I successfully installed pycuda,by install the Microsoft VC compiler for python
2.7, but when i executing the test code, it comes out a new error . The
tarceback is too long to read,I have no clue about it,could anybody offer me
some help again please?
I dont know how to set the .theanorc.txt, I copied it from net and edit it a
little,i dont know how to make all these thing rignt, ,intalling cuda 7.5 also
dont work.... I really need your help,have you met this situation? I use
vs2010, anaconda 2, use pip intall to install theano and pycuda,Use conda
install mingw libpython, os is win7 64bit,my gpu is GTX770m,and set all the env
varible to path,but while using cpu,the code can run correctly,but after
setting it to gpu, it comes out a lot error. I was so depressed，Thank you
advance.
from theano import function, config, shared, sandbox
import theano.tensor as T
import numpy
import time
vlen = 10 * 30 * 768 # 10 x #cores x # threads per core
iters = 1000
rng = numpy.random.RandomState(22)
x = shared(numpy.asarray(rng.rand(vlen), config.floatX))
f = function([], T.exp(x))
print f.maker.fgraph.toposort()
t0 = time.time()
for i in xrange(iters):
r = f()
t1 = time.time()
print 'Looping %d times took' % iters, t1 - t0, 'seconds'
print 'Result is', r
if numpy.any([isinstance(x.op, T.Elemwise) for x in
f.maker.fgraph.toposort()]):
print 'Used the cpu'
else:
print 'Used the gpu' C:\ProgramData\Anaconda2\python.exe
C:/Users/Administrator/Desktop/work/ntm-one-shot-master/tt.py
WARNING (theano.sandbox.cuda): The cuda backend is deprecated and will be
removed in the next release (v0.10). Please switch to the gpuarray backend.
You can get more information about how to switch at this URL:
https://github.com/Theano/Theano/wiki/Converting-to-the-new-gpu-back-end%28gpuarray%29
Using gpu device 0: GeForce GTX 770M (CNMeM is disabled, cuDNN not available)
1 #include <Python.h>
2 #include <iostream>
3 #include "theano_mod_helper.h"
4 #include "cuda_ndarray.cuh"
5 //////////////////////
6 //// Support Code
7 //////////////////////
8
9 static __global__ void
kernel_reduce_ccontig_node_544270fe7a21a748315f83abfe0913cc_0(
10 const unsigned int d0,
11 const float *A,
12 float * Z)
13 {
14 const int threadCount = blockDim.x;
15 const int threadNum = threadIdx.x;
16 extern __shared__ float buf[];
17 float myresult = 0;
18
19 if (warpSize != 32)
20 {
21 return; //TODO: set error code
22 }
23
24 for (int i0 = threadIdx.x; i0 < d0; i0 += blockDim.x)
25 {
26 myresult = myresult + A[i0];
27 }
28
29 __syncthreads(); // some kernel do multiple reduction.
30 buf[threadNum] = myresult;
31 __syncthreads();
32
33 // rest of function is handled by one warp
34 if (threadNum < warpSize)
35 {
36 //round up all the partial sums into the first `warpSize`
elements
37 for (int i = threadNum + warpSize; i < threadCount; i +=
warpSize)
38 {
39 myresult = myresult + buf[i];
40 }
41 buf[threadNum] = myresult;
42 /*Comment this optimization as it don't work on Fermi GPU.
