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.
94 lines
3.8 KiB
C++
94 lines
3.8 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 Helper functions for mapping CUTLASS concepts to cuBLAS.
|
|
*/
|
|
|
|
#pragma once
|
|
|
|
#if CUTLASS_ENABLE_CUBLAS
|
|
#include <cublas_v2.h>
|
|
|
|
#include "cutlass/cutlass.h"
|
|
#include "cutlass/library/library.h"
|
|
|
|
/////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
namespace cutlass {
|
|
namespace profiler {
|
|
|
|
/////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
/// Converts a cuBLAS status to cutlass::Status
|
|
Status get_cutlass_status(cublasStatus_t cublas);
|
|
|
|
/// Maps a CUTLASS tensor layout to a cuBLAS transpose operation
|
|
cublasOperation_t get_cublas_transpose_operation(library::LayoutTypeID layout);
|
|
|
|
/// Maps a CUTLASS numeric type to a cuBLAS data type enumeration
|
|
bool get_cublas_datatype(cublasDataType_t &data_type, library::NumericTypeID element_type);
|
|
|
|
/// Gets the cublas algorithm given threadblock tile dimensions and math opcode class
|
|
cublasGemmAlgo_t get_cublas_gemm_algo(
|
|
int cta_m,
|
|
int cta_n,
|
|
int cta_k,
|
|
library::OpcodeClassID opcode_class);
|
|
|
|
/// Returns a status if cuBLAS can satisfy a particular GEMM description
|
|
Status cublas_satisfies(library::GemmDescription const &desc);
|
|
|
|
/// This is a helper class to create cublasHandle_t automatically on CublasCreate object creation and
|
|
/// to destroy cublasHandle_t on CublasCreate object destruction.
|
|
/// Additionaly, it provides implicit cast from CublasCreate's object to cublasHandle_t's object
|
|
class CublasCreate {
|
|
private:
|
|
cublasHandle_t handle;
|
|
cublasStatus_t status;
|
|
|
|
public:
|
|
CublasCreate() {
|
|
status = cublasCreate(&handle);
|
|
}
|
|
|
|
~CublasCreate() {
|
|
cublasDestroy(handle);
|
|
}
|
|
|
|
/// Implicit cast CublasCreate object to cublasHandle_t
|
|
operator cublasHandle_t() const { return handle; }
|
|
|
|
/// returns cublasStatus_t for handle creation
|
|
cublasStatus_t get_cublas_create_status() { return status; }
|
|
};
|
|
|
|
/////////////////////////////////////////////////////////////////////////////////////////////////
|
|
|
|
} // namespace profiler
|
|
} // namespace cutlass
|
|
|
|
|
|
#endif // #if CUTLASS_ENABLE_CUBLAS
|