fb335f6a5f
CUTLASS 2.0 Substantially refactored for - Better performance, particularly for native Turing Tensor Cores - Robust and durable templates spanning the design space - Encapsulated functionality embodying modern C++11 programming techniques - Optimized containers and data types for efficient, generic, portable device code Updates to: - Quick start guide - Documentation - Utilities - CUTLASS Profiler Native Turing Tensor Cores - Efficient GEMM kernels targeting Turing Tensor Cores - Mixed-precision floating point, 8-bit integer, 4-bit integer, and binarized operands Coverage of existing CUTLASS functionality: - GEMM kernels targeting CUDA and Tensor Cores in NVIDIA GPUs - Volta Tensor Cores through native mma.sync and through WMMA API - Optimizations such as parallel reductions, threadblock rasterization, and intra-threadblock reductions - Batched GEMM operations - Complex-valued GEMMs Note: this commit and all that follow require a host compiler supporting C++11 or greater.
109 lines
3.7 KiB
C++
109 lines
3.7 KiB
C++
/***************************************************************************************************
|
|
* Copyright (c) 2017-2019, NVIDIA CORPORATION. All rights reserved.
|
|
*
|
|
* Redistribution and use in source and binary forms, with or without modification, are permitted
|
|
* provided that the following conditions are met:
|
|
* * Redistributions of source code must retain the above copyright notice, this list of
|
|
* conditions and the following disclaimer.
|
|
* * Redistributions in binary form must reproduce the above copyright notice, this list of
|
|
* conditions and the following disclaimer in the documentation and/or other materials
|
|
* provided with the distribution.
|
|
* * Neither the name of the NVIDIA CORPORATION nor the names of its contributors may be used
|
|
* to endorse or promote products derived from this software without specific prior written
|
|
* permission.
|
|
*
|
|
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR
|
|
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
|
|
* FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL NVIDIA CORPORATION BE LIABLE
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
|
|
* BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
|
|
* OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
|
|
* STRICT LIABILITY, OR TOR (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
|
|
* OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
*
|
|
**************************************************************************************************/
|
|
/* \file
|
|
\brief
|
|
*/
|
|
|
|
#pragma once
|
|
|
|
#include <map>
|
|
#include <string>
|
|
|
|
|
|
#include "cutlass/library/library.h"
|
|
|
|
#include "options.h"
|
|
#include "device_allocation.h"
|
|
|
|
namespace cutlass {
|
|
namespace profiler {
|
|
|
|
/////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
/// Collection of allocations on the device
|
|
class DeviceContext {
|
|
public:
|
|
|
|
//
|
|
// Type definitions
|
|
//
|
|
using AllocationMap = std::map<std::string, DeviceAllocation *>;
|
|
|
|
private:
|
|
//
|
|
// Data members
|
|
//
|
|
|
|
/// Memory allocations that exist (owning)
|
|
DeviceAllocationList device_memory_;
|
|
|
|
/// Non-owning set of named allocations
|
|
AllocationMap allocations_;
|
|
|
|
public:
|
|
|
|
/// Allocates memory of a given type, capacity (elements), and name
|
|
DeviceAllocation *allocate_block(
|
|
std::string const &name,
|
|
library::NumericTypeID type,
|
|
size_t capacity);
|
|
|
|
/// Allocates memory of a given type, capacity (elements), and name
|
|
DeviceAllocation *allocate_tensor(
|
|
std::string const &name,
|
|
library::NumericTypeID type,
|
|
library::LayoutTypeID layout_id,
|
|
std::vector<int> const &extent,
|
|
std::vector<int> const &stride = std::vector<int>());
|
|
|
|
/// Allocates memory of a given type, capacity (elements), and name
|
|
DeviceAllocation *allocate_tensor(
|
|
Options const &options,
|
|
std::string const &name,
|
|
library::NumericTypeID type,
|
|
library::LayoutTypeID layout_id,
|
|
std::vector<int> const &extent,
|
|
std::vector<int> const &stride = std::vector<int>());
|
|
|
|
/// Clears named allocations (but does not necessarily free memory)
|
|
void clear();
|
|
|
|
/// Frees all device memory allocations
|
|
void free();
|
|
|
|
/// Gets the allocation by name
|
|
DeviceAllocation &at(std::string const &name);
|
|
|
|
size_t size() const;
|
|
|
|
AllocationMap::iterator begin();
|
|
AllocationMap::iterator end();
|
|
};
|
|
|
|
/////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
} // namespace profiler
|
|
} // namespace cutlass
|