Do you support Giraph v1?

[khayyatzy]

Do you support Giraph v1?

[serafett]

Hi Zuhair,

We bulit Giraphx on Giraph v0.1. It is possible to modify it to use it also on Giraph v1 but currently we did not test it on Giraph v1. I think just copying the files will not be enough since there are additional modifications in Giraph v1.

You can download the attached folder to see the patch file for Giraphx and instructions on how to run it.

https://www.dropbox.com/s/gkj6a9vtnfg3tps/giraphx-all.zip

Best,

Serafettin

On Sunday, March 16, 2014 7:16 AM, Zuhair Khayyat [email protected] wrote:

Hi,
I would like to use your serialization algorithm on top of Giraph v1. Will it work? Which files I have to copy to the new Giraph?
—
Reply to this email directly or view it on GitHub.

[khayyatzy]

Hi,

Thanks for the file; it worked on Giraph 0.1. For the time being I will work on the old version of Giraph then later move to the newer version.

Where do you keep the implementation for the token ring based and "dining philosophers" based graph colouring? I would like to run on larger graphs and compare the difference between GraphX performance.

[serafett]

Use vj parameter to select a job type. Giraphx code is inside examples.Giraphx class. For example "-vj dGiraphx_coloring" runs dining philosopher for coloring problem. token ring coloring job runs using "compute_tGiraphx_coloring(Iterator, boolean)" but if you want to implement another algorithm, write the algorithm code in a method like "operate_tGiraph_coloring(Map<Long, String>)" and specify the method inside compute_tGiraphx method.

On Monday, March 17, 2014 7:39 AM, Zuhair Khayyat [email protected] wrote:

Hi,
Thanks for the file; it worked on Giraph 0.1. For the time being I will work on the old version of Giraph then later move to the newer version.
Where do you keep the implementation for the token ring based and "dining philosophers" based graph colouring? I would like to run on larger graphs and compare the difference between GraphX performance.
—
Reply to this email directly or view it on GitHub.

[khayyatzy]

Hi,

I understood from your paper you have a third algorithm for graph colouring other than token ring and "dining philosophers" algorithms. Do you have a third algorithm? or it is only two implementations?

Thank you

[serafett]

Yes only two implementations.

On Monday, March 17, 2014 11:36 AM, Zuhair Khayyat [email protected] wrote:

Hi,
I understood from your paper you have a third algorithm for graph colouring other than token ring and "dining philosophers" algorithms. Do you have a third algorithm? or it is only two implementations?
Thank you
—
Reply to this email directly or view it on GitHub.