Installation:

• 30oct09 - tesla c1060 installation/cuda installation
• Tesla C1060 installation Guide (.pdf)
• Cuda sdk_release_notes_linux (.txt)
• nvidia hardware driver installation guide/notes
  
• tesla specs (deviceQuery)
• bandwidth test (bandwidthTest)
  

hardware documentation:

• Tesla c1060 computing processor board (board specification) (.pdf)
• 
  

Software documentation:

• Cuda programming guide V2.3.1. 145 pages (.pdf)
• Cuda reference manual V2.3  156 pages (.pdf)
  
• Cuda C programming  best pracitices guide tookit V2.3 66 pages. (.pdf)
• Cuda development tools 2.3: installation and verification on Linux 16 pages (.pdf)
• Cuda CUBLAS library v2.3 130 pages (.pdf)
• Cuda CUFFT library v2.3 21 pages (.pdf)
  
• Cuda Debugger: cuda-gdb V2.3 beta 22 pages (.pdf)
• Sample programs release notes

List of gpus at AO

Last updated 23feb22

30oct09 tesla c1060 installation:

• installed board in fajardo, softare in /local/nvidia
  
• downloaded from nvidia site:
• NVIDIA-Linux-x86_64-190.42-pkg2.run .. nvidia driver
• cudasdk_2.3_linux.run ... cuda software developement kit
  
• cudatoolkit_2.3_linux_64_rhel5.3.run  .. cuda toolkit
• Installed driver
• as root from /local/nvidia. sh NVIDIA-Linux-x86_64-190.42-pkg2.run
• Questions.. also installed 32 bit libs
• Ran /usr/bin/nvidia-xconfig -a  to setup xorg.conf
• -a will install all gpus.. ends up booting board name in xorg.conf device for nvidia
• ended up manually editing xorg.conf file:
• Section "ServerLayout"
      Identifier     "Default Layout"
      Screen      0  "Screen0"
      Screen      1   "Screen1"  RightOf "Screen0"
      InputDevice    "Mouse0" "CorePointer"
      InputDevice    "Keyboard0" "CoreKeyboard"
  EndSection
• Section "Device"
      Identifier     "Tesla"Section "Screen"
      Identifier     "Screen1"
      Device         "Tesla"
      Monitor        "Monitor1"
      DefaultDepth    24
      SubSection     "Display"
          Depth       24
      EndSubSection
  EndSection
• Section "Screen"
      Identifier     "Screen1"
      Device         "Tesla"
      Monitor        "Monitor1"
      DefaultDepth    24
      SubSection     "Display"
          Depth       24
      EndSubSection
  EndSection
• Left screen0 as xgi chip main screen
  
  
• Ran /usr/bin/nvidia-settings to test setup
• doesn't work remote. tried -c screen but didn't work
• Never got it to work. Kept getting confused with xgi chip
• just ignore this.
• Install cuda toolkit:
• sh cudatoolkit_2.3_linux_64_rhel5.3.run
• used /usr/local/cuda ... this is where all the junk goes.
  
• Need /usr/local/cuda/bin in your path.
• Install cuda sdk
• sh cudasdk_2.3_linux.run
• Got installed in /root/NVIDIA_GPU_Computing_SDK
• Each user needs own copy
• copied directory to /home/phil (had to temp change /root permissions to search it).
• Check installation: nvcc -V  .. prints out version
• Make sdk binaries
• fajardo didn't have -lglut (opengl) copied from summer1.
• had to create links  xx.so -> xx.so.3 --> xx.so.3.nnn
• cd /home/phil/NVxx/C; make
• binaries installed in ./bin/release

Our tesla board:

• ./bin/release/deviceQuery
• CUDA Device Query (Runtime API) version (CUDART static linking)
  There is 1 device supporting CUDA
• 
  

  tesla c1060 specs

  CUDA Driver Version:	2.30
  CUDA Runtime Version:	2.30
  CUDA Capability Major revision number	1
  CUDA Capability Minor revision number	3
  Total amount of global memory	4294705152 bytes
  Number of multiprocessors:	30
  Number of cores	240
  Total amount of constant memory	65536
  Total amount of shared memory per block	16384 bytes
  Total number of registers available per block:	16384
  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:	1.30 GHz
  Concurrent copy and execution:	Yes
  Run time limit on kernels	Yes
  Integrated	No
  Support host page-locked memory mapping	Yes
  Compute mode	default  (multiple host threads can use this device simultaneously)

• Bandwidth test
• ./bin/release/bandwidthTest
• Running on......
        device 0:Tesla C1060
  Quick Mode
  Host to Device Bandwidth for Pageable memory
  Transfer Size (Bytes)   Bandwidth(MB/s)
   33554432               2043.3
  Quick Mode
  Device to Host Bandwidth for Pageable memory
  
  Transfer Size (Bytes)   Bandwidth(MB/s)
   33554432               1733.0
  Quick Mode
  Device to Device Bandwidth
  
  Transfer Size (Bytes)   Bandwidth(MB/s)
   33554432               73268.9
  
  
  

<- page up
home_~phil