43 TODO: find why it don't work or put the GPU compute capability into
the version
44 // no sync because only one warp is running
45 if(threadCount >32)
46 {buf[threadNum] = buf[threadNum] +
buf[threadNum+16];buf[threadNum] = buf[threadNum] +
buf[threadNum+8];buf[threadNum] = buf[threadNum] +
buf[threadNum+4];buf[threadNum] = buf[threadNum] +
buf[threadNum+2];buf[threadNum] = buf[threadNum] + buf[threadNum+1];
47 if (threadNum == 0)
48 {
49 Z[0] = buf[0];
50 }
51
52 }
53 else */
54 if (threadNum < 16)
55 {
56 //reduce so that threadNum 0 has the reduction of everything
57 if (threadNum + 16 < threadCount) buf[threadNum] =
buf[threadNum] + buf[threadNum+16];if (threadNum + 8 < threadCount)
buf[threadNum] = buf[threadNum] + buf[threadNum+8];if (threadNum + 4 <
threadCount) buf[threadNum] = buf[threadNum] + buf[threadNum+4];if (threadNum +
2 < threadCount) buf[threadNum] = buf[threadNum] + buf[threadNum+2];if
(threadNum + 1 < threadCount) buf[threadNum] = buf[threadNum] +
buf[threadNum+1];
58 if (threadNum == 0)
59 {
60 Z[0] = buf[0];
61 }
62 }
63 }
64
65 }
66
67
68 static __global__ void
kernel_reduce_1_node_544270fe7a21a748315f83abfe0913cc_0(
69 const unsigned int d0,
70 const float *A, const int sA0,
71 float * Z)
72 {
73 const int threadCount = blockDim.x;
74 const int threadNum = threadIdx.x;
75 extern __shared__ float buf[];
76 float myresult = 0;
77
78 if (warpSize != 32)
79 {
80 return; //TODO: set error code
81 }
82
83 for (int i0 = threadIdx.x; i0 < d0; i0 += blockDim.x)
84 {
85 myresult = myresult + A[i0 * sA0];
86 }
87
88 __syncthreads(); // some kernel do multiple reduction.
89 buf[threadNum] = myresult;
90 __syncthreads();
91
92 // rest of function is handled by one warp
93 if (threadNum < warpSize)
94 {
95 //round up all the partial sums into the first `warpSize`
elements
96 for (int i = threadNum + warpSize; i < threadCount; i +=
warpSize)
97 {
98 myresult = myresult + buf[i];
99 }
100 buf[threadNum] = myresult;
101 /*Comment this optimization as it don't work on Fermi GPU.
102 TODO: find why it don't work or put the GPU compute capability into
the version
103 // no sync because only one warp is running
104 if(threadCount >32)
105 {buf[threadNum] = buf[threadNum] +
buf[threadNum+16];buf[threadNum] = buf[threadNum] +
buf[threadNum+8];buf[threadNum] = buf[threadNum] +
buf[threadNum+4];buf[threadNum] = buf[threadNum] +
buf[threadNum+2];buf[threadNum] = buf[threadNum] + buf[threadNum+1];
106 if (threadNum == 0)
107 {
108 Z[0] = buf[0];
109 }
110
111 }
112 else */
113 if (threadNum < 16)
114 {
115 //reduce so that threadNum 0 has the reduction of everything
116 if (threadNum + 16 < threadCount) buf[threadNum] =
buf[threadNum] + buf[threadNum+16];if (threadNum + 8 < threadCount)
buf[threadNum] = buf[threadNum] + buf[threadNum+8];if (threadNum + 4 <
threadCount) buf[threadNum] = buf[threadNum] + buf[threadNum+4];if (threadNum +
2 < threadCount) buf[threadNum] = buf[threadNum] + buf[threadNum+2];if
(threadNum + 1 < threadCount) buf[threadNum] = buf[threadNum] +
buf[threadNum+1];
117 if (threadNum == 0)
118 {
119 Z[0] = buf[0];
120 }
121 }
122 }
123
124 }
125
126
127
128 namespace {
129 struct __struct_compiled_op_544270fe7a21a748315f83abfe0913cc {
130 PyObject* __ERROR;
131
132 PyObject* storage_V3;
133 PyObject* storage_V1;
134
135
136 __struct_compiled_op_544270fe7a21a748315f83abfe0913cc() {
137 // This is only somewhat safe because we:
138 // 1) Are not a virtual class
139 // 2) Do not use any virtual classes in the members
140 // 3) Deal with mostly POD and pointers
141
142 // If this changes, we would have to revise this, but for
143 // now I am tired of chasing segfaults because
144 // initialization code had an error and some pointer has
145 // a junk value.
