8000 Unable to run samples - cudaMalloc fills all the system memory · Issue #34 · NVIDIA/gvdb-voxels · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

Unable to run samples - cudaMalloc fills all the system memory #34

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Closed
AndreFrelicot opened this issue Oct 13, 2018 · 2 comments
Closed

Unable to run samples - cudaMalloc fills all the system memory #34

AndreFrelicot opened this issue Oct 13, 2018 · 2 comments

Comments

@AndreFrelicot
Copy link

I'm building on Windows 10 (tried on Linux Ubuntu also, same problem).

When I run a sample it blocks in function allocReduceStorage :
gvdb_1.1\shared_cudpp\src\cudpp\app\reduce_app.cu

[...]
void allocReduceStorage(CUDPPReducePlan *plan)
...
    case CUDPP_FLOAT:
-->        cudaMalloc(&plan->m_blockSums, blocks * sizeof(float));
        break;

System :
Version 10.0.17134 Build 17134
Processor i7-8750H
Graphics card: NVIDIA GeForce GTX 1070 with Max-Q Design
RAM: 32GB
GVDB: 1.1
CUDA: 10

`

  •   plan	0x000001f057310b30 {m_threadsPerBlock=256 m_maxBlocks=64 m_blockSums=0xcdcdcdcdcdcdcdcd }	CUDPPReducePlan *

  •   CUDPPPlan	{m_config={algorithm=CUDPP_REDUCE (3) op=CUDPP_MAX (3) datatype=CUDPP_FLOAT (6) ...} m_numElements=1000000 ...}	CUDPPPlan

  •   __vfptr	0x00007ffb96c87928 {cudpp_1915x64d.dll!void(* CUDPPReducePlan::`vftable'[2])()} {0x00007ffb96417662 {cudpp_1915x64d.dll!CUDPPReducePlan::`vector deleting destructor'(unsigned int)}}	void * *

  •   m_config	{algorithm=CUDPP_REDUCE (3) op=CUDPP_MAX (3) datatype=CUDPP_FLOAT (6) ...}	CUDPPConfiguration
  m_numElements	1000000	unsigned __int64
  m_numRows	1	unsigned __int64
  m_rowPitch	0	unsigned __int64

  •   m_planManager	0x000001f0571fe540 {m_deviceProps={name=0x000001f0571fe540 "GeForce GTX 1070 with Max-Q Design" uuid=...} }	CUDPPManager *
  m_threadsPerBlock	256	unsigned int
  m_maxBlocks	64	unsigned int
  m_blockSums	0xcdcdcdcdcdcdcdcd	void *

  •   (*((CUDPPPlan*)plan)).m_planManager->m_deviceProps	{name=0x000001f0571fe540 "GeForce GTX 1070 with Max-Q Design" uuid={bytes=0x000001f0571fe640 "ÀæÙþ?ÆÐPÀÇ_¢ñ9¬3... } ...}	cudaDeviceProp

  •   name	0x000001f0571fe540 "GeForce GTX 1070 with Max-Q Design"	char[256]

  •   uuid	{bytes=0x000001f0571fe640 "ÀæÙþ?ÆÐPÀÇ_¢ñ9¬3... }	CUuuid_st

  •   luid	0x000001f0571fe650 "�û"	char[8]
  luidDeviceNodeMask	1	unsigned int
  totalGlobalMem	8589934592	unsigned __int64
  sharedMemPerBlock	49152	unsigned __int64
  regsPerBlock	65536	int
  warpSize	32	int
  memPitch	2147483647	unsigned __int64
  maxThreadsPerBlock	1024	int

  •   maxThreadsDim	0x000001f0571fe684 {1024, 1024, 64}	int[3]

  •   maxGridSize	0x000001f0571fe690 {2147483647, 65535, 65535}	int[3]
  clockRate	1265500	int
  totalConstMem	65536	unsigned __int64
  major	6	int
  minor	1	int
  textureAlignment	512	unsigned __int64
  texturePitchAlignment	32	unsigned __int64
  deviceOverlap	1	int
  multiProcessorCount	16	int
  kernelExecTimeoutEnabled	1	int
  integrated	0	int
  canMapHostMemory	1	int
  computeMode	0	int
  maxTexture1D	131072	int
  maxTexture1DMipmap	16384	int
  maxTexture1DLinear	134217728	int

