Skip to content

Use compile-time promotion to reduce optimized le op size & build time #3534

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

Closed
wants to merge 10 commits into from
Closed

Use compile-time promotion to reduce optimized le op size & build time #3534

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
cef74f0
Use compile-time promotion to reduce max/min size & build time (#3459)
swolchok May 8, 2024
689ffaf
Use compile-time promotion to reduce floor_divide size & build time (…
swolchok May 8, 2024
ccb58e3
Use compile-time promotion to reduce remainder size & build time (#3458)
swolchok May 8, 2024
61485b5
Use compile-time promotion to reduce fmod size & build time (#3456)
swolchok May 8, 2024
1a13a29
support Half in minimum and clamp (#3457)
swolchok May 8, 2024
74f4fc8
use utils:{min,max}_override in {min,max}imum ops (#3453)
swolchok May 8, 2024
9be3d6d
Use compile-time promotion to reduce bitwise op size & build time (#3…
swolchok May 8, 2024
7bd0547
Use compile-time promotion to reduce optimized mul op size & build ti…
swolchok May 8, 2024
eaff700
Use compile-time promotion to reduce optimized add/sub op size & buil…
swolchok May 8, 2024
816b50c
Use compile-time promotion to reduce optimized le op size & build tim…
swolchok May 8, 2024
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
83 changes: 63 additions & 20 deletions kernels/optimized/cpu/op_add.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,55 @@
namespace torch {
namespace executor {
namespace native {
namespace {

template <
bool can_cast,
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct AddInner;

template <
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct AddInner<true, CTYPE_A, CTYPE_B, CTYPE_IN, CTYPE_OUT> {
static void
run(const Tensor& a, const Tensor& b, CTYPE_IN alpha_val, Tensor& out) {
apply_binary_elementwise_fn<CTYPE_A, CTYPE_B, CTYPE_OUT>(
// NOLINTNEXTLINE(facebook-hte-ConstantArgumentPassByValue)
[alpha_val](const CTYPE_A val_a, const CTYPE_B val_b) {
CTYPE_IN a_casted = static_cast<CTYPE_IN>(val_a);
CTYPE_IN b_casted = static_cast<CTYPE_IN>(val_b);
CTYPE_IN value = a_casted + alpha_val * b_casted;

return static_cast<CTYPE_OUT>(value);
},
a,
b,
out);
}
};

template <typename CTYPE_IN>
struct ReportCanCastBug {
static void run(const Tensor&, const Tensor&, CTYPE_IN, Tensor&) {
ET_DCHECK_MSG(false, "BUG: canCast should have been checked above");
}
};

template <
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct AddInner<false, CTYPE_A, CTYPE_B, CTYPE_IN, CTYPE_OUT>
: public ReportCanCastBug<CTYPE_IN> {};

} // namespace

using Tensor = exec_aten::Tensor;
using ScalarType = exec_aten::ScalarType;
Expand Down Expand Up @@ -69,26 +118,20 @@ Tensor& opt_add_out(

ET_SWITCH_REALHB_TYPES(a_type, ctx, "add.out", CTYPE_A, [&]() {
ET_SWITCH_REALHB_TYPES(b_type, ctx, "add.out", CTYPE_B, [&]() {
ET_SWITCH_REALB_TYPES(common_type, ctx, "add.out", CTYPE_IN, [&]() {
ET_SWITCH_REALHB_TYPES(out_type, ctx, "add.out", CTYPE_OUT, [&]() {
CTYPE_IN alpha_val;
ET_KERNEL_CHECK(
ctx,
utils::extract_scalar(alpha, &alpha_val),
InvalidArgument, );

apply_binary_elementwise_fn<CTYPE_A, CTYPE_B, CTYPE_OUT>(
[alpha_val](const CTYPE_A val_a, const CTYPE_B val_b) {
CTYPE_IN a_casted = static_cast<CTYPE_IN>(val_a);
CTYPE_IN b_casted = static_cast<CTYPE_IN>(val_b);
CTYPE_IN value = a_casted + alpha_val * b_casted;

return static_cast<CTYPE_OUT>(value);
},
a,
b,
out);
});
using CTYPE_IN = typename torch::executor::
promote_types<CTYPE_A, CTYPE_B, /*half_to_float*/ true>::type;
ET_DCHECK(CppTypeToScalarType<CTYPE_IN>::value == common_type);
ET_SWITCH_REALHB_TYPES(out_type, ctx, "add.out", CTYPE_OUT, [&]() {
CTYPE_IN alpha_val;
ET_KERNEL_CHECK(
ctx, utils::extract_scalar(alpha, &alpha_val), InvalidArgument, );