146 memset(this, 0, sizeof(*this));
147 }
148 ~__struct_compiled_op_544270fe7a21a748315f83abfe0913cc(void) {
149 cleanup();
150 }
151
152 int init(PyObject* __ERROR, PyObject* storage_V3, PyObject*
storage_V1) {
153 Py_XINCREF(storage_V3);
154 Py_XINCREF(storage_V1);
155 this->storage_V3 = storage_V3;
156 this->storage_V1 = storage_V1;
157
158
159
160
161 this->__ERROR = __ERROR;
162 return 0;
163 }
164 void cleanup(void) {
165 __label_1:
166
167 double __DUMMY_1;
168 __label_3:
169
170 double __DUMMY_3;
171 __label_6:
172
173 double __DUMMY_6;
174
175 Py_XDECREF(this->storage_V3);
176 Py_XDECREF(this->storage_V1);
177 }
178 int run(void) {
179 int __failure = 0;
180
181 PyObject* py_V1;
182 CudaNdarray * V1;
183 PyObject* py_V3;
184 CudaNdarray * V3;
185 {
186
187 py_V1 = PyList_GET_ITEM(storage_V1, 0);
188 {Py_XINCREF(py_V1);}
189
190 if (py_V1 == Py_None)
191 {
192 V1 = NULL;
193 }
194 else
195 {
196
197 assert(py_V1->ob_refcnt >= 2); // There should be at least one ref
from the container object,
198 // and one ref from the local scope.
199
200 if (CudaNdarray_Check(py_V1))
201 {
202 //fprintf(stderr, "c_extract CNDA object w refcnt %p %i\n",
py_V1, (py_V1->ob_refcnt));
203 V1 = (CudaNdarray*)py_V1;
204 //std::cerr << "c_extract " << V1 << '\n';
205
206
207 if (V1->nd != 0)
208 {
209 PyErr_Format(PyExc_RuntimeError,
210 "c_extract: Some CudaNdarray has rank %i,
it was supposed to have rank 0",
211 V1->nd);
212 V1 = NULL;
213 {
214 __failure = 2;
215 if (!PyErr_Occurred()) {
216 PyErr_SetString(PyExc_RuntimeError,
217 "Unexpected error in an Op's C code. "
218 "No Python exception was set.");
219 }
220 goto __label_2;};
221 }
222 //std::cerr << "c_extract " << V1 << " nd check passed\n";
223
224
225 assert(V1);
226 Py_INCREF(py_V1);
227 }
228 else if (py_V1 == Py_None)
229 {
230 PyErr_SetString(PyExc_TypeError,
231 "expected a CudaNdarray, not None");
232 V1 = NULL;
233 {
234 __failure = 2;
235 if (!PyErr_Occurred()) {
236 PyErr_SetString(PyExc_RuntimeError,
237 "Unexpected error in an Op's C code. "
238 "No Python exception was set.");
239 }
240 goto __label_2;};
241 }
242 else
243 {
244 //fprintf(stderr, "FAILING c_extract CNDA object w refcnt
%p %i\n", py_V1, (py_V1->ob_refcnt));
245 PyErr_SetString(PyExc_TypeError, "Argument not a
CudaNdarray");
246 V1 = NULL;
247 {
248 __failure = 2;
249 if (!PyErr_Occurred()) {
250 PyErr_SetString(PyExc_RuntimeError,
251 "Unexpected error in an Op's C code. "
252 "No Python exception was set.");
253 }
254 goto __label_2;};
255 }
256 //std::cerr << "c_extract done " << V1 << '\n';
257
258
259 }
260
261 {
262
263 py_V3 = PyList_GET_ITEM(storage_V3, 0);
264 {Py_XINCREF(py_V3);}
265
266 assert(py_V3->ob_refcnt >= 2); // There should be at least one ref
from the container object,
267 // and one ref from the local scope.
