Files
composable_kernel/include/ck/utility/tuple_helper.hpp
arai713 2e3183af4f Codegen hipRTC compilation (#1579)
* updating codegen build for MIOpen access: adding .cmake for codegen component

* updating CMake

* adding in header guards for some headers due to issues with hiprtc compilation in MIOpen

* some more header guards

* putting env file in header guard

* cleaning up some includes

* updated types file for hiprtc purposes

* fixed types file: bit-wise/memcpy issue

* updating multiple utility files to deal with standard header inclusion for hiprtc

* added some more header guards in the utility files, replacing some standard header functionality

* added some more header guards

* fixing some conflicts in utility files, another round of header guards

* fixing errors in data type file

* resolved conflict errors in a few utility files

* added header guards/replicated functionality in device files

* resolved issues with standard headers in device files: device_base and device_grouped_conv_fwd_multiple_abd

* resolved issues with standard headers in device files: device_base.hpp, device_grouped_conv_fwd_multiple_abd.hpp, device_grouped_conv_fwd_multiple_abd_xdl_cshuffle.hpp

* added header guards for gridwise gemm files: gridwise_gemm_multiple_abd_xdl_cshuffle.hpp and gridwise_gemm_multiple_d_xdl_cshuffle.hpp

* fixed issue with numerics header, removed from transform_conv_fwd_to_gemm and added to device_column_to_image_impl, device_grouped_conv_fwd_multiple_abd_xdl_cshuffle, device_grouped_conv_fwd_multiple_abd_xdl_cshuffle_v3, device_image_to_column_impl

* replaced standard header usage and added header guards in block to ctile map and gridwise_gemm_pipeline_selector

* resolved errors in device_gemm_xdl_splitk_c_shuffle files in regards to replacement of standard headers in previous commit

* added replicated functionality for standard header methods in utility files

* replaced standard header functionality in threadwise tensor slice transfer files and added header guards in element_wise_operation.hpp

* temp fix for namespace error in MIOpen

* remove standard header usage in codegen device op

* removed standard header usage in elementwise files, resolved namespace errors

* formatting fix

* changed codegen argument to ON for testing

* temporarily removing codegen compiler flag for testing purposes

* added codegen flag again, set default to ON

* set codegen flag default back to OFF

* replaced enable_if_t standard header usage in data_type.hpp

* added some debug prints to pinpoint issues in MIOpen

* added print outs to debug in MIOpen

* removed debug print outs from device op

* resolved stdexcept include error

* formatting fix

* adding includes to new fp8 file to resolve ck::enable_if_t errors

* made changes to amd_wave_read_first_lane

* updated functionality in type utility file

* fixed end of file issue

* resovled errors in type utility file, added functionality to array utility file

* fixed standard header usage replication in data_type file, resolves error with failing examples on navi3x

* formatting fix

* replaced standard header usage in amd_ck_fp8 file

* added include to random_gen file

* removed and replicated standard header usage from data_type and type_convert files for fp8 changes

* replicated standard unsigned integer types in random_gen

* resolved comments from review: put calls to reinterpret_cast for size_t in header guards

* updated/added copyright headers

* removed duplicate header

* fixed typo in header guard

* updated copyright headers

---------

Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com>
2025-01-31 09:48:39 -08:00

