You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hama.apache.org by "Edward J. Yoon" <ed...@apache.org> on 2011/09/01 07:27:40 UTC
Hama 0.4 (r.1163903) bench results
I ran the Bench example on 16 nodes Hama cluster, today.
* Hama 0.4 (r.1163903) and 3 tasks per node:
| 16 bytes | 1000 | 512 | 307.129 seconds |
* Hama 0.2 and 1 task per node:
| 16 bytes | 1000 | 512 | 507.837 seconds |
After fix HAMA-387, I'm plan to run this test on 1,024 nodes (10,000+
cores) Hama cluster.
--
Best Regards, Edward J. Yoon
@eddieyoon
Re: Hama 0.4 (r.1163903) bench results
Posted by Thomas Jungblut <th...@googlemail.com>.
Wow great, I think we can target asynchronous message transfer in
0.5.0 then. This is going to give us a speed boost as well as message
compression.
I'm so happy :)
2011/9/1 Edward J. Yoon <ed...@apache.org>:
> I ran the Bench example on 16 nodes Hama cluster, today.
>
> * Hama 0.4 (r.1163903) and 3 tasks per node:
> | 16 bytes | 1000 | 512 | 307.129 seconds |
>
> * Hama 0.2 and 1 task per node:
> | 16 bytes | 1000 | 512 | 507.837 seconds |
>
> After fix HAMA-387, I'm plan to run this test on 1,024 nodes (10,000+
> cores) Hama cluster.
>
> --
> Best Regards, Edward J. Yoon
> @eddieyoon
>
--
Thomas Jungblut
Berlin
mobile: 0170-3081070
business: thomas.jungblut@testberichte.de
private: thomas.jungblut@gmail.com