268
269 if (CudaNdarray_Check(py_V3))
270 {
271 //fprintf(stderr, "c_extract CNDA object w refcnt %p %i\n",
py_V3, (py_V3->ob_refcnt));
272 V3 = (CudaNdarray*)py_V3;
273 //std::cerr << "c_extract " << V3 << '\n';
274
275
276 if (V3->nd != 1)
277 {
278 PyErr_Format(PyExc_RuntimeError,
279 "c_extract: Some CudaNdarray has rank %i,
it was supposed to have rank 1",
280 V3->nd);
281 V3 = NULL;
282 {
283 __failure = 4;
284 if (!PyErr_Occurred()) {
285 PyErr_SetString(PyExc_RuntimeError,
286 "Unexpected error in an Op's C code. "
287 "No Python exception was set.");
288 }
289 goto __label_4;};
290 }
291 //std::cerr << "c_extract " << V3 << " nd check passed\n";
292
293
294 assert(V3);
295 Py_INCREF(py_V3);
296 }
297 else if (py_V3 == Py_None)
298 {
299 PyErr_SetString(PyExc_TypeError,
300 "expected a CudaNdarray, not None");
301 V3 = NULL;
302 {
303 __failure = 4;
304 if (!PyErr_Occurred()) {
305 PyErr_SetString(PyExc_RuntimeError,
306 "Unexpected error in an Op's C code. "
307 "No Python exception was set.");
308 }
309 goto __label_4;};
310 }
311 else
312 {
313 //fprintf(stderr, "FAILING c_extract CNDA object w refcnt
%p %i\n", py_V3, (py_V3->ob_refcnt));
314 PyErr_SetString(PyExc_TypeError, "Argument not a
CudaNdarray");
315 V3 = NULL;
316 {
317 __failure = 4;
318 if (!PyErr_Occurred()) {
319 PyErr_SetString(PyExc_RuntimeError,
320 "Unexpected error in an Op's C code. "
321 "No Python exception was set.");
322 }
323 goto __label_4;};
324 }
325 //std::cerr << "c_extract done " << V3 << '\n';
326
327
328 {
329 // Op class GpuCAReduce
330
331 if (V3->nd != 1)
332 {
333 PyErr_Format(PyExc_TypeError,
334 "required nd=1, got nd=%i", V3->nd);
335 {
336 __failure = 5;
337 if (!PyErr_Occurred()) {
338 PyErr_SetString(PyExc_RuntimeError,
339 "Unexpected error in an Op's C code. "
340 "No Python exception was set.");
341 }
342 goto __label_5;};
343 }
344
345
346 if ( !V1
347 || (V1->nd != 0)
348
349
350 )
351 {
352
353 int *new_dims=NULL;
354
355 Py_XDECREF(V1);
356 V1 = (CudaNdarray*) CudaNdarray_NewDims(0, new_dims);
357 if (NULL == V1)
358 {
359 {
360 __failure = 5;
361 if (!PyErr_Occurred()) {
362 PyErr_SetString(PyExc_RuntimeError,
363 "Unexpected error in an Op's C code. "
364 "No Python exception was set.");
365 }
366 goto __label_5;};
367 }
368 }
369
370
371 if (CudaNdarray_SIZE(V1) && ! CudaNdarray_SIZE(V3)){
372 cudaMemset(V1->devdata, 0, CudaNdarray_SIZE(V1) *
sizeof(float));
373 }
374 else if (CudaNdarray_SIZE(V1))
375 {
376
377 if(CudaNdarray_is_c_contiguous( V3)){
378
379 {
380 if(CudaNdarray_SIZE(V3)==0){
381 cudaMemset(V1->devdata, 0, CudaNdarray_SIZE(V1) *
sizeof(float));
382 }else{
383 int verbose = 0;
384 dim3 n_threads(
385 std::min(CudaNdarray_SIZE(V3),
386 (size_t) NUM_VECTOR_OP_THREADS_PER_BLOCK));
387 dim3 n_blocks(1);
388 if (verbose) printf("running
kernel_reduce_ccontig_node_544270fe7a21a748315f83abfe0913cc_0"
389 " n_threads.x=%d, size=%d, ndim=%d\n",
390 n_threads.x,CudaNdarray_SIZE(V3),V3->nd);
391 int n_shared = sizeof(float) * n_threads.x;
392
kernel_reduce_ccontig_node_544270fe7a21a748315f83abfe0913cc_0<<<n_blocks,
n_threads, n_shared>>>(
393 CudaNdarray_SIZE(V3),
394 CudaNdarray_DEV_DATA(V3),
395 CudaNdarray_DEV_DATA(V1));
396 CNDA_THREAD_SYNC;
397 cudaError_t sts = cudaGetLastError();
398 if (cudaSuccess != sts)
399 {
400 PyErr_Format(PyExc_RuntimeError,
401 "Cuda error: %s: %s."
