1 /*
2 * Copyright (c) 2024, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation. Oracle designates this
8 * particular file as subject to the "Classpath" exception as provided
9 * by Oracle in the LICENSE file that accompanied this code.
10 *
11 * This code is distributed in the hope that it will be useful, but WITHOUT
12 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
13 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
14 * version 2 for more details (a copy is included in the LICENSE file that
15 * accompanied this code).
16 *
17 * You should have received a copy of the GNU General Public License version
18 * 2 along with this work; if not, write to the Free Software Foundation,
19 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
20 *
21 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
22 * or visit www.oracle.com if you need additional information or have any
23 * questions.
24 */
25 #pragma once
26 #define CUDA_TYPES
27 #ifdef __APPLE__
28
29 #define LongUnsignedNewline "%llu\n"
30 #define Size_tNewline "%lu\n"
31 #define LongHexNewline "(0x%llx)\n"
32 #define alignedMalloc(size, alignment) memalign(alignment, size)
33 #define SNPRINTF snprintf
34 #else
35
36 #include <malloc.h>
37
38 #define LongHexNewline "(0x%lx)\n"
39 #define LongUnsignedNewline "%lu\n"
40 #define Size_tNewline "%lu\n"
41 #if defined (_WIN32)
42 #include "windows.h"
43 #define alignedMalloc(size, alignment) _aligned_malloc(size, alignment)
44 #define SNPRINTF _snprintf
45 #else
46 #define alignedMalloc(size, alignment) memalign(alignment, size)
47 #define SNPRINTF snprintf
48 #endif
49 #endif
50
51 #include <iostream>
52 #include <cuda.h>
53 #include <builtin_types.h>
54
55 #include "shared.h"
56
57 #include <fstream>
58 #include <thread>
59
60 struct WHERE{
61 const char* f;
62 int l;
63 cudaError_enum e;
64 const char* t;
65 void report() const{
66 if (e == CUDA_SUCCESS){
67 // std::cout << t << " OK at " << f << " line " << l << std::endl;
68 }else {
69 const char *buf;
70 cuGetErrorName(e, &buf);
71 std::cerr << t << " CUDA error = " << e << " " << buf <<std::endl<< " " << f << " line " << l << std::endl;
72 exit(-1);
73 }
74 }
75 };
76
77 class PtxSource final : public Text {
78 public:
79 PtxSource();
80 explicit PtxSource(size_t len);
81 PtxSource(size_t len, char *text);
82 PtxSource(size_t len, char *text, bool isCopy);
83 explicit PtxSource(char *text);
84 ~PtxSource() override = default;
85 };
86
87 class CudaSource final :public Text {
88 public:
89 CudaSource(size_t len, char *text, bool isCopy);
90 explicit CudaSource(size_t len);
91 explicit CudaSource(char* text);
92 CudaSource();
93 ~CudaSource() override = default;
94 };
95
96 class CudaBackend final : public Backend {
97 public:
98 class CudaQueue final : public Backend::Queue {
99 public:
100 std::thread::id streamCreationThread;
101 CUstream cuStream;
102 explicit CudaQueue(Backend *backend);
103 void init();
104 void wait() override;
105
106 void release() override;
107
108 void computeStart() override;
109
110 void computeEnd() override;
111
112 void copyToDevice(Buffer *buffer) override;
113
114 void copyFromDevice(Buffer *buffer) override;
115
116 void dispatch(KernelContext *kernelContext, CompilationUnit::Kernel *kernel) override;
117
118 ~CudaQueue() override;
119
120 };
121
122 class CudaBuffer final : public Buffer {
123 public:
124 CUdeviceptr devicePtr;
125 CudaBuffer(Backend *backend, BufferState *bufferState);
126 ~CudaBuffer() override;
127 };
128
129 class CudaModule final : public CompilationUnit {
130 CUmodule module;
131 CudaSource cudaSource;
132 PtxSource ptxSource;
133 Log log;
134
135 public:
136 class CudaKernel final : public Kernel {
137
138 public:
139 bool setArg(KernelArg *arg) override;
140 bool setArg(KernelArg *arg, Buffer *buffer) override;
141 CudaKernel(Backend::CompilationUnit *program, char* name, CUfunction function);
142 ~CudaKernel() override;
143 static CudaKernel * of(long kernelHandle);
144 static CudaKernel * of(Backend::CompilationUnit::Kernel *kernel);
145
146 CUfunction function;
147 void *argslist[100]{};
148 };
149 CudaModule(Backend *backend, char *cudaSrc, char *log, bool ok, CUmodule module);
150 ~CudaModule() override;
151 static CudaModule * of(long moduleHandle);
152 //static CudaModule * of(CompilationUnit *compilationUnit);
153 Kernel *getKernel(int nameLen, char *name) override;
154 CudaKernel *getCudaKernel(char *name);
155 CudaKernel *getCudaKernel(int nameLen, char *name);
156 bool programOK();
157 };
158
159 private:
160 CUresult initStatus;
161 CUdevice device;
162 CUcontext context;
163 public:
164 void info() override;
165 CudaModule * compile(const CudaSource *cudaSource);
166 CudaModule * compile(const CudaSource &cudaSource);
167 CudaModule * compile(const PtxSource *ptxSource);
168 CudaModule * compile(const PtxSource &ptxSource);
169 static PtxSource *nvcc(const CudaSource *cudaSource);
170 CompilationUnit * compile(int len, char *source) override;
171 void computeStart() override;
172 void computeEnd() override;
173 CudaBuffer * getOrCreateBuffer(BufferState *bufferState) override;
174 bool getBufferFromDeviceIfDirty(void *memorySegment, long memorySegmentLength) override;
175
176 explicit CudaBackend(int mode);
177
178 ~CudaBackend() override;
179 static CudaBackend * of(long backendHandle);
180 static CudaBackend * of(Backend *backend);
181 };
182
183