AddInner<
can_cast<CTYPE_IN, CTYPE_OUT>::value,
CTYPE_A,
CTYPE_B,
CTYPE_IN,
CTYPE_OUT>::run(a, b, alpha_val, out);
});
});
});
Expand Down
35 changes: 15 additions & 20 deletions kernels/optimized/cpu/op_le.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -53,31 +53,26 @@ Tensor& opt_le_tensor_out(
a.numel());
});
} else {
ScalarType common_type = promoteTypes(a_type, b_type);
ET_SWITCH_REAL_TYPES_AND(
Bool, a_type, ctx, "le.Tensor_out", CTYPE_A, [&]() {
ET_SWITCH_REAL_TYPES_AND(
Bool, b_type, ctx, "le.Tensor_out", CTYPE_B, [&]() {
using CTYPE_IN = typename torch::executor::
promote_types<CTYPE_A, CTYPE_B>::type;
ET_DCHECK(
CppTypeToScalarType<CTYPE_IN>::value ==
promoteTypes(a_type, b_type));
ET_SWITCH_REAL_TYPES_AND(
Bool, common_type, ctx, "le.Tensor_out", CTYPE_IN, [&]() {
ET_SWITCH_REAL_TYPES_AND(
Bool,
out_type,
ctx,
"le.Tensor_out",
CTYPE_OUT,
[&]() {
const size_t n = a.numel();
const CTYPE_A* a_data = a.const_data_ptr<CTYPE_A>();
const CTYPE_B* b_data = b.const_data_ptr<CTYPE_B>();
CTYPE_OUT* out_data =
out.mutable_data_ptr<CTYPE_OUT>();
for (auto i = 0; i < n; ++i) {
out_data[i] = static_cast<CTYPE_OUT>(
static_cast<CTYPE_IN>(a_data[i]) <=
static_cast<CTYPE_IN>(b_data[i]));
}
});
Bool, out_type, ctx, "le.Tensor_out", CTYPE_OUT, [&]() {
const size_t n = a.numel();
const CTYPE_A* a_data = a.const_data_ptr<CTYPE_A>();
const CTYPE_B* b_data = b.const_data_ptr<CTYPE_B>();
CTYPE_OUT* out_data = out.mutable_data_ptr<CTYPE_OUT>();
for (auto i = 0; i < n; ++i) {
out_data[i] = static_cast<CTYPE_OUT>(
static_cast<CTYPE_IN>(a_data[i]) <=
static_cast<CTYPE_IN>(b_data[i]));
}
});
});
});
Expand Down
73 changes: 59 additions & 14 deletions kernels/optimized/cpu/op_mul.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,6 +41,50 @@ bool can_use_optimized_path(
(a.numel() == b.numel() && a.numel() == out.numel()));
return can_use_optimized_path;
}

template <
bool can_cast,
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct MulInner;

template <
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct MulInner<true, CTYPE_A, CTYPE_B, CTYPE_IN, CTYPE_OUT> {
static void run(const Tensor& a, const Tensor& b, Tensor& out) {
apply_binary_elementwise_fn<CTYPE_A, CTYPE_B, CTYPE_OUT>(
// NOLINTNEXTLINE(facebook-hte-ConstantArgumentPassByValue)
[](const CTYPE_A val_a, const CTYPE_B val_b) {
CTYPE_IN a_casted = static_cast<CTYPE_IN>(val_a);
CTYPE_IN b_casted = static_cast<CTYPE_IN>(val_b);
CTYPE_IN value = a_casted * b_casted;

return static_cast<CTYPE_OUT>(value);
},
a,
b,
out);
}
};

struct ReportCanCastBug {
static void run(const Tensor&, const Tensor&, Tensor&) {
ET_DCHECK_MSG(false, "BUG: canCast should have been checked above");
}
};

template <
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct MulInner<false, CTYPE_A, CTYPE_B, CTYPE_IN, CTYPE_OUT>
: public ReportCanCastBug {};
} // namespace

Tensor& opt_mul_out(
Expand Down Expand Up @@ -86,20 +130,21 @@ Tensor& opt_mul_out(

ET_SWITCH_REALHB_TYPES(a_type, ctx, "mul.out", CTYPE_A, [&]() {
ET_SWITCH_REALHB_TYPES(b_type, ctx, "mul.out", CTYPE_B, [&]() {
ET_SWITCH_REALB_TYPES(common_type, ctx, "mul.out", CTYPE_IN, [&]() {
ET_SWITCH_REALHB_TYPES(out_type, ctx, "mul.out", CTYPE_OUT, [&]() {
apply_binary_elementwise_fn<CTYPE_A, CTYPE_B, CTYPE_OUT>(
[](const CTYPE_A val_a, const CTYPE_B val_b) {
CTYPE_IN a_casted = static_cast<CTYPE_IN>(val_a);
CTYPE_IN b_casted = static_cast<CTYPE_IN>(val_b);
CTYPE_IN value = a_casted * b_casted;

