Hugging Face's logo

Spaces:
Kano001
/
GameServerX
Running

App Files Community
GameServerX
File size: 13,021 Bytes 
c61ccee
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
 
#pragma once

/*

  Provides a subset of cuSPARSE functions as templates:



    csrgeam2<scalar_t>(...)



  where scalar_t is double, float, c10::complex<double> or c10::complex<float>.

  The functions are available in at::cuda::sparse namespace.

*/

#include <ATen/cuda/CUDAContext.h>
#include <ATen/cuda/CUDASparse.h>

namespace at::cuda::sparse {

#define CUSPARSE_CSRGEAM2_BUFFERSIZE_ARGTYPES(scalar_t)             \
  cusparseHandle_t handle, int m, int n, const scalar_t *alpha,     \
      const cusparseMatDescr_t descrA, int nnzA,                    \
      const scalar_t *csrSortedValA, const int *csrSortedRowPtrA,   \
      const int *csrSortedColIndA, const scalar_t *beta,            \
      const cusparseMatDescr_t descrB, int nnzB,                    \
      const scalar_t *csrSortedValB, const int *csrSortedRowPtrB,   \
      const int *csrSortedColIndB, const cusparseMatDescr_t descrC, \
      const scalar_t *csrSortedValC, const int *csrSortedRowPtrC,   \
      const int *csrSortedColIndC, size_t *pBufferSizeInBytes

template <typename scalar_t>
inline void csrgeam2_bufferSizeExt(

    CUSPARSE_CSRGEAM2_BUFFERSIZE_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::csrgeam2_bufferSizeExt: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void csrgeam2_bufferSizeExt<float>(
    CUSPARSE_CSRGEAM2_BUFFERSIZE_ARGTYPES(float));
template <>
void csrgeam2_bufferSizeExt<double>(
    CUSPARSE_CSRGEAM2_BUFFERSIZE_ARGTYPES(double));
template <>
void csrgeam2_bufferSizeExt<c10::complex<float>>(
    CUSPARSE_CSRGEAM2_BUFFERSIZE_ARGTYPES(c10::complex<float>));
template <>
void csrgeam2_bufferSizeExt<c10::complex<double>>(
    CUSPARSE_CSRGEAM2_BUFFERSIZE_ARGTYPES(c10::complex<double>));

#define CUSPARSE_CSRGEAM2_NNZ_ARGTYPES()                                      \
  cusparseHandle_t handle, int m, int n, const cusparseMatDescr_t descrA,     \
      int nnzA, const int *csrSortedRowPtrA, const int *csrSortedColIndA,     \
      const cusparseMatDescr_t descrB, int nnzB, const int *csrSortedRowPtrB, \
      const int *csrSortedColIndB, const cusparseMatDescr_t descrC,           \
      int *csrSortedRowPtrC, int *nnzTotalDevHostPtr, void *workspace

template <typename scalar_t>
inline void csrgeam2Nnz(CUSPARSE_CSRGEAM2_NNZ_ARGTYPES()) {
  TORCH_CUDASPARSE_CHECK(cusparseXcsrgeam2Nnz(
      handle,
      m,
      n,
      descrA,
      nnzA,
      csrSortedRowPtrA,
      csrSortedColIndA,
      descrB,
      nnzB,
      csrSortedRowPtrB,
      csrSortedColIndB,
      descrC,
      csrSortedRowPtrC,
      nnzTotalDevHostPtr,
      workspace));
}

#define CUSPARSE_CSRGEAM2_ARGTYPES(scalar_t)                                 \
  cusparseHandle_t handle, int m, int n, const scalar_t *alpha,              \
      const cusparseMatDescr_t descrA, int nnzA,                             \
      const scalar_t *csrSortedValA, const int *csrSortedRowPtrA,            \
      const int *csrSortedColIndA, const scalar_t *beta,                     \
      const cusparseMatDescr_t descrB, int nnzB,                             \
      const scalar_t *csrSortedValB, const int *csrSortedRowPtrB,            \
      const int *csrSortedColIndB, const cusparseMatDescr_t descrC,          \
      scalar_t *csrSortedValC, int *csrSortedRowPtrC, int *csrSortedColIndC, \
      void *pBuffer

template <typename scalar_t>
inline void csrgeam2(CUSPARSE_CSRGEAM2_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::csrgeam2: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void csrgeam2<float>(CUSPARSE_CSRGEAM2_ARGTYPES(float));
template <>
void csrgeam2<double>(CUSPARSE_CSRGEAM2_ARGTYPES(double));
template <>
void csrgeam2<c10::complex<float>>(
    CUSPARSE_CSRGEAM2_ARGTYPES(c10::complex<float>));
template <>
void csrgeam2<c10::complex<double>>(
    CUSPARSE_CSRGEAM2_ARGTYPES(c10::complex<double>));

