Skip to content

203 use const instead of const reference on device function parameters that are computed inside the kernel #204

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
morousg merged 3 commits into from
Feb 18, 2026
Merged

203 use const instead of const reference on device function parameters that are computed inside the kernel #204

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
fd8274d
Making get_opt generic like get
morousg Feb 18, 2026
952eff1
Simplified get wrapper and get_opt implementations
morousg Feb 18, 2026
e536e14
Made all values that reside in registers, to be passed as const value…
morousg Feb 18, 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
4 changes: 2 additions & 2 deletions include/fused_kernel/algorithms/basic_ops/algebraic.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -38,7 +38,7 @@ namespace fk {
using Parent = BinaryOperation<float3, M3x3Float, float3, MxVFloat3<BinaryType>>;
DECLARE_BINARY_PARENT

FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input, const ParamsType& params) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input, const ParamsType& params) {
const float3 xOut = input * params.x;
const float3 yOut = input * params.y;
const float3 zOut = input * params.z;
Expand All @@ -54,7 +54,7 @@ namespace fk {
FK_STATIC_STRUCT(MxVFloat3, SelfType)
using Parent = UnaryOperation<Tuple<float3, M3x3Float>, float3, MxVFloat3<UnaryType>>;
DECLARE_UNARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input) {
const float3 xOut = get<0>(input) * get<1>(input).x;
const float3 yOut = get<0>(input) * get<1>(input).y;
const float3 zOut = get<0>(input) * get<1>(input).z;
Expand Down
10 changes: 5 additions & 5 deletions include/fused_kernel/algorithms/basic_ops/arithmetic.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ namespace fk {
FK_STATIC_STRUCT(Add, SelfType)
using Parent = BinaryOperation<I, P, O, Add<I, P, O, BinaryType>>;
DECLARE_BINARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input, const ParamsType& params) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input, const ParamsType& params) {
return input + params;
}
};
Expand All @@ -43,7 +43,7 @@ namespace fk {
FK_STATIC_STRUCT(Add, SelfType)
using Parent = UnaryOperation<Tuple<I1, I2>, O, Add<I1, I2, O, UnaryType>>;
DECLARE_UNARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input) {
return get<0>(input) + get<1>(input);
}
};
Expand All @@ -56,7 +56,7 @@ namespace fk {
FK_STATIC_STRUCT(Sub, SelfType)
using Parent = BinaryOperation<I, P, O, Sub<I, P, O>>;
DECLARE_BINARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input, const ParamsType& params) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input, const ParamsType& params) {
return input - params;
}
};
Expand All @@ -69,7 +69,7 @@ namespace fk {
FK_STATIC_STRUCT(Mul, SelfType)
using Parent = BinaryOperation<I, P, O, Mul<I, P, O>>;
DECLARE_BINARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input, const ParamsType& params) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input, const ParamsType& params) {
return input * params;
}
};
Expand All @@ -82,7 +82,7 @@ namespace fk {
FK_STATIC_STRUCT(Div, SelfType)
using Parent = BinaryOperation<I, P, O, Div<I, P, O>>;
DECLARE_BINARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input, const ParamsType& params) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input, const ParamsType& params) {
return input / params;
}
};
Expand Down
2 changes: 1 addition & 1 deletion include/fused_kernel/algorithms/basic_ops/cast.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,7 @@ namespace fk {
FK_STATIC_STRUCT(Cast, SelfType)
using Parent = UnaryOperation<I, O, Cast<I, O>>;
DECLARE_UNARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input) {
return cxp::cast<OutputType>::f(input);
}
};
Expand Down
18 changes: 9 additions & 9 deletions include/fused_kernel/algorithms/basic_ops/logical.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,7 @@ namespace fk {
FK_STATIC_STRUCT(IsEven, SelfType)
using Parent = UnaryOperation<I, bool, IsEven<I>>;
DECLARE_UNARY_PARENT
FK_HOST_DEVICE_FUSE auto exec(const InputType& input) {
FK_HOST_DEVICE_FUSE auto exec(const InputType input) {
return cxp::is_even::f(input);
}
};
Expand All @@ -42,7 +42,7 @@ namespace fk {
FK_STATIC_STRUCT(Max, SelfType)
using Parent = BinaryOperation<I, P, O, Max<I, P, O, BinaryType>>;
DECLARE_BINARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input, const ParamsType& params) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input, const ParamsType& params) {
return cxp::max::f(input, params);
}
};
Expand All @@ -55,7 +55,7 @@ namespace fk {
FK_STATIC_STRUCT(Max, SelfType)
using Parent = UnaryOperation<Tuple<I, P>, O, Max<I, P, O, UnaryType>>;
DECLARE_UNARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input) {
return cxp::max::f(get<0>(input), get<1>(input));
}
};
Expand All @@ -68,7 +68,7 @@ namespace fk {
FK_STATIC_STRUCT(Min, SelfType)
using Parent = BinaryOperation<I, P, O, Min<I, P, O, BinaryType>>;
DECLARE_BINARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input, const ParamsType& params) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input, const ParamsType& params) {
return cxp::min::f(input, params);
}
};
Expand All @@ -81,7 +81,7 @@ namespace fk {
FK_STATIC_STRUCT(Min, SelfType)
using Parent = UnaryOperation<Tuple<I, P>, O, Min<I, P, O, UnaryType>>;
DECLARE_UNARY_PARENT
FK_HOST_DEVICE_FUSE OutputType exec(const InputType& input) {
FK_HOST_DEVICE_FUSE OutputType exec(const InputType input) {
return cxp::min::f(get<0>(input), get<1>(input));
}
};
Expand All @@ -95,21 +95,21 @@ namespace fk {
using Parent = UnaryOperation<Tuple<I1, I2>, bool, Equal<I1, I2>>;
DECLARE_UNARY_PARENT
template <int N = cn<I1>>
FK_HOST_DEVICE_FUSE std::enable_if_t<N==1, OutputType> exec(const InputType& input) {
FK_HOST_DEVICE_FUSE std::enable_if_t<N==1, OutputType> exec(const InputType input) {
return get<0>(input) == get<1>(input);
}
template <int N = cn<I1>>
FK_HOST_DEVICE_FUSE std::enable_if_t<N == 2, OutputType> exec(const InputType& input) {
FK_HOST_DEVICE_FUSE std::enable_if_t<N == 2, OutputType> exec(const InputType input) {
const auto result = get<0>(input) == get<1>(input);
return result.x && result.y;
}
template <int N = cn<I1>>
FK_HOST_DEVICE_FUSE std::enable_if_t<N == 3, OutputType> exec(const InputType& input) {
FK_HOST_DEVICE_FUSE std::enable_if_t<N == 3, OutputType> exec(const InputType input) {
const auto result = get<0>(input) == get<1>(input);
return result.x && result.y && result.z;
}
template <int N = cn<I1>>
FK_HOST_DEVICE_FUSE std::enable_if_t<N == 4, OutputType> exec(const InputType& input) {
FK_HOST_DEVICE_FUSE std::enable_if_t<N == 4, OutputType> exec(const InputType input) {
const auto result = get<0>(input) == get<1>(input);
return result.x && result.y && result.z && result.w;
}
Expand Down
Loading