402 " (grid: %i x %i; block: %i x %i x %i)\n",
403
"kernel_reduce_ccontig_node_544270fe7a21a748315f83abfe0913cc_0",
404 cudaGetErrorString(sts),
405 n_blocks.x,
406 n_blocks.y,
407 n_threads.x,
408 n_threads.y,
409 n_threads.z);
410 {
411 __failure = 5;
412 if (!PyErr_Occurred()) {
413 PyErr_SetString(PyExc_RuntimeError,
414 "Unexpected error in an Op's C code. "
415 "No Python exception was set.");
416 }
417 goto __label_5;};
418 }
419 }
420 }
421
422 }else{
423
424 {
425 int verbose = 0;
426 dim3 n_threads(
427 std::min(CudaNdarray_HOST_DIMS(V3)[0],
428 NUM_VECTOR_OP_THREADS_PER_BLOCK));
429 dim3 n_blocks(1);
430
431 if (verbose)
432 printf("running
kernel_reduce_1_node_544270fe7a21a748315f83abfe0913cc_0\n");
433 int n_shared = sizeof(float) * n_threads.x * n_threads.y *
n_threads.z;
434 if (verbose>1)
435 printf("n_threads.x=%d, n_threads.y=%d, n_threads.z=%d,"
436 " nb_threads=%d, n_blocks.x=%d, n_blocks.y=%d,"
437 " nb_block=%d, n_shared=%d, shape=(%d)\n",
438 n_threads.x,n_threads.y,n_threads.z,
439 n_threads.x*n_threads.y*n_threads.z,
440 n_blocks.x,n_blocks.y,
441 n_blocks.x*n_blocks.y, n_shared,
CudaNdarray_HOST_DIMS(V3)[0]);
442
kernel_reduce_1_node_544270fe7a21a748315f83abfe0913cc_0<<<n_blocks, n_threads,
n_shared>>>(
443
444
445 CudaNdarray_HOST_DIMS(V3)[0],
446
447
448 CudaNdarray_DEV_DATA(V3)
449
450
451 ,CudaNdarray_HOST_STRIDES(V3)[0]
452
453
454 ,CudaNdarray_DEV_DATA(V1)
455
456
457 );
458 CNDA_THREAD_SYNC;
459 cudaError_t sts = cudaGetLastError();
460 if (cudaSuccess != sts)
461 {
462 PyErr_Format(PyExc_RuntimeError,
463 "Cuda error: %s: %s."
464 " (grid: %i x %i; block: %i x %i x %i)"
465 " shape=(%d) \n",
466
"kernel_reduce_1_node_544270fe7a21a748315f83abfe0913cc_0",
467 cudaGetErrorString(sts),
468 n_blocks.x,
469 n_blocks.y,
470 n_threads.x,
471 n_threads.y,
472 n_threads.z,
473 CudaNdarray_HOST_DIMS(V3)[0]);
474 {
475 __failure = 5;
476 if (!PyErr_Occurred()) {
477 PyErr_SetString(PyExc_RuntimeError,
478 "Unexpected error in an Op's C code. "
479 "No Python exception was set.");
480 }
481 goto __label_5;};
482 }
483
484
485 }
486
487 }
488
489 }
490
491 __label_5:
492
493 double __DUMMY_5;
494
495 }
496 __label_4:
497
498 //std::cerr << "cleanup " << py_V3 << " " << V3 << "\n";
499 //fprintf(stderr, "c_cleanup CNDA py_object w refcnt %p %i\n",
py_V3, (py_V3->ob_refcnt));
500 if (V3)
501 {
502 //fprintf(stderr, "c_cleanup CNDA cn_object w refcnt %p %i\n",
V3, (V3->ob_refcnt));
503 Py_XDECREF(V3);
504 }
505 //std::cerr << "cleanup done" << py_V3 << "\n";
506
507 {Py_XDECREF(py_V3);}
508
509 double __DUMMY_4;
510
511 }
512 __label_2:
513
514 if (!__failure) {
515
516 //std::cerr << "sync\n";
517 if (NULL == V1) {
518 // failure: sync None to storage
519 Py_XDECREF(py_V1);
520 py_V1 = Py_None;
521 Py_INCREF(py_V1);
522 }
523 else
524 {
525 if (py_V1 != (PyObject*)V1)
526 {
527 Py_XDECREF(py_V1);
528 py_V1 = (PyObject*)V1;
529 Py_INCREF(py_V1);
530 }
531 assert(py_V1->ob_refcnt);
532 }
533
534 PyObject* old = PyList_GET_ITEM(storage_V1, 0);
535 {Py_XINCREF(py_V1);}
536 PyList_SET_ITEM(storage_V1, 0, py_V1);
537 {Py_XDECREF(old);}
538 }
539
540 //std::cerr << "cleanup " << py_V1 << " " << V1 << "\n";
541 //fprintf(stderr, "c_cleanup CNDA py_object w refcnt %p %i\n",
py_V1, (py_V1->ob_refcnt));
542 if (V1)
543 {
544 //fprintf(stderr, "c_cleanup CNDA cn_object w refcnt %p %i\n",
V1, (V1->ob_refcnt));
545 Py_XDECREF(V1);
546 }
547 //std::cerr << "cleanup done" << py_V1 << "\n";
548
549 {Py_XDECREF(py_V1);}
550
551 double __DUMMY_2;
552
553 }
554
555
556 if (__failure) {
557 // When there is a failure, this code puts the exception
558 // in __ERROR.