  •   maxTexture2D	0x000001f0571fe6e4 {131072, 65536}	int[2]

  •   maxTexture2DMipmap	0x000001f0571fe6ec {32768, 32768}	int[2]

  •   maxTexture2DLinear	0x000001f0571fe6f4 {131072, 65000, 2097120}	int[3]

  •   maxTexture2DGather	0x000001f0571fe700 {32768, 32768}	int[2]

  •   maxTexture3D	0x000001f0571fe708 {16384, 16384, 16384}	int[3]

  •   maxTexture3DAlt	0x000001f0571fe714 {8192, 8192, 32768}	int[3]
  maxTextureCubemap	32768	int

  •   maxTexture1DLayered	0x000001f0571fe724 {32768, 2048}	int[2]

  •   maxTexture2DLayered	0x000001f0571fe72c {32768, 32768, 2048}	int[3]

  •   maxTextureCubemapLayered	0x000001f0571fe738 {32768, 2046}	int[2]
  maxSurface1D	32768	int

  •   maxSurface2D	0x000001f0571fe744 {131072, 65536}	int[2]

  •   maxSurface3D	0x000001f0571fe74c {16384, 16384, 16384}	int[3]

  •   maxSurface1DLayered	0x000001f0571fe758 {32768, 2048}	int[2]

  •   maxSurface2DLayered	0x000001f0571fe760 {32768, 32768, 2048}	int[3]
  maxSurfaceCubemap	32768	int

  •   maxSurfaceCubemapLayered	0x000001f0571fe770 {32768, 2046}	int[2]
  surfaceAlignment	512	unsigned __int64
  concurrentKernels	1	int
  ECCEnabled	0	int
  pciBusID	1	int
  pciDeviceID	0	int
  pciDomainID	0	int
  tccDriver	0	int
  asyncEngineCount	5	int
  unifiedAddressing	1	int
  memoryClockRate	4004000	int
  memoryBusWidth	256	int
  l2CacheSize	2097152	int
  maxThreadsPerMultiProcessor	2048	int
  streamPrioritiesSupported	1	int
  globalL1CacheSupported	1	int
  localL1CacheSupported	1	int
  sharedMemPerMultiprocessor	98304	unsigned __int64
  regsPerMultiprocessor	65536	int
  managedMemory	1	int
  isMultiGpuBoard	0	int
  multiGpuBoardGroupID	0	int
  hostNativeAtomicSupported	0	int
  singleToDoublePrecisionPerfRatio	32	int
  pageableMemoryAccess	0	int
  concurrentManagedAccess	0	int
  computePreemptionSupported	0	int
  canUseHostPointerForRegisteredMem	0	int
  cooperativeLaunch	0	int
  cooperativeMultiDeviceLaunch	0	int
  sharedMemPerBlockOptin	0	unsigned __int64
  pageableMemoryAccessUsesHostPageTables	0	int
  directManagedMemAccessFromHost	0	int

`
@AndreFrelicot AndreFrelicot mentioned this issue Nov 16, 2018
Closed
@oursnoir
Copy link
oursnoir commented Nov 20, 2018 •
edited
Loading

Hello,

I had the exact same error (also eating up all my CPU RAM) under linux (ubuntu 18.04 with a fresh cuda 10 install from the nvidia repos).
The issue was resolved when I built the cudpp library in shared_cudpp and enforcing the compilation with architecture flags that match my GPU.
NB: I think this issue is related to #26 .
Best,

@AndreFrelicot
Copy link
Author

Thank you, I'll test that

http://arnon.dk/matching-sm-architectures-arch-and-gencode-for-various-nvidia-cards/

https://github.com/tpruvot/ccminer/wiki/Compatibility

@nathanchrs nathanchrs mentioned this issue Jun 7, 2019
Closed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@AndreFrelicot @oursnoir
0