[ art / civ / cult / cyb / diy / drg / feels / layer / lit / λ / q / r / sci / sec / tech / w / zzz ] archive provided by lainchan.jp

lainchan archive - /tech/ - 36164



File: 1491311240794.png (128.97 KB, 278x300, IMG_6052.png)

No.36164

I currently have 2 computers and one android phone (planning on buying another laptop or RPi). How can I modify the three to work on the processes running on my primary computer? For the sake of the argument, lets assume that I am creating a 200GiB encrypted container. How can the architectural difference between ARM and AMD-64 be handled while I am creating this container?

  No.36165

From what I've seen, clusters are generally given to compute parts of a full problem. If those depend on one another though, you have blocking and lose benefit of parallelism. By virtue of the problem type, you generally need to implement domain-specific solutions and can't really make a one-size-fit-all solution.

You might be interested in these:
http://bonsai.hgc.jp/~mdehoon/software/cluster/software.htm
https://computing.llnl.gov/tutorials/linux_clusters/

Apache Hadoop and Spark may also be worth looking into.

  No.36170

File: 1491321786213.png (61.3 KB, 200x98, Untitled-1-122jwp8.png)


  No.36177

If you want to learn about hardware pralellisation, learning about old school beowulf clusters may just be what you need.

  No.36179

I'm not sure how your example could work, generally you need specialized software for distributed computing. If you have actual problems that would benefit from a distributed architecture, I recommend writing programs for them with MPI and managing your nodes and tasks using Slurm.