559 PyObject* err_type = NULL;
560 PyObject* err_msg = NULL;
561 PyObject* err_traceback = NULL;
562 PyErr_Fetch(&err_type, &err_msg, &err_traceback);
563 if (!err_type) {err_type = Py_None;Py_INCREF(Py_None);}
564 if (!err_msg) {err_msg = Py_None; Py_INCREF(Py_None);}
565 if (!err_traceback) {err_traceback = Py_None;
Py_INCREF(Py_None);}
566 PyObject* old_err_type = PyList_GET_ITEM(__ERROR, 0);
567 PyObject* old_err_msg = PyList_GET_ITEM(__ERROR, 1);
568 PyObject* old_err_traceback = PyList_GET_ITEM(__ERROR, 2);
569 PyList_SET_ITEM(__ERROR, 0, err_type);
570 PyList_SET_ITEM(__ERROR, 1, err_msg);
571 PyList_SET_ITEM(__ERROR, 2, err_traceback);
572 {Py_XDECREF(old_err_type);}
573 {Py_XDECREF(old_err_msg);}
574 {Py_XDECREF(old_err_traceback);}
575 }
576 // The failure code is returned to index what code block failed.
577 return __failure;
578
579 }
580 };
581 }
582
583
584 static int
__struct_compiled_op_544270fe7a21a748315f83abfe0913cc_executor(__struct_compiled_op_544270fe7a21a748315f83abfe0913cc*
self) {
585 return self->run();
586 }
587
588 static void
__struct_compiled_op_544270fe7a21a748315f83abfe0913cc_destructor(void*
executor, void* self) {
589 delete
((__struct_compiled_op_544270fe7a21a748315f83abfe0913cc*)self);
590 }
591
592 //////////////////////
593 //// Functions
594 //////////////////////
595 static PyObject * instantiate(PyObject * self, PyObject *argtuple) {
596 assert(PyTuple_Check(argtuple));
597 if (3 != PyTuple_Size(argtuple)){
598 PyErr_Format(PyExc_TypeError, "Wrong number of arguments, expected 3,
got %i", (int)PyTuple_Size(argtuple));
599 return NULL;
600 }
601 __struct_compiled_op_544270fe7a21a748315f83abfe0913cc* struct_ptr = new
__struct_compiled_op_544270fe7a21a748315f83abfe0913cc();
602 if (struct_ptr->init( PyTuple_GET_ITEM(argtuple,
0),PyTuple_GET_ITEM(argtuple, 1),PyTuple_GET_ITEM(argtuple, 2) ) != 0) {
603 delete struct_ptr;
604 return NULL;
605 }
606 PyObject* thunk =
PyCObject_FromVoidPtrAndDesc((void*)(&__struct_compiled_op_544270fe7a21a748315f83abfe0913cc_executor),
struct_ptr, __struct_compiled_op_544270fe7a21a748315f83abfe0913cc_destructor);
607 return thunk; }
608
609 //////////////////////
610 //// Module init
611 //////////////////////
612 static PyMethodDef MyMethods[] = {
613 {"instantiate", instantiate, METH_VARARGS, "undocumented"} ,
614 {NULL, NULL, 0, NULL}
615 };
616 PyMODINIT_FUNC init544270fe7a21a748315f83abfe0913cc(void){
617 (void) Py_InitModule("544270fe7a21a748315f83abfe0913cc", MyMethods);
618 }
619
===============================
c:\program files\nvidia gpu computing
toolkit\cuda\v8.0\include\math_functions.h(1): error: expected a declaration
c:\program files\nvidia gpu computing
toolkit\cuda\v8.0\include\cuda_surface_types.h(91): warning: parsing restarts
here after previous syntax error
c:\program files\nvidia gpu computing
toolkit\cuda\v8.0\include\cuda_surface_types.h(94): error: surface is not a
template
c:\program files\nvidia gpu computing
toolkit\cuda\v8.0\include\cuda_surface_types.h(97): error: explicit type is
missing ("int" assumed)
