56fd68e7dd
CCCL (CUDA C++ Core Libraries) provides: - CUB: device/block/warp-level GPU primitives (reduce, scan, sort, topk) - Thrust: high-level parallel algorithms (transform_reduce, sort, scan) - libcudacxx: CUDA C++ standard library (atomics, barriers, memory) - cudax: experimental features (memory resources, allocators) - Tuning policies: per-SM hardware-specific algorithm parameters Competition optimization vectors mapped to CCCL: - Output TPS (83% weight): warp_reduce, block_reduce, device_topk - Input TPS (14% weight): device_scan, block_load, prefetch - Cache TPS (3% weight): prefix caching strategy patterns - Memory (0.9 util): pooled/cached/buddy allocators Source: https://github.com/NVIDIA/cccl (shallow clone, HEAD only) License: Apache-2.0
137 lines
3.2 KiB
C++
137 lines
3.2 KiB
C++
//===----------------------------------------------------------------------===//
|
|
//
|
|
// Part of CUDA Experimental in CUDA C++ Core Libraries,
|
|
// under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
// SPDX-FileCopyrightText: Copyright (c) 2025 NVIDIA CORPORATION & AFFILIATES.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#pragma once
|
|
|
|
#include <cuda/std/type_traits>
|
|
|
|
#include <format>
|
|
#include <string>
|
|
|
|
#include "../util/errors.h"
|
|
#include "../util/types.h"
|
|
|
|
extern const char* jit_template_header_contents;
|
|
|
|
template <typename Arg>
|
|
struct parameter_mapping;
|
|
|
|
template <typename Tpl>
|
|
struct template_id
|
|
{};
|
|
|
|
// tagged_arg is needed to pass storage type information to the parameter
|
|
// mapping. This is needed because different args may have different storage
|
|
// types.
|
|
template <typename StorageT, typename T>
|
|
struct tagged_arg
|
|
{
|
|
using storage_type = StorageT;
|
|
using value_type = T;
|
|
T value;
|
|
};
|
|
|
|
template <typename T>
|
|
struct is_tagged_arg : cuda::std::false_type
|
|
{};
|
|
|
|
template <typename StorageT, typename T>
|
|
struct is_tagged_arg<tagged_arg<StorageT, T>> : cuda::std::true_type
|
|
{};
|
|
|
|
template <typename T>
|
|
struct arg_traits
|
|
{
|
|
using storage_type = storage_t;
|
|
using value_type = T;
|
|
|
|
static constexpr const T& unwrap(const T& value)
|
|
{
|
|
return value;
|
|
}
|
|
|
|
static constexpr auto wrap(const T& value)
|
|
{
|
|
return tagged_arg<storage_t, T>{value};
|
|
}
|
|
};
|
|
|
|
template <typename StorageT, typename T>
|
|
struct arg_traits<tagged_arg<StorageT, T>>
|
|
{
|
|
using storage_type = StorageT;
|
|
using value_type = T;
|
|
|
|
static constexpr const T& unwrap(const tagged_arg<StorageT, T>& value)
|
|
{
|
|
return value.value;
|
|
}
|
|
|
|
static constexpr const auto& wrap(const tagged_arg<StorageT, T>& value)
|
|
{
|
|
return value;
|
|
}
|
|
};
|
|
|
|
template <typename T>
|
|
struct mapping_arg_type
|
|
{
|
|
using type = T;
|
|
};
|
|
|
|
template <typename StorageT, typename T>
|
|
struct mapping_arg_type<tagged_arg<StorageT, T>>
|
|
{
|
|
using type = T;
|
|
};
|
|
|
|
struct specialization
|
|
{
|
|
std::string type_name;
|
|
std::string aux_code = "";
|
|
};
|
|
|
|
template <typename Tag,
|
|
typename Traits,
|
|
typename... Args,
|
|
typename = Traits::template type<void, parameter_mapping<typename mapping_arg_type<Args>::type>::archetype...>>
|
|
specialization get_specialization(template_id<Traits> id, Args... args)
|
|
{
|
|
#ifdef __CUDA_ARCH__
|
|
return specialization{};
|
|
#else
|
|
if constexpr (requires { Traits::template special<Tag>(args...); })
|
|
{
|
|
if (auto result = Traits::template special<Tag>(args...))
|
|
{
|
|
return *result;
|
|
}
|
|
}
|
|
|
|
std::string tag_name;
|
|
check(cccl_type_name_from_nvrtc<Tag>(&tag_name));
|
|
|
|
auto map = [&](auto arg) {
|
|
using arg_t = cuda::std::decay_t<decltype(arg)>;
|
|
using map_t = typename mapping_arg_type<arg_t>::type;
|
|
return parameter_mapping<map_t>::map(id, arg);
|
|
};
|
|
|
|
auto aux = [&](auto arg) {
|
|
using arg_t = cuda::std::decay_t<decltype(arg)>;
|
|
using map_t = typename mapping_arg_type<arg_t>::type;
|
|
return parameter_mapping<map_t>::aux(id, arg);
|
|
};
|
|
|
|
return {std::format("{}<{}{}>", Traits::name, tag_name, ((", " + map(args)) + ...)),
|
|
std::format("struct {};", tag_name) + (aux(args) + ...)};
|
|
#endif
|
|
}
|