cuda_arc

cuda_arc, examples which show how the Virginia Tech Advanced Research Computing (ARC) computer clusters can be used to execute jobs with CUDA, which is an extension to the C language that enables parallel programming on a Graphics Processing Unit (GPU);

Licensing:

The information on this web page is distributed under the MIT license.

Related Data and Programs:

cplex_slurm_arc, examples which uses the slurm() job scheduler to submit a cplex() job to Virginia Tech's Advanced Research Computing (ARC) computer cluster.

cuda_loop, a C code which shows how, in a CUDA program running on a Graphics Processing Unit (GPU), the choice of block and thread factors determines the allocation of tasks to processors.

Reference:

1. John Cheng, Max Grossman, Ty McKercher,
   Professional CUDA C Programming,
   John Wiley, 2014,
   ISBN: 978-1-118-73932-7.
2. Jason Sanders, Edward Kandrot,
   CUDA by Example,
   Addison Wesley, 2010,
   ISBN: 978-0-13-138768-3,
   LC: QA76.76.A65S255 2010.

Source Code:

COLLATZ searches for the longest Collatz sequence among the integer 1 through N.

• collatz.cu, is the CUDA program.
• collatz_hokiespeed.sh, commands to run the example on the ARC hokiespeed cluster.
• collatz_hokiespeed.txt, the output file.
• collatz_newriver.sh, commands to run the example on the ARC newriver cluster.
• collatz_newriver.txt, the output file.
• collatz_local.sh, commands to run the example locally.
• collatz_local.txt, the output file.

DENOISE process an image by applying an NEWS median filter.

• denoise.cu, is the CUDA program.
• denoise_hokiespeed.sh, commands to run the example on the ARC HokieSpeed cluster.
• denoise_newriver.sh, commands to run the example on the ARC NewRiver cluster.
• balloons_noisy.ascii.ppm, an ASCII PPM image, with noise.
• balloons_noisy.png, a PNG version of the image.
• balloons_median.ascii.ppm, the image, after median filtering.
• balloons_median.png, a PNG version of the image.

HELLO has 10 CUDA threads say "Hello!".

• hello.cu, is the CUDA program.
• hello_hokiespeed.sh, commands to run the example on the ARC hokiespeed cluster.
• hello_hokiespeed.txt, the output file.
• hello_newriver.sh, commands to run the example on the ARC newriver cluster.
• hello_newriver.txt, the output file.
• hello_local.sh, commands to run the example locally.
• hello_local.txt, the output file.

PROPERTIES counts CUDA devices and queries their properties.

• properties.cu, is the CUDA program.
• properties_hokiespeed.sh, commands to run the example on the ARC hokiespeed cluster.
• properties_hokiespeed.txt, the output file.
• properties_newriver.sh, commands to run the example on the ARC newriver cluster.
• properties_newriver.txt, the output file.
• properties_local.sh, commands to run the example locally.
• properties_local.txt, the output file.

VECADD does vector addition C=A+B.

• vecadd.cu, is the CUDA program.
• vecadd_hokiespeed.sh, commands to run the example on the ARC hokiespeed cluster.
• vecadd_hokiespeed.txt, the output file.
• vecadd_newriver.sh, commands to run the example on the ARC newriver cluster.
• vecadd_newriver.txt, the output file.
• vecadd_local.sh, commands to run the example locally.
• vecadd_local.txt, the output file.

You can go up one level to the EXAMPLES directory.