[CCCL] 瘦身 + 补全: 移除 cudax/python/libcudacxx-tests 冗余文件, 新增 c2h 测试助手 + cmake 构建系统 + 8 个 CUDA thrust examples
变更摘要:
- 删除: cudax/ (783 files, 7.2M) — 实验性组件,竞赛不需要
- 删除: python/ (226 files, 2.0M) — Python 绑定,竞赛不需要
- 删除: libcudacxx/{test,benchmarks,codegen,cmake,share} (4432 files, 31M)
保留: libcudacxx/include/ (1463 headers, cuda::std 编译依赖)
- 新增: c2h/ (27 files) — CUB Catch2 测试辅助头文件,编译 243 个测试必需
- 新增: cmake/ (29 files) — CCCL 原生 CMake 构建系统
- 新增: thrust/examples/cuda/ (7 files) + cpp_integration/ (1 file)
async_reduce, custom_temporary_allocation, explicit_cuda_stream,
global_device_vector, range_view, unwrap_pointer, wrap_pointer, device
结果: cccl_upstream 从 74M→35M (瘦身 53%), 核心内容 100% 保留:
27/27 tuning headers, 78 benchmarks, 243 tests,
60 thrust examples, 18 CUB examples, 全部编译头文件
This commit is contained in:
@@ -1,94 +0,0 @@
|
||||
//===----------------------------------------------------------------------===//
|
||||
//
|
||||
// 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) 2024 NVIDIA CORPORATION & AFFILIATES.
|
||||
//
|
||||
//===----------------------------------------------------------------------===//
|
||||
|
||||
#include <cuda/experimental/execution.cuh>
|
||||
|
||||
#include "common/checked_receiver.cuh" // IWYU pragma: keep
|
||||
#include "common/dummy_scheduler.cuh" // IWYU pragma: keep
|
||||
#include "common/error_scheduler.cuh" // IWYU pragma: keep
|
||||
#include "common/stopped_scheduler.cuh" // IWYU pragma: keep
|
||||
#include "common/utility.cuh" // IWYU pragma: keep
|
||||
|
||||
namespace ex = cuda::experimental::execution;
|
||||
|
||||
namespace
|
||||
{
|
||||
C2H_TEST("simple task_scheduler test", "[scheduler][task_scheduler]")
|
||||
{
|
||||
ex::task_scheduler sched{dummy_scheduler{}};
|
||||
STATIC_CHECK(ex::scheduler<decltype(sched)>);
|
||||
auto sndr = sched.schedule();
|
||||
STATIC_CHECK(ex::sender<decltype(sndr)>);
|
||||
auto op = ex::connect(cuda::std::move(sndr), checked_value_receiver{});
|
||||
ex::start(op);
|
||||
// The receiver checks that it's called
|
||||
}
|
||||
|
||||
C2H_TEST("task_scheduler starts work on the correct execution context", "[scheduler][task_scheduler]")
|
||||
{
|
||||
ex::thread_context ctx;
|
||||
ex::task_scheduler sched{ctx.get_scheduler()};
|
||||
auto sndr = ex::starts_on(sched, ex::just() | ex::then([] {
|
||||
return ::std::this_thread::get_id();
|
||||
}));
|
||||
auto [tid] = ex::sync_wait(cuda::std::move(sndr)).value();
|
||||
CHECK(tid == ctx.get_id());
|
||||
}
|
||||
|
||||
#if !_CCCL_HOST_COMPILATION()
|
||||
static __device__ bool g_called = false;
|
||||
#else
|
||||
static bool g_called = false;
|
||||
#endif
|
||||
|
||||
template <class Sndr>
|
||||
struct protect : private Sndr
|
||||
{
|
||||
using sender_concept = ex::sender_t;
|
||||
_CCCL_HOST_DEVICE_API explicit protect(Sndr sndr)
|
||||
: Sndr{cuda::std::move(sndr)}
|
||||
{}
|
||||
using Sndr::connect;
|
||||
using Sndr::get_completion_signatures;
|
||||
using Sndr::get_env;
|
||||
};
|
||||
|
||||
struct test_domain
|
||||
{
|
||||
_CCCL_TEMPLATE(class Sndr, class Env)
|
||||
_CCCL_REQUIRES(ex::sender_for<Sndr, ex::bulk_chunked_t>)
|
||||
_CCCL_HOST_DEVICE_API auto transform_sender(ex::set_value_t, Sndr sndr, const Env&) const
|
||||
{
|
||||
return ex::then(protect{cuda::std::move(sndr)}, []() noexcept {
|
||||
g_called = true;
|
||||
});
|
||||
}
|
||||
};
|
||||
|
||||
C2H_TEST("bulk_unchunked dispatches correctly through task_scheduler", "[scheduler][task_scheduler]")
|
||||
{
|
||||
ex::task_scheduler sched{dummy_scheduler<test_domain>{}};
|
||||
auto sndr = ex::on(sched, ex::just(-1) | ex::bulk_chunked(ex::par_unseq, 100, [](int, int, int&) {}));
|
||||
g_called = false;
|
||||
auto [val] = ex::sync_wait(cuda::std::move(sndr)).value();
|
||||
CHECK(val == -1);
|
||||
CHECK(g_called);
|
||||
}
|
||||
|
||||
C2H_TEST("bulk dispatches correctly through task_scheduler", "[scheduler][task_scheduler]")
|
||||
{
|
||||
ex::task_scheduler sched{dummy_scheduler<test_domain>{}};
|
||||
auto sndr = ex::on(sched, ex::just(-1) | ex::bulk(ex::par_unseq, 100, [](int, int&) {}));
|
||||
g_called = false;
|
||||
auto [val] = ex::sync_wait(cuda::std::move(sndr)).value();
|
||||
CHECK(val == -1);
|
||||
CHECK(g_called);
|
||||
}
|
||||
} // namespace
|
||||
Reference in New Issue
Block a user