blob: 17dadebe975b936b7d5d7a78ac69b890d9c8e7ac (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
|
/* Copyright 2017 The TensorFlow Authors. All Rights Reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
==============================================================================*/
#ifndef TENSORFLOW_COMPILER_XLA_SERVICE_CPU_COMPILER_FUNCTOR_H_
#define TENSORFLOW_COMPILER_XLA_SERVICE_CPU_COMPILER_FUNCTOR_H_
#include "external/llvm/include/llvm/IR/LegacyPassManager.h"
#include "external/llvm/include/llvm/IR/Module.h"
#include "external/llvm/include/llvm/Object/ObjectFile.h"
#include "external/llvm/include/llvm/Target/TargetMachine.h"
#include "tensorflow/compiler/xla/service/cpu/disassembler.h"
#include "tensorflow/core/platform/logging.h"
namespace xla {
namespace cpu {
// Functor class for compiling an LLVM module down to an object file. For use by
// Orc JIT compile layer.
class CompilerFunctor {
public:
// Describes the set of vector intrinsics available to the generated code.
struct VectorIntrinsics {
bool sse_intrinsics;
bool avx_intrinsics;
};
// Returns a VectorIntrinsics where all intrinsics are available.
static VectorIntrinsics AllIntrinsics();
explicit CompilerFunctor(llvm::TargetMachine* target_machine,
const Disassembler* disassembler, int opt_level,
const VectorIntrinsics& available_intrinsics)
: target_machine_(target_machine),
disassembler_(CHECK_NOTNULL(disassembler)),
opt_level_(opt_level),
available_intrinsics_(available_intrinsics) {}
// Compile a Module to an ObjectFile.
llvm::object::OwningBinary<llvm::object::ObjectFile> operator()(
llvm::Module& module) const; // NOLINT
private:
// Populates the given pass managers based on the optimization level.
void AddOptimizationPasses(
llvm::legacy::PassManagerBase* module_passes,
llvm::legacy::FunctionPassManager* function_passes) const;
llvm::TargetMachine* target_machine_;
const Disassembler* disassembler_;
const unsigned opt_level_;
const VectorIntrinsics available_intrinsics_;
};
} // namespace cpu
} // namespace xla
#endif // TENSORFLOW_COMPILER_XLA_SERVICE_CPU_COMPILER_FUNCTOR_H_
|