return static_cast<CTYPE_OUT>(value);
},
a,
b,
out);
});
using CTYPE_IN = typename torch::executor::
promote_types<CTYPE_A, CTYPE_B, /*half_to_float*/ true>::type;
ET_DCHECK(CppTypeToScalarType<CTYPE_IN>::value == common_type);
ET_SWITCH_REALHB_TYPES(out_type, ctx, "mul.out", CTYPE_OUT, [&]() {
apply_binary_elementwise_fn<CTYPE_A, CTYPE_B, CTYPE_OUT>(
[](const CTYPE_A val_a, const CTYPE_B val_b) {
CTYPE_IN a_casted = static_cast<CTYPE_IN>(val_a);
CTYPE_IN b_casted = static_cast<CTYPE_IN>(val_b);
CTYPE_IN value = a_casted * b_casted;

return static_cast<CTYPE_OUT>(value);
},
a,
b,
out);
});
});
});
Expand Down
82 changes: 62 additions & 20 deletions kernels/optimized/cpu/op_sub.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -17,6 +17,55 @@
namespace torch {
namespace executor {
namespace native {
namespace {

template <
bool can_cast,
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct SubInner;

template <
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct SubInner<true, CTYPE_A, CTYPE_B, CTYPE_IN, CTYPE_OUT> {
static void
run(const Tensor& a, const Tensor& b, CTYPE_IN alpha_val, Tensor& out) {
apply_binary_elementwise_fn<CTYPE_A, CTYPE_B, CTYPE_OUT>(
// NOLINTNEXTLINE(facebook-hte-ConstantArgumentPassByValue)
[alpha_val](const CTYPE_A val_a, const CTYPE_B val_b) {
CTYPE_IN a_casted = static_cast<CTYPE_IN>(val_a);
CTYPE_IN b_casted = static_cast<CTYPE_IN>(val_b);
CTYPE_IN value = a_casted - alpha_val * b_casted;

return static_cast<CTYPE_OUT>(value);
},
a,
b,
out);
}
};

template <typename CTYPE_IN>
struct ReportCanCastBug {
static void run(const Tensor&, const Tensor&, CTYPE_IN, Tensor&) {
ET_DCHECK_MSG(false, "BUG: canCast should have been checked above");
}
};

template <
typename CTYPE_A,
typename CTYPE_B,
typename CTYPE_IN,
typename CTYPE_OUT>
struct SubInner<false, CTYPE_A, CTYPE_B, CTYPE_IN, CTYPE_OUT>
: public ReportCanCastBug<CTYPE_IN> {};

} // namespace

using Tensor = exec_aten::Tensor;
using ScalarType = exec_aten::ScalarType;
Expand Down Expand Up @@ -72,26 +121,19 @@ Tensor& opt_sub_out(

ET_SWITCH_REALH_TYPES(a_type, ctx, "sub.out", CTYPE_A, [&]() {
ET_SWITCH_REALH_TYPES(b_type, ctx, "sub.out", CTYPE_B, [&]() {
ET_SWITCH_REAL_TYPES(common_type, ctx, "sub.out", CTYPE_IN, [&]() {
ET_SWITCH_REALH_TYPES(out_type, ctx, "sub.out", CTYPE_OUT, [&]() {
CTYPE_IN alpha_val;
ET_KERNEL_CHECK(
ctx,
utils::extract_scalar(alpha, &alpha_val),
InvalidArgument, );

apply_binary_elementwise_fn<CTYPE_A, CTYPE_B, CTYPE_OUT>(
[alpha_val](const CTYPE_A val_a, const CTYPE_B val_b) {
CTYPE_IN a_casted = static_cast<CTYPE_IN>(val_a);
CTYPE_IN b_casted = static_cast<CTYPE_IN>(val_b);
CTYPE_IN value = a_casted - alpha_val * b_casted;

return static_cast<CTYPE_OUT>(value);
},
a,
b,
out);
});
using CTYPE_IN = typename torch::executor::
promote_types<CTYPE_A, CTYPE_B, /*half_to_float*/ true>::type;
ET_DCHECK(CppTypeToScalarType<CTYPE_IN>::value == common_type);
ET_SWITCH_REALH_TYPES(out_type, ctx, "sub.out", CTYPE_OUT, [&]() {
CTYPE_IN alpha_val;
ET_KERNEL_CHECK(
ctx, utils::extract_scalar(alpha, &alpha_val), InvalidArgument, );
SubInner<
can_cast<CTYPE_IN, CTYPE_OUT>::value,
CTYPE_A,
CTYPE_B,
CTYPE_IN,
CTYPE_OUT>::run(a, b, alpha_val, out);
});
});
});
Expand Down
Loading
Loading