-
Notifications
You must be signed in to change notification settings - Fork 0
/
cuStack.cu
307 lines (245 loc) · 8.16 KB
/
cuStack.cu
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
/*---------------------------------------------------.
| STACK and Memory Management Functions |
`---------------------------------------------------*/
/*-------------------------------------.
| How to Test Stack Implementaion |
`-------------------------------------*/
/*
__global__ void myStackTesting( byte* buffer )
{
int i;
int stackID;
stackID = allocateStack();
PUSH( stackID , (byte)'3' );
PUSH( stackID , (byte)'o' );
PUSH( stackID , (byte)'m' );
PUSH( stackID , (byte)'a' );
PUSH( stackID , (byte)'r' );
PUSH( stackID , (byte)'z' );
//PUSH( stack , sp , 0 , '\0' );
buffer[0] = 0;
for(i = 0; i < 6; i++)
{
buffer[i] = POP( stackID );
}
buffer[6] = ' ';
buffer[7] = 'I';
buffer[8] = 'D';
buffer[9] = ' ';
buffer[10] = '=';
buffer[11] = ' ';
buffer[12] = stackID + '0';
buffer[13] = 0;
deallocateStack( stackID );
}
*/
__device__ int stackMainLock = 0;
__device__ void superCudaMemcpy( byte *destination , byte *source , unsigned int len)
{
unsigned int counter;
for(counter = 0; counter < len; counter++)
destination[counter] = source[counter];
}
__device__ byte PUSH( int segment , uint16 DATA )
{
// if segment is not in range
if( segment < 0 || segment >= mainStack.segments ) return PUSH_FAIL;
// precaution for overflow
if(mainStack.sp[segment] - mainStack.lowerLimit[segment] < sizeof(DATA))
return PUSH_FAIL;
//Push a value on the stack for the respectful element.
mainStack.sp[segment] = mainStack.sp[segment] - sizeof(DATA);
*((uint16*)mainStack.sp[segment]) = DATA;
return PUSH_SUCCESS;
}
__device__ byte PUSH( int segment , unsigned int DATA )
{
if( segment == STACK_ID_INVALID ) return PUSH_FAIL;
//Push a value on the stack for the respectful element.
mainStack.sp[segment] = mainStack.sp[segment] - sizeof(DATA);
*((unsigned int*)mainStack.sp[segment]) = DATA;
return PUSH_SUCCESS;
}
__device__ byte PUSH( int segment , byte DATA )
{
// if segment is not in range
if( segment < 0 || segment >= mainStack.segments ) return PUSH_FAIL;
//Push a value on the stack for the respectful element.
mainStack.sp[segment]--;
*(mainStack.sp[segment]) = DATA;
return PUSH_SUCCESS;
}
__device__ byte POP( int segment )
{
//Pop a value from the stack and return it.
byte DATA;
// if segment is not in range
if( segment < 0 || segment >= mainStack.segments ) return 0;
DATA = *(mainStack.sp[segment]);
mainStack.sp[segment]++;
return DATA;
}
__device__ unsigned short POPshort( int segment )
{
//Pop a value from the stack and return it.
unsigned short DATA;
// if segment is not in range
if( segment < 0 || segment >= mainStack.segments ) return 0;
DATA = *((unsigned short*)(mainStack.sp[segment]));
mainStack.sp[segment] = mainStack.sp[segment] + sizeof(DATA);
return DATA;
}
__device__ unsigned int POPint( int segment )
{
// Pop a value from the stack and return it.
unsigned int DATA;
// if segment is not in range
if( segment < 0 || segment >= mainStack.segments ) return 0;
DATA = *((unsigned int*)mainStack.sp[segment]);
mainStack.sp[segment] = mainStack.sp[segment] + sizeof(DATA);
return DATA;
}
__device__ byte PEEK( int segment )
{
// Peek at the top byte of the stack.
