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
|
#ifndef GSR_CUDA_H
#define GSR_CUDA_H
#include "overclock.h"
#include <stddef.h>
#include <stdbool.h>
// To prevent hwcontext_cuda.h from including cuda.h
#define CUDA_VERSION 11070
#define CU_CTX_SCHED_AUTO 0
#if defined(_WIN64) || defined(__LP64__)
typedef unsigned long long CUdeviceptr_v2;
#else
typedef unsigned int CUdeviceptr_v2;
#endif
typedef CUdeviceptr_v2 CUdeviceptr;
typedef int CUresult;
typedef int CUdevice_v1;
typedef CUdevice_v1 CUdevice;
typedef struct CUctx_st *CUcontext;
typedef struct CUstream_st *CUstream;
typedef struct CUarray_st *CUarray;
#define CUDA_SUCCESS 0
typedef enum CUgraphicsMapResourceFlags_enum {
CU_GRAPHICS_MAP_RESOURCE_FLAGS_NONE = 0x00,
CU_GRAPHICS_MAP_RESOURCE_FLAGS_READ_ONLY = 0x01,
CU_GRAPHICS_MAP_RESOURCE_FLAGS_WRITE_DISCARD = 0x02
} CUgraphicsMapResourceFlags;
typedef enum CUgraphicsRegisterFlags_enum {
CU_GRAPHICS_REGISTER_FLAGS_NONE = 0x00,
CU_GRAPHICS_REGISTER_FLAGS_READ_ONLY = 0x01,
CU_GRAPHICS_REGISTER_FLAGS_WRITE_DISCARD = 0x02,
CU_GRAPHICS_REGISTER_FLAGS_SURFACE_LDST = 0x04,
CU_GRAPHICS_REGISTER_FLAGS_TEXTURE_GATHER = 0x08
} CUgraphicsRegisterFlags;
typedef enum CUmemorytype_enum {
CU_MEMORYTYPE_HOST = 0x01, /**< Host memory */
CU_MEMORYTYPE_DEVICE = 0x02, /**< Device memory */
CU_MEMORYTYPE_ARRAY = 0x03, /**< Array memory */
CU_MEMORYTYPE_UNIFIED = 0x04 /**< Unified device or host memory */
} CUmemorytype;
typedef struct CUDA_MEMCPY2D_st {
size_t srcXInBytes; /**< Source X in bytes */
size_t srcY; /**< Source Y */
CUmemorytype srcMemoryType; /**< Source memory type (host, device, array) */
const void *srcHost; /**< Source host pointer */
CUdeviceptr srcDevice; /**< Source device pointer */
CUarray srcArray; /**< Source array reference */
size_t srcPitch; /**< Source pitch (ignored when src is array) */
size_t dstXInBytes; /**< Destination X in bytes */
size_t dstY; /**< Destination Y */
CUmemorytype dstMemoryType; /**< Destination memory type (host, device, array) */
void *dstHost; /**< Destination host pointer */
CUdeviceptr dstDevice; /**< Destination device pointer */
CUarray dstArray; /**< Destination array reference */
size_t dstPitch; /**< Destination pitch (ignored when dst is array) */
size_t WidthInBytes; /**< Width of 2D memory copy in bytes */
size_t Height; /**< Height of 2D memory copy */
} CUDA_MEMCPY2D_v2;
typedef CUDA_MEMCPY2D_v2 CUDA_MEMCPY2D;
typedef struct CUgraphicsResource_st *CUgraphicsResource;
typedef struct gsr_cuda gsr_cuda;
struct gsr_cuda {
gsr_overclock overclock;
bool do_overclock;
void *library;
CUcontext cu_ctx;
CUresult (*cuInit)(unsigned int Flags);
CUresult (*cuDeviceGetCount)(int *count);
CUresult (*cuDeviceGet)(CUdevice *device, int ordinal);
CUresult (*cuCtxCreate_v2)(CUcontext *pctx, unsigned int flags, CUdevice dev);
CUresult (*cuCtxDestroy_v2)(CUcontext ctx);
CUresult (*cuCtxPushCurrent_v2)(CUcontext ctx);
CUresult (*cuCtxPopCurrent_v2)(CUcontext *pctx);
CUresult (*cuGetErrorString)(CUresult error, const char **pStr);
CUresult (*cuMemcpy2D_v2)(const CUDA_MEMCPY2D *pCopy);
CUresult (*cuMemcpy2DAsync_v2)(const CUDA_MEMCPY2D *pcopy, CUstream hStream);
CUresult (*cuStreamSynchronize)(CUstream hStream);
CUresult (*cuGraphicsGLRegisterImage)(CUgraphicsResource *pCudaResource, unsigned int image, unsigned int target, unsigned int flags);
CUresult (*cuGraphicsEGLRegisterImage)(CUgraphicsResource *pCudaResource, void *image, unsigned int flags);
CUresult (*cuGraphicsResourceSetMapFlags)(CUgraphicsResource resource, unsigned int flags);
CUresult (*cuGraphicsMapResources)(unsigned int count, CUgraphicsResource *resources, CUstream hStream);
CUresult (*cuGraphicsUnmapResources)(unsigned int count, CUgraphicsResource *resources, CUstream hStream);
CUresult (*cuGraphicsUnregisterResource)(CUgraphicsResource resource);
CUresult (*cuGraphicsSubResourceGetMappedArray)(CUarray *pArray, CUgraphicsResource resource, unsigned int arrayIndex, unsigned int mipLevel);
};
bool gsr_cuda_load(gsr_cuda *self, Display *display, bool overclock);
void gsr_cuda_unload(gsr_cuda *self);
#endif /* GSR_CUDA_H */
|
