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
|
// This code contains NVIDIA Confidential Information and is disclosed to you
// under a form of NVIDIA software license agreement provided separately to you.
//
// Notice
// NVIDIA Corporation and its licensors retain all intellectual property and
// proprietary rights in and to this software and related documentation and
// any modifications thereto. Any use, reproduction, disclosure, or
// distribution of this software and related documentation without an express
// license agreement from NVIDIA Corporation is strictly prohibited.
//
// ALL NVIDIA DESIGN SPECIFICATIONS, CODE ARE PROVIDED "AS IS.". NVIDIA MAKES
// NO WARRANTIES, EXPRESSED, IMPLIED, STATUTORY, OR OTHERWISE WITH RESPECT TO
// THE MATERIALS, AND EXPRESSLY DISCLAIMS ALL IMPLIED WARRANTIES OF NONINFRINGEMENT,
// MERCHANTABILITY, AND FITNESS FOR A PARTICULAR PURPOSE.
//
// Information and code furnished is believed to be accurate and reliable.
// However, NVIDIA Corporation assumes no responsibility for the consequences of use of such
// information or for any infringement of patents or other rights of third parties that may
// result from its use. No license is granted by implication or otherwise under any patent
// or patent rights of NVIDIA Corporation. Details are subject to change without notice.
// This code supersedes and replaces all information previously supplied.
// NVIDIA Corporation products are not authorized for use as critical
// components in life support devices or systems without express written approval of
// NVIDIA Corporation.
//
// Copyright (c) 2008-2016 NVIDIA Corporation. All rights reserved.
#ifndef PXCUDACONTEXTMANAGER_PXGPUCOPYDESCQUEUE_H
#define PXCUDACONTEXTMANAGER_PXGPUCOPYDESCQUEUE_H
#include "foundation/PxPreprocessor.h"
#if PX_SUPPORT_GPU_PHYSX
#include "foundation/PxAssert.h"
#include "task/PxTaskDefine.h"
#include "task/PxGpuDispatcher.h"
#include "cudamanager/PxGpuCopyDesc.h"
#include "cudamanager/PxCudaContextManager.h"
/* forward decl to avoid including <cuda.h> */
typedef struct CUstream_st* CUstream;
namespace physx
{
PX_PUSH_PACK_DEFAULT
/// \brief Container class for queueing PxGpuCopyDesc instances in pinned (non-pageable) CPU memory
class PxGpuCopyDescQueue
{
public:
/// \brief PxGpuCopyDescQueue constructor
PxGpuCopyDescQueue(PxGpuDispatcher& d)
: mDispatcher(d)
, mBuffer(0)
, mStream(0)
, mReserved(0)
, mOccupancy(0)
, mFlushed(0)
{
}
/// \brief PxGpuCopyDescQueue destructor
~PxGpuCopyDescQueue()
{
if (mBuffer)
{
mDispatcher.getCudaContextManager()->getMemoryManager()->free(PxCudaBufferMemorySpace::T_PINNED_HOST, (size_t) mBuffer);
}
}
/// \brief Reset the enqueued copy descriptor list
///
/// Must be called at least once before any copies are enqueued, and each time the launched
/// copies are known to have been completed. The recommended use case is to call this at the
/// start of each simulation step.
void reset(CUstream stream, uint32_t reserveSize)
{
if (reserveSize > mReserved)
{
if (mBuffer)
{
mDispatcher.getCudaContextManager()->getMemoryManager()->free(
PxCudaBufferMemorySpace::T_PINNED_HOST,
(size_t) mBuffer);
mReserved = 0;
}
mBuffer = (PxGpuCopyDesc*) mDispatcher.getCudaContextManager()->getMemoryManager()->alloc(
PxCudaBufferMemorySpace::T_PINNED_HOST,
reserveSize * sizeof(PxGpuCopyDesc),
PX_ALLOC_INFO("PxGpuCopyDescQueue", GPU_UTIL));
if (mBuffer)
{
mReserved = reserveSize;
}
}
mOccupancy = 0;
mFlushed = 0;
mStream = stream;
}
/// \brief Enqueue the specified copy descriptor, or launch immediately if no room is available
void enqueue(PxGpuCopyDesc& desc)
{
PX_ASSERT(desc.isValid());
if (desc.bytes == 0)
{
return;
}
if (mOccupancy < mReserved)
{
mBuffer[ mOccupancy++ ] = desc;
}
else
{
mDispatcher.launchCopyKernel(&desc, 1, mStream);
}
}
/// \brief Launch all copies queued since the last flush or reset
void flushEnqueued()
{
if (mOccupancy > mFlushed)
{
mDispatcher.launchCopyKernel(mBuffer + mFlushed, mOccupancy - mFlushed, mStream);
mFlushed = mOccupancy;
}
}
private:
PxGpuDispatcher& mDispatcher;
PxGpuCopyDesc* mBuffer;
CUstream mStream;
uint32_t mReserved;
uint32_t mOccupancy;
uint32_t mFlushed;
void operator=(const PxGpuCopyDescQueue&); // prevent a warning...
};
PX_POP_PACK
} // end physx namespace
#endif // PX_SUPPORT_GPU_PHYSX
#endif // PXCUDACONTEXTMANAGER_PXGPUCOPYDESCQUEUE_H
|