'ion cuda'에 해당되는 글 2건

  1. 2012.06.02 CUDA devicequery - ION 330
  2. 2012.05.22 nvidia ion cuda core와 h.264 library
리플 룩 ion330 모델에 내장된 ion에 대한 devicequery이다.
2개의 MP가 존재해서 총 16개의 CUDA core가 존재한다.

~/NVIDIA_GPU_Computing_SDK/C/bin/linux/release$ ./deviceQuery
[deviceQuery] starting...

./deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

Found 1 CUDA Capable device(s)

Device 0: "ION"
  CUDA Driver Version / Runtime Version          4.2 / 4.2
  CUDA Capability Major/Minor version number:    1.1
  Total amount of global memory:                 254 MBytes (266010624 bytes)
  ( 2) Multiprocessors x (  8) CUDA Cores/MP:    16 CUDA Cores
  GPU Clock rate:                                1100 MHz (1.10 GHz)
  Memory Clock rate:                             800 Mhz
  Memory Bus Width:                              64-bit
  Max Texture Dimension Size (x,y,z)             1D=(8192), 2D=(65536,32768), 3D=(2048,2048,2048)
  Max Layered Texture Size (dim) x layers        1D=(8192) x 512, 2D=(8192,8192) x 512
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       16384 bytes
  Total number of registers available per block: 8192
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  768
  Maximum number of threads per block:           512
  Maximum sizes of each dimension of a block:    512 x 512 x 64
  Maximum sizes of each dimension of a grid:     65535 x 65535 x 1
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             256 bytes
  Concurrent copy and execution:                 No with 0 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            Yes
  Support host page-locked memory mapping:       Yes
  Concurrent kernel execution:                   No
  Alignment requirement for Surfaces:            Yes
  Device has ECC support enabled:                No
  Device is using TCC driver mode:               No
  Device supports Unified Addressing (UVA):      No
  Device PCI Bus ID / PCI location ID:           3 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 4.2, CUDA Runtime Version = 4.2, NumDevs = 1, Device = ION
[deviceQuery] test results...
PASSED

> exiting in 3 seconds:
3...2...1...done! 

그나저나.. 대역폭에서 내장형 그래픽이라 메인메모리를 공유하는데 왜 대역폭에서 이렇게 차이가 날까?
~/NVIDIA_GPU_Computing_SDK/C/bin/linux/release$ ./bandwidthTest
[bandwidthTest] starting...

./bandwidthTest Starting...

Running on...

 Device 0: ION
 Quick Mode

 Host to Device Bandwidth, 1 Device(s), Paged memory
   Transfer Size (Bytes)        Bandwidth(MB/s)
   33554432                     887.0

 Device to Host Bandwidth, 1 Device(s), Paged memory
   Transfer Size (Bytes)        Bandwidth(MB/s)
   33554432                     735.9

 Device to Device Bandwidth, 1 Device(s)
   Transfer Size (Bytes)        Bandwidth(MB/s)
   33554432                     5345.2

[bandwidthTest] test results...
PASSED

> exiting in 3 seconds: 3...2...1...done! 

8800GT 에 비하면 확실히 nbody 에서의 연산속도와 fps가 많이 떨어지는 느낌
(8800GT에서는 150fps에 50GFLOP/s 정도 나옴)

2010/11/02 - [Programming/openCL / CUDA] - CUDA 예제파일 실행결과 + SLI


+ 리눅스에서 nvidia 드라이버 버전 보는 방법
$ cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86 Kernel Module  295.40  Thu Apr  5 21:28:09 PDT 2012
GCC version:  gcc version 4.6.3 (Ubuntu/Linaro 4.6.3-1ubuntu5)

[링크 : http://www.nvnews.net/vbulletin/showthread.php?t=127289]   


신고

'Programming > openCL & CUDA' 카테고리의 다른 글

cuda 메모리별 접근 소요 클럭 사이클  (0) 2012.06.05
ubuntu 에서 vectorAdd 직접 컴파일 하기  (0) 2012.06.03
CUDA devicequery - ION 330  (0) 2012.06.02
cuda 5 preview  (0) 2012.06.02
nvidia ion cuda core와 h.264 library  (0) 2012.05.22
CUDA API 메모리 종류  (0) 2012.05.18
Posted by 구차니

댓글을 달아 주세요

2개의 mp를 내장하고 있는데
음. H.264 인코딩/디코딩 라이브러리 제한이 몇개부터드라...

scalar processor는 cuda core인데 1개의 mp에 8개 있고, 4개의 mp가 32 scalar 프로세서이니 
h.264 인코딩은 ion에서는 불가능 할 것으로 보인다

MPEG-2/VC-1 support 
 Decode Acceleration for G8x, G9x (Requires Compute 1.1 or higher) 
 Full Bitstream Decode for MCP79, MCP89, G98, GT2xx, GF1xx 
MPEG-2 CUDA accelerated decode with a GPUs with 8+ SMs (64 CUDA cores).  (Windows) 
 Supports HD (1080i/p) playback including Bluray content 
 R185+ (Windows), R260+ (Linux) 
 
H.264/AVCHD support 
 Baseline, Main, and High Profile, up to Level 4.1  
 Full Bitstream Decoding in hardware including HD (1080i/p) Bluray content 
 Supports B-Frames, bitrates up to 45 mbps 
 Available on NVIDIA GPUs:  G8x, G9x, MCP79, MCP89, G98, GT2xx, GF1xx 
 R185+ (Windows), R260+ (Linux) 

[출처 : CUDA_VideoDecoder_Library.pdf] 

 Supported on all CUDA-enabled GPUs with 32 scalar processor cores or more 
[출처 : CUDA_VideoEncoder_Library.pdf] 
[링크 :  http://www.vpac.org/files/GPU-Slides/05.CudaOptimization.pdf ]

Device 0: "ION"
  CUDA Driver Version:                                             2.30
  CUDA Runtime Version:                                           2.30
  CUDA Capability Major revision number:                 1
  CUDA Capability Minor revision number:                 1
  Total amount of global memory:                                 268435456 bytes
  Number of multiprocessors:                                     2
  Number of cores:                                                         16
  Total amount of constant memory:                         65536 bytes
  Total amount of shared memory per block:         16384 bytes
  Total number of registers available per block: 8192
  Warp size:                                                                     32
  Maximum number of threads per block:             512
  Maximum sizes of each dimension of a block:   512 x 512 x 64
  Maximum sizes of each dimension of a grid:     65535 x 65535 x 1
  Maximum memory pitch:                                           262144 bytes
  Texture alignment:                                                     256 bytes
  Clock rate:                                                                   1.10 GHz
  Concurrent copy and execution:                                 No
  Run time limit on kernels:                                     No
  Integrated:                                                                   Yes
  Support host page-locked memory mapping:         Yes
  Compute mode:                                                           Default (multiple host threads can use this devi
 
[링크 :  http://forums.nvidia.com/index.php?showtopic=100288 ]  

[링크 : http://www.nvidia.com/object/picoatom_specifications.html ]
[링크 : http://en.wikipedia.org/wiki/Nvidia_Ion ]
신고

'Programming > openCL & CUDA' 카테고리의 다른 글

CUDA devicequery - ION 330  (0) 2012.06.02
cuda 5 preview  (0) 2012.06.02
nvidia ion cuda core와 h.264 library  (0) 2012.05.22
CUDA API 메모리 종류  (0) 2012.05.18
Interoperability (상호운용성)  (0) 2012.05.04
cuda 내장변수  (0) 2012.04.30
Posted by 구차니

댓글을 달아 주세요