199 lines
5.6 KiB
C++

// SPDX-License-Identifier: MIT
// Copyright (c) 2018-2025, Advanced Micro Devices, Inc. All rights reserved.
#pragma once
#include "functional4.hpp"
#include "tuple.hpp"
#ifndef CK_CODE_GEN_RTC
#include "is_detected.hpp"
#endif
namespace ck {
template <typename F, index_t N>
__host__ __device__ constexpr auto generate_tuple(F&& f, Number<N>)
{
return unpack([&f](auto&&... xs) { return make_tuple(f(xs)...); },
typename arithmetic_sequence_gen<0, N, 1>::type{});
}
template <typename F, index_t N>
__host__ __device__ constexpr auto generate_tie(F&& f, Number<N>)
{
return unpack([&f](auto&&... xs) { return tie(f(xs)...); },
typename arithmetic_sequence_gen<0, N, 1>::type{});
}
// tx and ty are tuple of references, return type of will tuple of referennce (not rvalue)
template <typename... X, typename... Y>
__host__ __device__ constexpr auto concat_tuple_of_reference(const Tuple<X&...>& tx,
const Tuple<Y&...>& ty)
{
return unpack2(
[&](auto&&... zs) { return Tuple<decltype(zs)...>{ck::forward<decltype(zs)>(zs)...}; },
tx,
ty);
}
template <typename... X, typename... Y>
__host__ __device__ constexpr auto concat_tuple(const Tuple<X...>& tx, const Tuple<Y...>& ty)
{
return unpack2(
[&](auto... zs) { return Tuple<decltype(zs)...>{ck::forward<decltype(zs)>(zs)...}; },
tx,
ty);
}
// Support any number of tuples to concat (also 1)
template <typename... X>
__host__ __device__ constexpr auto concat_tuple(const Tuple<X...>& tx)
{
return tx;
}
template <typename... X, typename... Tuples>
__host__ __device__ constexpr auto concat_tuple(const Tuple<X...>& tx, const Tuples&... tuples)
{
return concat_tuple(tx, concat_tuple(tuples...));
}
namespace detail {
template <typename F, typename X, index_t... Is>
__host__ __device__ constexpr auto transform_tuples_impl(F f, const X& x, Sequence<Is...>)
{
return make_tuple(f(x.At(Number<Is>{}))...);
}
template <typename F, typename X, typename Y, index_t... Is>
__host__ __device__ constexpr auto
transform_tuples_impl(F f, const X& x, const Y& y, Sequence<Is...>)
{
return make_tuple(f(x.At(Number<Is>{}), y.At(Number<Is>{}))...);
}
template <typename F, typename X, typename Y, typename Z, index_t... Is>
__host__ __device__ constexpr auto
transform_tuples_impl(F f, const X& x, const Y& y, const Z& z, Sequence<Is...>)
{
return make_tuple(f(x.At(Number<Is>{}), y.At(Number<Is>{}), z.At(Number<Is>{}))...);
}
} // namespace detail
template <typename F, typename X>
__host__ __device__ constexpr auto transform_tuples(F f, const X& x)
{
return detail::transform_tuples_impl(
f, x, typename arithmetic_sequence_gen<0, X::Size(), 1>::type{});
}
template <typename F, typename X, typename Y>
__host__ __device__ constexpr auto transform_tuples(F f, const X& x, const Y& y)
{
return detail::transform_tuples_impl(
f, x, y, typename arithmetic_sequence_gen<0, X::Size(), 1>::type{});
}
template <typename F, typename X, typename Y, typename Z>
__host__ __device__ constexpr auto transform_tuples(F f, const X& x, const Y& y, const Z& z)
{
return detail::transform_tuples_impl(
f, x, y, z, typename arithmetic_sequence_gen<0, X::Size(), 1>::type{});
}
// By default unroll to the flatten
template <index_t Depth = 0, index_t MaxDepth = -1>
__host__ __device__ constexpr auto UnrollNestedTuple(const Tuple<>& element)
{
return element;
}
template <index_t Depth = 0, index_t MaxDepth = -1, typename T>
__host__ __device__ constexpr auto UnrollNestedTuple(const T& element)
{
return make_tuple(element);
}
template <index_t Depth = 0, index_t MaxDepth = -1, typename... Ts>
__host__ __device__ constexpr auto UnrollNestedTuple(const Tuple<Ts...>& tuple)
{
if constexpr(Depth == MaxDepth)
{
return tuple;
}
else
{
return unpack(
[&](auto&&... ts) {
return concat_tuple(UnrollNestedTuple<Depth + 1, MaxDepth>(ts)...);
},
tuple);
}
}
template <typename... Ts>
__host__ __device__ constexpr auto TupleReverse(const Tuple<Ts...>& tuple)
{
return generate_tuple(
[&](auto i) {
using Idx = Number<Tuple<Ts...>::Size() - i - 1>;
return tuple.At(Idx{});
},
Number<Tuple<Ts...>::Size()>{});
}
// Reduce tuple values in specific range using Function
template <index_t Idx, index_t End, typename F, typename... Ts>
__host__ __device__ constexpr auto TupleReduce(F&& f, const Tuple<Ts...>& tuple)
{
static_assert(Idx < End, "Wrong parameters for TupleReduce");
if constexpr(Idx + 1 == End)
{
return tuple.At(Number<Idx>{});
}
else
{
return f(tuple.At(Number<Idx>{}), TupleReduce<Idx + 1, End>(f, tuple));
}
}
#ifndef CK_CODE_GEN_RTC
template <typename T>
using is_tuple = decltype(ck::declval<T&>().IsTuple());
#endif
template <typename... Ts>
__host__ __device__ constexpr auto IsNestedTuple(const Tuple<Ts...>&)
{
#ifndef CK_CODE_GEN_RTC
return (is_detected<is_tuple, Ts>::value || ...);
#endif
}
template <index_t depth = 0, typename T>
__host__ __device__ constexpr auto TupleDepth(const T&)
{
return depth;
}
template <index_t depth = 0, typename... Ts>
__host__ __device__ constexpr auto TupleDepth(const Tuple<Ts...>&)
{
return math::max(TupleDepth<depth + 1>(Ts{})...);
}
template <index_t from, index_t to, typename... Ts>
__host__ __device__ constexpr auto TupleSlice(const Tuple<Ts...>& tuple)
{
return generate_tuple(
[&](auto i) {
using Idx = Number<from + i>;
return tuple.At(Idx{});
},
Number<to - from>{});
}
} // namespace ck