Nothing here since 18th, but Krembil don't send me short deadlines. IBM used to and I didn't fail to return them in time. I expect re-sends in 2 days (24th).

Mike

It is now over 6 days since the last batch were issue and I haven't seen any re-sends yet. Has anyone?

Mike

Got a few ARP1's since my last post (in a different forum). Currently have 7 RESENDS, In Progress all ..._138_2 or ...139_2

Previous one was received on 5/22

Hi Mike, Adri, and all:

I have wondered about the analysis strategies and workunit deployment tactics ARP uses for a couple of years now.

Do you know if the cell grid is square or hexagonal or some other shape? Do the cells have overlap zones?

And the one that vexes me most: It seems to me that successfully validated cells in a generation would have their outputs influence the adjacent cells in the next generation. If so, doesn't an unvalidated cell propagate a ring of unvalidated cells that grows with each successive generation? When the original unvalidated cell is eventually successfully validated, it seems to me there should be a release of workunits for that cell plus the adjacent cells for the next generation. This release of adjacent cells does not seem to be happening in your tracking of the advance of the most overdue cells.

If you know, please advise why a cell can advance through generations without requiring recalculation of the surrounding cells.

Thanks, Richard

I wonder if last day's contents of completed.txt are correct. It shows:

+--------------+---------------------+----------------+-------------+
| number_units | avg_hrs_to_complete | max_generation | type        |
+--------------+---------------------+----------------+-------------+
|            1 |               185.0 | 125            | Accelerated |
|           12 |               223.4 | 141            | Normal      |
+--------------+---------------------+----------------+-------------+

The max value for the Extreme category, according to state.txt, is 130:

+--------------+----------------+-------------+
| number_units | max_generation | type        |
+--------------+----------------+-------------+
|           83 | 135            | Accelerated |
|          401 | 130            | Extreme     |
|        35123 | 145            | Normal      |
+--------------+----------------+-------------+

You would say that type 'Accelerated' in completed.txt should be 'Extreme' now. However, it seems that none in the Accelerated category were completed lately. The effect seems to be that the data from the Extreme category have been transferred to the Accelerated category in some way. Surprising. That's what I gather.

Adri