Dear all, No one was present at last Tuesday phone meeting. I think that maybe we should continue the discussion for a while only offline on the mailing list. What do you think? What do you also think about what I wrote some days ago (I copy it here)? Just a reminder for today's phone meeting. We could discuss the GPU accounting suggested by John Gordon. I think that, among the proposed solutions, I prefer the second one: 2. Repeat the computeusageblock - there is no identifier that would distinguish between 2 instances of the block. If there were then we could have two instances, one for CPU and one for GPU. The fields like, cpu count and total CPU would be different in each and could be treated separately or together because they are recorded together in the same UR. Another problem is that we do not identify the type of processor so we could not parse to distinguish between multiple blocks. I think we could just slightly modify the present definition by adding one or two field on the ComputeUsageBlock. Something like CPUArchitecture or CPUFamily. Then if both CPU and GPU are used the UR could just contain two block of this type one that include the usage for the CPU and one for the GPU. Also the UR document would require minor changes. Anyway I would wait for the end of the public comment on UR2.0 and then release a UR2.1 if necessary. While we are working to the new version we could also think to application accounting. This could be maybe solved by just adding a new block. Probably we should start investigating what are the fields required for it. Cheers, Andrea On 04/30/2013 12:41 PM, Andrea Cristofori wrote:
Dear all,
Just a reminder for today's phone meeting. We could discuss the GPU accounting suggested by John Gordon. I think that, among the proposed solutions, I prefer the second one:
2. Repeat the computeusageblock - there is no identifier that would distinguish between 2 instances of the block. If there were then we could have two instances, one for CPU and one for GPU. The fields like, cpu count and total CPU would be different in each and could be treated separately or together because they are recorded together in the same UR. Another problem is that we do not identify the type of processor so we could not parse to distinguish between multiple blocks.
I think we could just slightly modify the present definition by adding one or two field on the ComputeUsageBlock. Something like CPUArchitecture or CPUFamily. Then if both CPU and GPU are used the UR could just contain two block of this type one that include the usage for the CPU and one for the GPU. Also the UR document would require minor changes.
Anyway I would wait for the end of the public comment on UR2.0 and then release a UR2.1 if necessary. While we are working to the new version we could also think to application accounting. This could be maybe solved by just adding a new block. Probably we should start investigating what are the fields required for it.
Cheers, Andrea
Il 16/04/2013 14:37, Andrea Cristofori ha scritto:
Dear all,
Looks like today many of us could not join the phone meeting so I think it is better to postpone it. Next one will be in two weeks, the 30th of April at 15:00.
In the meanwhile if you have any idea or thought about what John Gordon proposed please send an email to the list so we can start the discussion via email and think about it before the meeting.
Regards Andrea
-- ur-wg mailing list ur-wg@ogf.org https://www.ogf.org/mailman/listinfo/ur-wg
-- Andrea Cristofori INFN-CNAF Viale Berti Pichat 6/2 40127 Bologna Italy Tel. : +39-051-6092920 Skype: andrea-cnaf