#define CUSPARSE_BSRMM_ARGTYPES(scalar_t)                                    \
  cusparseHandle_t handle, cusparseDirection_t dirA,                         \
      cusparseOperation_t transA, cusparseOperation_t transB, int mb, int n, \
      int kb, int nnzb, const scalar_t *alpha,                               \
      const cusparseMatDescr_t descrA, const scalar_t *bsrValA,              \
      const int *bsrRowPtrA, const int *bsrColIndA, int blockDim,            \
      const scalar_t *B, int ldb, const scalar_t *beta, scalar_t *C, int ldc

template <typename scalar_t>
inline void bsrmm(CUSPARSE_BSRMM_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::bsrmm: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void bsrmm<float>(CUSPARSE_BSRMM_ARGTYPES(float));
template <>
void bsrmm<double>(CUSPARSE_BSRMM_ARGTYPES(double));
template <>
void bsrmm<c10::complex<float>>(CUSPARSE_BSRMM_ARGTYPES(c10::complex<float>));
template <>
void bsrmm<c10::complex<double>>(CUSPARSE_BSRMM_ARGTYPES(c10::complex<double>));

#define CUSPARSE_BSRMV_ARGTYPES(scalar_t)                                    \
  cusparseHandle_t handle, cusparseDirection_t dirA,                         \
      cusparseOperation_t transA, int mb, int nb, int nnzb,                  \
      const scalar_t *alpha, const cusparseMatDescr_t descrA,                \
      const scalar_t *bsrValA, const int *bsrRowPtrA, const int *bsrColIndA, \
      int blockDim, const scalar_t *x, const scalar_t *beta, scalar_t *y

template <typename scalar_t>
inline void bsrmv(CUSPARSE_BSRMV_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::bsrmv: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void bsrmv<float>(CUSPARSE_BSRMV_ARGTYPES(float));
template <>
void bsrmv<double>(CUSPARSE_BSRMV_ARGTYPES(double));
template <>
void bsrmv<c10::complex<float>>(CUSPARSE_BSRMV_ARGTYPES(c10::complex<float>));
template <>
void bsrmv<c10::complex<double>>(CUSPARSE_BSRMV_ARGTYPES(c10::complex<double>));

#if AT_USE_HIPSPARSE_TRIANGULAR_SOLVE()

#define CUSPARSE_BSRSV2_BUFFER_ARGTYPES(scalar_t)                 \
  cusparseHandle_t handle, cusparseDirection_t dirA,              \
      cusparseOperation_t transA, int mb, int nnzb,               \
      const cusparseMatDescr_t descrA, scalar_t *bsrValA,         \
      const int *bsrRowPtrA, const int *bsrColIndA, int blockDim, \
      bsrsv2Info_t info, int *pBufferSizeInBytes

template <typename scalar_t>
inline void bsrsv2_bufferSize(CUSPARSE_BSRSV2_BUFFER_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::bsrsv2_bufferSize: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void bsrsv2_bufferSize<float>(CUSPARSE_BSRSV2_BUFFER_ARGTYPES(float));
template <>
void bsrsv2_bufferSize<double>(CUSPARSE_BSRSV2_BUFFER_ARGTYPES(double));
template <>
void bsrsv2_bufferSize<c10::complex<float>>(
    CUSPARSE_BSRSV2_BUFFER_ARGTYPES(c10::complex<float>));
template <>
void bsrsv2_bufferSize<c10::complex<double>>(
    CUSPARSE_BSRSV2_BUFFER_ARGTYPES(c10::complex<double>));

#define CUSPARSE_BSRSV2_ANALYSIS_ARGTYPES(scalar_t)               \
  cusparseHandle_t handle, cusparseDirection_t dirA,              \
      cusparseOperation_t transA, int mb, int nnzb,               \
      const cusparseMatDescr_t descrA, const scalar_t *bsrValA,   \
      const int *bsrRowPtrA, const int *bsrColIndA, int blockDim, \
      bsrsv2Info_t info, cusparseSolvePolicy_t policy, void *pBuffer

template <typename scalar_t>
inline void bsrsv2_analysis(CUSPARSE_BSRSV2_ANALYSIS_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::bsrsv2_analysis: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void bsrsv2_analysis<float>(CUSPARSE_BSRSV2_ANALYSIS_ARGTYPES(float));
template <>
void bsrsv2_analysis<double>(CUSPARSE_BSRSV2_ANALYSIS_ARGTYPES(double));
template <>
void bsrsv2_analysis<c10::complex<float>>(
    CUSPARSE_BSRSV2_ANALYSIS_ARGTYPES(c10::complex<float>));
template <>
void bsrsv2_analysis<c10::complex<double>>(
    CUSPARSE_BSRSV2_ANALYSIS_ARGTYPES(c10::complex<double>));

