "_rand_z");
debug(spv::OpName, rand_w_, "_rand_w");
SType gtmp_type = get_pointer_type(t_int32_, spv::StorageClassStorageBuffer);
Value rand_gtmp_ = new_value(gtmp_type, ValueKind::kVariablePtr);
debug(spv::OpName, rand_gtmp_, "rand_gtmp");
Expand Down Expand Up @@ -928,7 +928,7 @@ void IRBuilder::init_random_function(Value global_tmp_) {
SType pint_type = this->get_pointer_type(t_uint32_, spv::StorageClassInput);
Value tmp0 = new_value(pint_type, ValueKind::kVariablePtr);
ib_.begin(spv::OpAccessChain)
.add_seq(pint_type, tmp0, gl_global_invocation_id,
.add_seq(pint_type, tmp0, gl_global_invocation_id_,
uint_immediate_number(t_uint32_, 0))
.commit(&func_header_);
Value tmp1 = load_var(tmp0, t_uint32_);
Expand Down Expand Up @@ -961,10 +961,10 @@ void IRBuilder::init_random_function(Value global_tmp_) {
ib_.begin(spv::OpIMul)
.add_seq(t_uint32_, tmp8, _1000000007u, tmp7)
.commit(&func_header_);
store_var(_rand_x_, tmp8);
store_var(_rand_y_, _362436069u);
store_var(_rand_z_, _521288629u);
store_var(_rand_w_, _88675123u);
store_var(rand_x_, tmp8);
store_var(rand_y_, _362436069u);
store_var(rand_z_, _521288629u);
store_var(rand_w_, _88675123u);
// Yes, this is not an atomic operation, but just fine since no matter
// how RAND_STATE changes, `gl_GlobalInvocationID.x` can still help
// us to set different seeds for different threads.
Expand Down
24 changes: 12 additions & 12 deletions taichi/codegen/spirv/spirv_ir_builder.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -346,11 +346,11 @@ class IRBuilder {
ib_.add(v);
}
}
if (gl_global_invocation_id.id != 0) {
ib_.add(gl_global_invocation_id);
if (gl_global_invocation_id_.id != 0) {
ib_.add(gl_global_invocation_id_);
}
if (gl_num_work_groups.id != 0) {
ib_.add(gl_num_work_groups);
if (gl_num_work_groups_.id != 0) {
ib_.add(gl_num_work_groups_);
}
ib_.commit(&entry_);
ib_.begin(spv::OpExecutionMode)
Expand Down Expand Up @@ -476,7 +476,7 @@ class IRBuilder {
return val;
}

Value get_const_(const SType &dtype, const uint64_t *pvalue, bool cache);
Value get_const(const SType &dtype, const uint64_t *pvalue, bool cache);
SType declare_primitive_type(DataType dt);

void init_random_function(Value global_tmp_);
Expand Down Expand Up @@ -509,16 +509,16 @@ class IRBuilder {
SType t_void_func_;
// gl compute shader related type(s) and variables
SType t_v3_uint_;
Value gl_global_invocation_id;
Value gl_num_work_groups;
Value gl_work_group_size;
Value gl_global_invocation_id_;
Value gl_num_work_groups_;
Value gl_work_group_size_;

// Random function and variables
bool init_rand_{false};
Value _rand_x_;
Value _rand_y_;
Value _rand_z_;
Value _rand_w_; // per-thread local variable
Value rand_x_;
Value rand_y_;
Value rand_z_;
Value rand_w_; // per-thread local variable

// map from value to its pointer type
std::map<std::pair<uint32_t, spv::StorageClass>, SType> pointer_type_tbl_;
Expand Down
Loading

0 comments on commit a93d985

Please sign in to comment.