c:\program files\nvidia gpu computing
toolkit\cuda\v8.0\include\device_functions.h(1): error: expected a declaration
C:\Program Files\NVIDIA GPU Computing
Toolkit\CUDA\v8.0\include\cuda_runtime.h(1420): warning: parsing restarts here
after previous syntax error
C:\Program Files\NVIDIA GPU Computing
Toolkit\CUDA\v8.0\include\cuda_runtime.h(1882): error: surface is not a template
C:\Program Files\NVIDIA GPU Computing
Toolkit\CUDA\v8.0\include\cuda_runtime.h(1911): error: surface is not a template
C:\ProgramData\Anaconda2\lib\site-packages\theano\sandbox\cuda\cuda_ndarray.cuh(331):
error: identifier "log2" is undefined
mod.cu(14): error: identifier "blockDim" is undefined
mod.cu(15): error: identifier "threadIdx" is undefined
mod.cu(19): error: identifier "warpSize" is undefined
mod.cu(29): error: identifier "__syncthreads" is undefined
mod.cu(73): error: identifier "blockDim" is undefined
mod.cu(74): error: identifier "threadIdx" is undefined
mod.cu(78): error: identifier "warpSize" is undefined
mod.cu(88): error: identifier "__syncthreads" is undefined
15 errors detected in the compilation of
"C:/Users/ADMINI~1/AppData/Local/Temp/tmpxft_000054a8_00000000-10_mod.cpp1.ii".
Traceback (most recent call last):
File "C:/Users/Administrator/Desktop/work/ntm-one-shot-master/tt.py", line 1,
in <module>
from theano import function, config, shared, sandbox
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\__init__.py", line
116, in <module>
theano.sandbox.cuda.tests.test_driver.test_nvidia_driver1()
File
"C:\ProgramData\Anaconda2\lib\site-packages\theano\sandbox\cuda\tests\test_driver.py",
line 32, in test_nvidia_driver1
profile=False)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\compile\function.py",
line 326, in function
output_keys=output_keys)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\compile\pfunc.py",
line 486, in pfunc
output_keys=output_keys)
File
"C:\ProgramData\Anaconda2\lib\site-packages\theano\compile\function_module.py",
line 1808, in orig_function
defaults)
File
"C:\ProgramData\Anaconda2\lib\site-packages\theano\compile\function_module.py",
line 1674, in create
nvcc warning : nvcc support for Microsoft Visual Studio 2010 and earlier has
been deprecated and is no longer being maintained
mod.cu
['nvcc', '-shared', '-O3', '-arch=sm_30', '-Xlinker', '/DEBUG', '-D
HAVE_ROUND', '-m64', '-Xcompiler',
'-DCUDA_NDARRAY_CUH=18715462c72ed6afcd7ca5d52813ce90,-DNPY_NO_DEPRECATED_API=NPY_1_7_API_VERSION,/Zi,/MD',
'-I"C:\\Users\\Administrator\\AppData\\Local\\Theano\\compiledir_Windows-7-6.1.7601-SP1-Intel64_Family_6_Model_60_Stepping_3_GenuineIntel-2.7.13-64\\cuda_ndarray"',
'-I"C:\\ProgramData\\Anaconda2\\lib\\site-packages\\numpy\\core\\include"',
'-I"C:\\ProgramData\\Anaconda2\\include"',
'-I"C:\\ProgramData\\Anaconda2\\lib\\site-packages\\theano\\gof"',
'-I"C:\\ProgramData\\Anaconda2\\lib\\site-packages\\theano\\sandbox\\cuda"',
'-L"C:\\Users\\Administrator\\AppData\\Local\\Theano\\compiledir_Windows-7-6.1.7601-SP1-Intel64_Family_6_Model_60_Stepping_3_GenuineIntel-2.7.13-64\\cuda_ndarray"',