byte DATA;
// if segment is not in range
if( segment < 0 || segment >= mainStack.segments ) return 0;
DATA = *(mainStack.sp[segment]);
return DATA;
}
__global__ void initializeDeviceStack( stack mStack )
{
int i;
superCudaMemcpy( (byte*)&mainStack , (byte*)&mStack , sizeof(stack) );
for(i = 0; i < mainStack.segments ; i++)
{
mainStack.sp[i] = (byte*)(mainStack.buffer + (mainStack.chunk * (i + 1)));
mainStack.lowerLimit[i] = (byte*)(mainStack.buffer + (mainStack.chunk * i));
mainStack.upperLimit[i] = (byte*)(mainStack.buffer + (mainStack.chunk * (i + 1)));
mainStack.allocated[i] = STACK_SEGMENT_UNALLCOATED;
}
}
__global__ void deinitializeDeviceStack()
{
int i;
mainStack.buffer = 0;
for(i = 0; i < mainStack.segments ; i++)
{
mainStack.sp[i] = 0;
mainStack.lowerLimit[i] = 0;
mainStack.upperLimit[i] = 0;
mainStack.allocated[i] = STACK_SEGMENT_UNALLCOATED;
}
}
__device__ int allocateStack()
{
int i = 0;
while( atomicCAS(&stackMainLock, 0, 1) != 0 );
for(i = 0; i < mainStack.segments; i++)
{
if(mainStack.allocated[i] == STACK_SEGMENT_UNALLCOATED)
{
mainStack.allocated[i] = STACK_SEGMENT_ALLCOATED;
mainStack.sp[i] = mainStack.upperLimit[i];
while( atomicCAS(&stackMainLock, 1, 0) != 1 );
return i;
}
}
while( atomicCAS(&stackMainLock, 1, 0) != 1 );
#if DEV_DEBUG_STACK
cuPrintf("NO STACK SPACE AVAILABLE FOR ME... I WILL DIE !!\n");
#endif
return STACK_ID_INVALID;
}
__device__ int allocateStack( int neededID , byte keepData )
{
int i = 0;
// if segment is not in range
if( neededID < 0 || neededID >= mainStack.segments ) return STACK_ID_INVALID;
for(i = 0; i < mainStack.segments; i++)
{
if(mainStack.allocated[i] == STACK_SEGMENT_UNALLCOATED)
{
mainStack.allocated[i] = STACK_SEGMENT_ALLCOATED;
if(keepData != 0)
mainStack.sp[i] = mainStack.upperLimit[i];
return i;
}
}
return STACK_ID_INVALID;
}
__device__ byte copyStack( int destination , int source )
{
// if segment is not in range
if( destination < 0 || destination >= mainStack.segments\
|| source < 0 || source >= mainStack.segments )\
return COPYSTACK_FAIL;
superCudaMemcpy( (byte*)mainStack.lowerLimit[destination] , (byte*)mainStack.lowerLimit[source] , mainStack.chunk );
mainStack.sp[destination] = mainStack.upperLimit[destination] - (mainStack.upperLimit[source] - mainStack.sp[source]);
return COPYSTACK_SUCCESS;
}
__device__ void deallocateStack( int stackID )
{
// if segment is not in range
if( stackID < 0 || stackID >= mainStack.segments ) return;
mainStack.allocated[stackID] = STACK_SEGMENT_UNALLCOATED;
#if DEV_DEBUG_STACK
//cuPrintf("DEALLOCATED ID %d\n", stackID);
#endif
}
void deinitializeStack()
{
if( mStack.buffer != 0 )
{
HANDLE_FREE( cudaFree, mStack.buffer );
HANDLE_FREE( cudaFree, (byte*)mStack.sp );
HANDLE_FREE( cudaFree, (byte*)mStack.lowerLimit );
HANDLE_FREE( cudaFree, (byte*)mStack.upperLimit );
HANDLE_FREE( cudaFree, mStack.allocated );
}
deinitializeDeviceStack<<<1,1>>>();
cudaThreadSynchronize();
}
void initializeStack( int chunk , int segments )
{
// it is necessary to make sure these are freed if we're going to \
call initializeStack more than once in the code
if( mStack.buffer != 0 )
{
HANDLE_FREE( cudaFree, mStack.buffer );
HANDLE_FREE( cudaFree, (byte*)mStack.sp );
HANDLE_FREE( cudaFree, (byte*)mStack.lowerLimit );
HANDLE_FREE( cudaFree, (byte*)mStack.upperLimit );
HANDLE_FREE( cudaFree, mStack.allocated );
}
//Allocate Stack Space
cudaMalloc( (void **)&(mStack.buffer) , chunk * segments );
cudaMalloc( (void **)&(mStack.sp) , segments * sizeof(byte*) );
cudaMalloc( (void **)&(mStack.lowerLimit) , segments * sizeof(byte*) );
cudaMalloc( (void **)&(mStack.upperLimit) , segments * sizeof(byte*) );
cudaMalloc( (void **)&(mStack.allocated) , segments * sizeof(byte) );
mStack.chunk = chunk;
mStack.segments = segments;
initializeDeviceStack<<<1,1>>>( mStack );
cudaThreadSynchronize();
}
void initializeStack( stack defaultStack , int chunk , int segments )
{
if( defaultStack.buffer != 0 )
{
HANDLE_FREE( cudaFree, defaultStack.buffer );
HANDLE_FREE( cudaFree, (byte*)defaultStack.sp );
HANDLE_FREE( cudaFree, (byte*)defaultStack.lowerLimit );
HANDLE_FREE( cudaFree, (byte*)defaultStack.upperLimit );
}
//Allocate Stack Space
cudaMalloc( (void **)&(defaultStack.buffer) , chunk * segments );
cudaMalloc( (void **)&(defaultStack.sp) , segments * sizeof(byte*) );
cudaMalloc( (void **)&(defaultStack.lowerLimit) , segments * sizeof(byte*) );
cudaMalloc( (void **)&(defaultStack.upperLimit) , segments * sizeof(byte*) );
defaultStack.chunk = chunk;
defaultStack.segments = segments;
initializeDeviceStack<<<1,1>>>( defaultStack );
cudaThreadSynchronize();
}