#define CUSPARSE_BSRSV2_SOLVE_ARGTYPES(scalar_t)                           \
  cusparseHandle_t handle, cusparseDirection_t dirA,                       \
      cusparseOperation_t transA, int mb, int nnzb, const scalar_t *alpha, \
      const cusparseMatDescr_t descrA, const scalar_t *bsrValA,            \
      const int *bsrRowPtrA, const int *bsrColIndA, int blockDim,          \
      bsrsv2Info_t info, const scalar_t *x, scalar_t *y,                   \
      cusparseSolvePolicy_t policy, void *pBuffer

template <typename scalar_t>
inline void bsrsv2_solve(CUSPARSE_BSRSV2_SOLVE_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::bsrsv2_solve: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void bsrsv2_solve<float>(CUSPARSE_BSRSV2_SOLVE_ARGTYPES(float));
template <>
void bsrsv2_solve<double>(CUSPARSE_BSRSV2_SOLVE_ARGTYPES(double));
template <>
void bsrsv2_solve<c10::complex<float>>(
    CUSPARSE_BSRSV2_SOLVE_ARGTYPES(c10::complex<float>));
template <>
void bsrsv2_solve<c10::complex<double>>(
    CUSPARSE_BSRSV2_SOLVE_ARGTYPES(c10::complex<double>));

#define CUSPARSE_BSRSM2_BUFFER_ARGTYPES(scalar_t)                            \
  cusparseHandle_t handle, cusparseDirection_t dirA,                         \
      cusparseOperation_t transA, cusparseOperation_t transX, int mb, int n, \
      int nnzb, const cusparseMatDescr_t descrA, scalar_t *bsrValA,          \
      const int *bsrRowPtrA, const int *bsrColIndA, int blockDim,            \
      bsrsm2Info_t info, int *pBufferSizeInBytes

template <typename scalar_t>
inline void bsrsm2_bufferSize(CUSPARSE_BSRSM2_BUFFER_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::bsrsm2_bufferSize: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void bsrsm2_bufferSize<float>(CUSPARSE_BSRSM2_BUFFER_ARGTYPES(float));
template <>
void bsrsm2_bufferSize<double>(CUSPARSE_BSRSM2_BUFFER_ARGTYPES(double));
template <>
void bsrsm2_bufferSize<c10::complex<float>>(
    CUSPARSE_BSRSM2_BUFFER_ARGTYPES(c10::complex<float>));
template <>
void bsrsm2_bufferSize<c10::complex<double>>(
    CUSPARSE_BSRSM2_BUFFER_ARGTYPES(c10::complex<double>));

#define CUSPARSE_BSRSM2_ANALYSIS_ARGTYPES(scalar_t)                          \
  cusparseHandle_t handle, cusparseDirection_t dirA,                         \
      cusparseOperation_t transA, cusparseOperation_t transX, int mb, int n, \
      int nnzb, const cusparseMatDescr_t descrA, const scalar_t *bsrValA,    \
      const int *bsrRowPtrA, const int *bsrColIndA, int blockDim,            \
      bsrsm2Info_t info, cusparseSolvePolicy_t policy, void *pBuffer

template <typename scalar_t>
inline void bsrsm2_analysis(CUSPARSE_BSRSM2_ANALYSIS_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::bsrsm2_analysis: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void bsrsm2_analysis<float>(CUSPARSE_BSRSM2_ANALYSIS_ARGTYPES(float));
template <>
void bsrsm2_analysis<double>(CUSPARSE_BSRSM2_ANALYSIS_ARGTYPES(double));
template <>
void bsrsm2_analysis<c10::complex<float>>(
    CUSPARSE_BSRSM2_ANALYSIS_ARGTYPES(c10::complex<float>));
template <>
void bsrsm2_analysis<c10::complex<double>>(
    CUSPARSE_BSRSM2_ANALYSIS_ARGTYPES(c10::complex<double>));

#define CUSPARSE_BSRSM2_SOLVE_ARGTYPES(scalar_t)                             \
  cusparseHandle_t handle, cusparseDirection_t dirA,                         \
      cusparseOperation_t transA, cusparseOperation_t transX, int mb, int n, \
      int nnzb, const scalar_t *alpha, const cusparseMatDescr_t descrA,      \
      const scalar_t *bsrValA, const int *bsrRowPtrA, const int *bsrColIndA, \
      int blockDim, bsrsm2Info_t info, const scalar_t *B, int ldb,           \
      scalar_t *X, int ldx, cusparseSolvePolicy_t policy, void *pBuffer

template <typename scalar_t>
inline void bsrsm2_solve(CUSPARSE_BSRSM2_SOLVE_ARGTYPES(scalar_t)) {
  TORCH_INTERNAL_ASSERT(
      false,
      "at::cuda::sparse::bsrsm2_solve: not implemented for ",
      typeid(scalar_t).name());
}

template <>
void bsrsm2_solve<float>(CUSPARSE_BSRSM2_SOLVE_ARGTYPES(float));
template <>
void bsrsm2_solve<double>(CUSPARSE_BSRSM2_SOLVE_ARGTYPES(double));
template <>
void bsrsm2_solve<c10::complex<float>>(
    CUSPARSE_BSRSM2_SOLVE_ARGTYPES(c10::complex<float>));
template <>
void bsrsm2_solve<c10::complex<double>>(
    CUSPARSE_BSRSM2_SOLVE_ARGTYPES(c10::complex<double>));

#endif // AT_USE_HIPSPARSE_TRIANGULAR_SOLVE

} // namespace at::cuda::sparse