'-L"C:\\ProgramData\\Anaconda2\\libs"', '-L"C:\\ProgramData\\Anaconda2"',
'-o',
'C:\\Users\\Administrator\\AppData\\Local\\Theano\\compiledir_Windows-7-6.1.7601-SP1-Intel64_Family_6_Model_60_Stepping_3_GenuineIntel-2.7.13-64\\tmpmenaes\\544270fe7a21a748315f83abfe0913cc.pyd',
'mod.cu', '-lcudart', '-lcublas', '-lcuda_ndarray', '-lpython27']
input_storage=input_storage_lists, storage_map=storage_map)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\link.py", line
699, in make_thunk
storage_map=storage_map)[:3]
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\vm.py", line
1047, in make_all
impl=impl))
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\op.py", line 935,
in make_thunk
no_recycling)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\op.py", line 839,
in make_c_thunk
output_storage=node_output_storage)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\cc.py", line
1190, in make_thunk
keep_lock=keep_lock)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\cc.py", line
1131, in __compile__
keep_lock=keep_lock)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\cc.py", line
1586, in cthunk_factory
key=key, lnk=self, keep_lock=keep_lock)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\cmodule.py", line
1159, in module_from_key
module = lnk.compile_cmodule(location)
File "C:\ProgramData\Anaconda2\lib\site-packages\theano\gof\cc.py", line
1489, in compile_cmodule
preargs=preargs)
File
"C:\ProgramData\Anaconda2\lib\site-packages\theano\sandbox\cuda\nvcc_compiler.py",
line 405, in compile_str
'for cmd', ' '.join(cmd))
Exception: ('The following error happened while compiling the node',
GpuCAReduce{add}{1}(<CudaNdarrayType(float32, vector)>), '\n', 'nvcc return
status', 2, 'for cmd', 'nvcc -shared -O3 -arch=sm_30 -Xlinker /DEBUG -D
HAVE_ROUND -m64 -Xcompiler
-DCUDA_NDARRAY_CUH=18715462c72ed6afcd7ca5d52813ce90,-DNPY_NO_DEPRECATED_API=NPY_1_7_API_VERSION,/Zi,/MD
-I"C:\\Users\\Administrator\\AppData\\Local\\Theano\\compiledir_Windows-7-6.1.7601-SP1-Intel64_Family_6_Model_60_Stepping_3_GenuineIntel-2.7.13-64\\cuda_ndarray"
-I"C:\\ProgramData\\Anaconda2\\lib\\site-packages\\numpy\\core\\include"
-I"C:\\ProgramData\\Anaconda2\\include"
-I"C:\\ProgramData\\Anaconda2\\lib\\site-packages\\theano\\gof"
-I"C:\\ProgramData\\Anaconda2\\lib\\site-packages\\theano\\sandbox\\cuda"
-L"C:\\Users\\Administrator\\AppData\\Local\\Theano\\compiledir_Windows-7-6.1.7601-SP1-Intel64_Family_6_Model_60_Stepping_3_GenuineIntel-2.7.13-64\\cuda_ndarray"
-L"C:\\ProgramData\\Anaconda2\\libs" -L"C:\\ProgramData\\Anaconda2" -o
C:\\Users\\Administrator\\AppData\\Local\\Theano\\compiledir_Windows-7-6.1.7601-SP1-Intel64_Family_6_Model_60_Stepping_3_GenuineIntel-2.7.13-64\\tmpmenaes\\544270fe7a21a748315f83abfe0913cc.pyd
mod.cu -lcudart -lcublas -lcuda_ndarray -lpython27',
'[GpuCAReduce{add}{1}(<CudaNdarrayType(float32, vector)>)]')
Process finished with exit code 1
_______________________________________________
PyCUDA mailing list
PyCUDA@tiker.net
https://lists.tiker.net/listinfo/pycuda
