Question for Uplinger/techs: On a hunch, does the 7.05 version number indicate you're now using the v7 API, and if so what is the exact number? (Saw someone speccing out something for GFAM 6.12 referring to 7.1.0 which is strange [to me], deviating from past protocol.

Got and crunched some CPU HCC betas, some with single task, plus some double-deckers. All seem OK.

The intial estimates of CPU time were close to the actual values for both types, but the times for the doubles were the same as for the singles. Is that supposed to happen? Don't the double WUs contain twice as much work?

[OT] I think it's been suggested in the forum before, but is it worth pursuing the CPU version of HCC now that the GPU crowd are gobbling up the work so rapidly? If the feed priority for HCC CPU tasks was decreased so that most of them go to the people who have only CPU-HCC selected, this would divert CPU resources to the other projects and be a much more efficient way to use those resources.
(For that reason I haven't crunched any HCC tasks since GPU beta testing started. I don't have any suitable GPUs and don't anticipate buying one unless the estimated end date of the HCC project is extended past its current ~ +120days by a considerable amount). [/OT]

"Don't the double WUs contain twice as much work?" Rhetoric?

The question here is, Rickjb, why [as I posted earlier in this thread], the single *beta* showed normal runtime and the "double-deckers" showed twice the TTC from point of arrival and of course, as we know from near 5 years of HCC project crunching, the average being quite steady.

As to the Beta, except for one double imaged sitting in PV, all were declared valid. One was a repair job btw, that arrived in the midst of the regulars.

The question was asked before [there's a Topic for it]. Since many have no access to GPU computing [it's really a specialist event], the proper is to me to assure that who computed CPU from the start [Nov.6, 2007] are allowed to compute for HCC to the end and hope myself to make the contribution goal I've set before the end. You are free to optimize your contributing resources. Others don't have that choice. It would be presumptive of WCG to take HCC CPU computing away, because it's more efficient to the volunteered resource. [We're eagerly waiting for a future choice that says]

1. Automate to send my comps work to it's best potential.
2. Send my comps the heavy duty work and only lighter when none is available.

Something along those lines, but those volunteering the resources, to be allowed a choice rather than a dictate.

We are using v7 API, as to what version, I'm not sure. That is a question for another tech. Also for GFAM, I am also not sure. Sorry for lack of information.

-Uplinger

Did you notice if the process went into zombie mode? Also, have you had issues with stuck work units in the past (even on other projects)?

The official stance right now on app_info is there is no official support. I personally have not had time to investigate the options and provide any useful information towards the conversation for app_info.

Thanks,
-Uplinger